paperID
stringlengths 36
36
| pwc_id
stringlengths 8
47
| arxiv_id
stringlengths 6
16
⌀ | nips_id
float64 | url_abs
stringlengths 18
329
| url_pdf
stringlengths 18
742
| title
stringlengths 8
325
| abstract
stringlengths 1
7.27k
⌀ | authors
stringlengths 2
7.06k
| published
stringlengths 10
10
⌀ | conference
stringlengths 12
47
⌀ | conference_url_abs
stringlengths 16
198
⌀ | conference_url_pdf
stringlengths 27
199
⌀ | proceeding
stringlengths 6
47
⌀ | taskID
stringlengths 7
1.44k
| areaID
stringclasses 688
values | embedding
stringlengths 9.26k
12.5k
| umap_embedding
stringlengths 29
44
|
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d42e3176-008e-45f7-ac40-d9306e4989c4 | trafficbots-towards-world-models-for | 2303.04116 | null | https://arxiv.org/abs/2303.04116v1 | https://arxiv.org/pdf/2303.04116v1.pdf | TrafficBots: Towards World Models for Autonomous Driving Simulation and Motion Prediction | Data-driven simulation has become a favorable way to train and test autonomous driving algorithms. The idea of replacing the actual environment with a learned simulator has also been explored in model-based reinforcement learning in the context of world models. In this work, we show data-driven traffic simulation can be formulated as a world model. We present TrafficBots, a multi-agent policy built upon motion prediction and end-to-end driving, and based on TrafficBots we obtain a world model tailored for the planning module of autonomous vehicles. Existing data-driven traffic simulators are lacking configurability and scalability. To generate configurable behaviors, for each agent we introduce a destination as navigational information, and a time-invariant latent personality that specifies the behavioral style. To improve the scalability, we present a new scheme of positional encoding for angles, allowing all agents to share the same vectorized context and the use of an architecture based on dot-product attention. As a result, we can simulate all traffic participants seen in dense urban scenarios. Experiments on the Waymo open motion dataset show TrafficBots can simulate realistic multi-agent behaviors and achieve good performance on the motion prediction task. | ['Luc van Gool', 'Fisher Yu', 'Dengxin Dai', 'Alexander Liniger', 'Zhejun Zhang'] | 2023-03-07 | null | null | null | null | ['motion-prediction'] | ['computer-vision'] | [-4.66619372e-01 8.33950862e-02 -1.09730829e-02 -4.61525112e-01
-2.90300250e-01 -3.34423542e-01 9.79896605e-01 -3.45089048e-01
-6.57230020e-01 7.27674127e-01 1.25077069e-01 -2.10724980e-01
-1.14041433e-01 -1.15790021e+00 -7.33652592e-01 -7.51540005e-01
-1.52885169e-01 1.13478410e+00 4.93304968e-01 -8.93058181e-01
9.19498503e-02 5.22254288e-01 -1.82573545e+00 8.69869068e-02
6.90153122e-01 3.21232975e-01 8.01625848e-01 8.08522820e-01
-7.76995420e-02 8.18245649e-01 -3.93228233e-01 -2.48185664e-01
2.62792498e-01 -3.77516687e-01 -5.55749357e-01 3.92501429e-02
8.49253237e-02 -3.31331342e-01 -7.26981521e-01 5.11774480e-01
5.28026521e-01 5.49682915e-01 4.85809505e-01 -1.92951727e+00
3.35808322e-02 6.13150358e-01 1.58935219e-01 -2.78771874e-02
2.64025867e-01 8.79764974e-01 6.12705767e-01 -3.47315550e-01
1.05702972e+00 1.44718838e+00 1.84850425e-01 9.67599094e-01
-1.20007479e+00 -3.36183429e-01 2.27507025e-01 7.53411055e-01
-1.09433961e+00 -1.07135199e-01 7.81818151e-01 -3.73972535e-01
1.06167889e+00 1.99435726e-02 1.06165266e+00 1.68134105e+00
5.40991068e-01 7.53758669e-01 8.68355632e-01 8.11342224e-02
7.36183465e-01 1.39699116e-01 -1.87026843e-01 8.36350918e-01
1.39414355e-01 6.47742629e-01 -3.24533135e-01 1.82180166e-01
4.53673124e-01 -2.82585919e-01 2.04961210e-01 -1.06253219e+00
-1.30499959e+00 1.00723660e+00 6.17718756e-01 -6.68331608e-02
-5.41221023e-01 7.18222082e-01 3.76296610e-01 1.37888253e-01
-1.11235097e-01 4.90174353e-01 -1.97281808e-01 -5.10939717e-01
-3.54676515e-01 1.08533752e+00 8.88646901e-01 1.10539663e+00
9.56608474e-01 3.87188554e-01 -3.92319381e-01 3.92320305e-01
2.97510624e-01 8.80973637e-01 5.26057124e-01 -1.41238856e+00
3.13360184e-01 3.51661295e-01 4.24040139e-01 -8.68961513e-01
-8.14192295e-01 -3.48924875e-01 -3.39016706e-01 6.35973692e-01
3.48070323e-01 -3.38811368e-01 -9.08499599e-01 1.96225762e+00
3.14493507e-01 5.37641108e-01 2.22538128e-01 1.03967667e+00
4.19272780e-01 6.64594650e-01 2.80389875e-01 3.61775637e-01
1.20449984e+00 -1.31620419e+00 -6.52980924e-01 -2.67208159e-01
9.79451180e-01 2.58282642e-03 9.23990428e-01 2.48188153e-01
-9.78353798e-01 -7.00328410e-01 -9.84796166e-01 1.41027734e-01
-7.65432894e-01 -3.27520430e-01 5.92342198e-01 5.80990911e-01
-1.19939721e+00 4.98436391e-01 -1.06345129e+00 -4.26851511e-01
1.42063200e-01 3.10947776e-01 -4.13793594e-01 -4.41464931e-02
-1.17136264e+00 1.43073666e+00 3.79651159e-01 -2.92213619e-01
-1.56069589e+00 -2.56262362e-01 -1.02749598e+00 -5.32146059e-02
3.51967633e-01 -1.15091968e+00 1.27878761e+00 -3.66202027e-01
-1.84596550e+00 3.56959254e-01 -6.25539273e-02 -7.53132224e-01
6.39311731e-01 1.41364694e-01 -3.63003254e-01 -1.42105490e-01
2.73219705e-01 1.06488657e+00 5.68901062e-01 -1.41075099e+00
-7.66998827e-01 3.70020978e-02 2.55088359e-01 4.76181090e-01
2.14278892e-01 -6.65740013e-01 -4.67573494e-01 -1.39709398e-01
-4.26213861e-01 -1.37960231e+00 -8.00355792e-01 -1.31853595e-01
-2.19011270e-02 -1.49933591e-01 7.98749030e-01 -1.75923947e-02
7.49626875e-01 -1.99396884e+00 4.42035019e-01 1.26027226e-01
5.68147823e-02 1.33533448e-01 -6.19944751e-01 5.97466886e-01
3.10844451e-01 -2.83819497e-01 -7.48907626e-02 -5.85320592e-01
4.46888953e-01 6.99334800e-01 -1.13303542e-01 1.86460942e-01
2.67367549e-02 1.10929084e+00 -1.13939691e+00 -2.87766278e-01
5.94064713e-01 3.98499340e-01 -8.79080236e-01 1.94016501e-01
-6.91900074e-01 6.64890170e-01 -6.70837343e-01 -1.38684988e-01
5.78397334e-01 2.24434495e-01 -5.88472225e-02 2.29552120e-01
-3.27444136e-01 2.53451765e-01 -1.19481111e+00 1.93216085e+00
-8.02486479e-01 5.85971296e-01 -1.25158653e-01 -6.83283687e-01
8.87096524e-01 3.34093459e-02 3.64757538e-01 -1.02113819e+00
1.99928373e-01 2.94340681e-02 2.40316272e-01 -7.06378520e-01
8.20973217e-01 1.32526919e-01 -3.41632813e-01 3.02566975e-01
8.63957182e-02 -5.11467993e-01 5.24378896e-01 9.89035890e-02
1.15947080e+00 5.36138773e-01 -2.69465059e-01 -3.32036503e-02
3.43920231e-01 6.27814174e-01 3.30361128e-01 9.04474556e-01
-2.27918163e-01 2.96423614e-01 3.09991121e-01 -6.06190801e-01
-1.30056810e+00 -1.04949963e+00 3.28957170e-01 1.15609050e+00
4.34720337e-01 -3.88877213e-01 -8.56095016e-01 -4.06331092e-01
-5.45335002e-02 1.18535841e+00 -6.69750154e-01 -3.04015994e-01
-1.08058786e+00 -6.21094584e-01 3.24361265e-01 2.79295206e-01
4.95179057e-01 -1.46321130e+00 -1.06400275e+00 5.96136689e-01
-1.31108388e-01 -1.27506900e+00 -9.07249823e-02 1.78352222e-01
-2.86893368e-01 -6.61576927e-01 -3.84866863e-01 -6.81537509e-01
2.33380690e-01 2.69712478e-01 9.82463419e-01 -1.95249900e-01
2.62806378e-02 4.90997642e-01 -3.05024654e-01 -2.97088534e-01
-7.74305224e-01 2.39485890e-01 1.35688752e-01 1.43156713e-02
1.12704284e-01 -5.37472963e-01 -5.83946109e-01 3.39111894e-01
-5.94526887e-01 4.34378982e-01 4.30678099e-01 6.90038979e-01
1.72257185e-01 -3.13024163e-01 3.86450768e-01 -4.10924822e-01
6.06902361e-01 -5.95035970e-01 -6.98174894e-01 -2.26294219e-01
-3.45504671e-01 4.13675874e-01 7.71231174e-01 -4.07162905e-01
-9.87470925e-01 1.58074647e-01 -3.61490190e-01 -2.05288500e-01
-4.68319356e-01 2.14844465e-01 -2.49711767e-01 1.28081217e-01
6.75601780e-01 2.49042228e-01 1.24591619e-01 -7.13248998e-02
7.57041156e-01 2.87902117e-01 3.33140433e-01 -6.04842007e-01
7.61327565e-01 6.18371189e-01 3.27590108e-01 -7.34963655e-01
4.55090366e-02 -8.36314857e-02 -3.43076110e-01 -4.09246117e-01
1.07580054e+00 -7.78424561e-01 -1.11633635e+00 4.27898645e-01
-1.15707302e+00 -1.02098393e+00 -5.22528052e-01 7.47029543e-01
-1.43039274e+00 -1.65197775e-01 -1.61031932e-01 -5.09029090e-01
4.76444542e-01 -1.60481024e+00 9.86320555e-01 1.30232945e-01
-8.18206146e-02 -9.74583328e-01 4.34823543e-01 -5.40902279e-02
6.98912740e-01 1.36442378e-01 7.37918735e-01 -4.93557394e-01
-9.52117860e-01 9.12189782e-02 2.42899299e-01 -3.66223067e-01
-3.26953769e-01 -3.30851644e-01 -6.02858722e-01 -7.85249472e-02
-3.30671102e-01 -4.88371141e-02 6.70852423e-01 2.98997819e-01
7.75722623e-01 1.12467175e-02 -7.30616331e-01 6.47404909e-01
1.29030085e+00 3.29812080e-01 6.61795080e-01 6.32688522e-01
6.53847277e-01 7.96006143e-01 6.99994266e-01 4.44144100e-01
9.11767244e-01 1.18241584e+00 8.81738961e-01 1.62365288e-01
-2.08944544e-01 -5.72190940e-01 3.45721841e-01 3.07933390e-01
5.40975772e-04 -5.35401583e-01 -8.87762964e-01 6.33167684e-01
-2.15978122e+00 -1.30916297e+00 -2.67417014e-01 1.86234295e+00
8.97717178e-02 2.60362804e-01 3.59929949e-01 -4.00492221e-01
3.08346152e-01 2.44341061e-01 -4.64518845e-01 -6.12686038e-01
3.34208198e-02 -4.37468402e-02 5.75014532e-01 8.73722434e-01
-8.17295492e-01 1.36263323e+00 5.97784615e+00 7.82367766e-01
-1.18892634e+00 1.36322752e-01 7.93667585e-02 -1.35833532e-01
-4.09504026e-01 -6.10763021e-02 -8.52705419e-01 5.75163424e-01
1.25761139e+00 -1.36946902e-01 5.28626621e-01 9.92114246e-01
7.38197446e-01 -8.96462202e-02 -9.94934201e-01 6.88796520e-01
-2.93175131e-01 -1.41511083e+00 1.22359343e-01 1.35636821e-01
5.00624061e-01 3.88072044e-01 1.74582571e-01 8.37300241e-01
7.44773805e-01 -9.92350399e-01 8.09610486e-01 6.03407800e-01
1.89955726e-01 -6.70364439e-01 4.34375226e-01 6.96201980e-01
-9.68116522e-01 -2.56729692e-01 -3.55006456e-01 -9.25781205e-02
6.59868836e-01 -3.64625812e-01 -9.93562579e-01 3.78024369e-01
3.57898265e-01 5.39081931e-01 -4.39336836e-01 1.11969197e+00
-3.43865454e-02 1.61490619e-01 -2.78801948e-01 -5.67871630e-01
7.36229777e-01 -2.97889322e-01 8.20075691e-01 1.05309677e+00
3.77095789e-01 -3.51182640e-01 3.38439614e-01 7.83134639e-01
5.32386839e-01 -1.65772498e-01 -1.05103397e+00 6.04320228e-01
1.75685719e-01 1.09172082e+00 -5.03555477e-01 -4.17683572e-01
-1.86054751e-01 8.54507089e-01 3.73559475e-01 4.67107743e-01
-1.11025310e+00 -4.56272326e-02 1.05560839e+00 7.21249953e-02
5.14535964e-01 -6.49352312e-01 -3.39543424e-03 -8.37870479e-01
-3.66328150e-01 -5.99992394e-01 -2.70438284e-01 -9.35867429e-01
-6.27054036e-01 8.52570355e-01 4.65394288e-01 -1.35532153e+00
-8.04061115e-01 -6.84805453e-01 -6.04956985e-01 6.74987733e-01
-1.34793139e+00 -1.14203548e+00 -5.76827705e-01 5.95880508e-01
6.64873898e-01 -4.15804565e-01 8.27293396e-01 1.42168984e-01
-3.05183470e-01 1.13730714e-01 -6.40012696e-02 -2.54945189e-01
2.46284992e-01 -1.19575393e+00 9.48955774e-01 3.73715818e-01
-1.13853123e-02 1.15448736e-01 1.22186804e+00 -4.85366225e-01
-1.40324247e+00 -1.28046143e+00 5.50113857e-01 -6.14462972e-01
5.30783772e-01 -4.76839781e-01 -6.23287320e-01 6.47640944e-01
2.57561117e-01 -1.20181993e-01 6.12605363e-02 -3.99014205e-01
1.12971060e-01 -7.59914666e-02 -1.10356343e+00 1.30212164e+00
1.26685512e+00 7.11239278e-02 -2.77295023e-01 1.21122621e-01
6.46798491e-01 -4.21249956e-01 -1.60940185e-01 5.86937591e-02
4.49282557e-01 -1.05029678e+00 7.78219163e-01 -6.92175150e-01
2.13291328e-02 -4.71416771e-01 -1.36651859e-01 -1.79766309e+00
-4.98132497e-01 -5.31268418e-01 7.94698820e-02 5.25458217e-01
4.08672988e-01 -5.84673584e-01 9.63979125e-01 4.80664462e-01
-3.78382385e-01 -4.55883771e-01 -1.09365809e+00 -7.87019253e-01
1.02256313e-01 -7.11495161e-01 9.17694986e-01 3.12599957e-01
-8.59760121e-02 2.97188133e-01 -5.02574384e-01 7.92124793e-02
3.34416330e-01 -2.09479347e-01 1.26633465e+00 -8.16335201e-01
-3.05341870e-01 -5.87616742e-01 -6.65124595e-01 -1.30995214e+00
4.31134194e-01 -8.14315319e-01 3.02206725e-01 -1.60291183e+00
-2.23697439e-01 -6.32275641e-01 1.93584532e-01 7.10844621e-02
2.74390280e-01 -1.05688190e-02 3.82321805e-01 -3.39451671e-01
-6.19828343e-01 8.23112369e-01 1.51473153e+00 -8.38324502e-02
-3.18839788e-01 -4.10733558e-02 -1.21271618e-01 4.46362913e-01
1.03638887e+00 -4.11463857e-01 -7.75816560e-01 -4.62552875e-01
1.86007485e-01 4.22886580e-01 5.51887214e-01 -1.36893439e+00
3.29419822e-01 -4.15265113e-01 -1.97291300e-01 -6.26866043e-01
1.00751507e+00 -9.00103152e-01 2.65888393e-01 7.33406425e-01
-2.80206293e-01 3.09288442e-01 1.54672325e-01 5.57606518e-01
1.48689434e-01 -7.36404955e-02 5.66912115e-01 -2.00207427e-01
-1.11982429e+00 4.53117132e-01 -1.02110791e+00 -2.44550005e-01
1.35520744e+00 -1.78007841e-01 -3.71933699e-01 -5.43569922e-01
-8.88577402e-01 5.93669236e-01 5.13237119e-01 8.18120837e-01
5.85865557e-01 -1.33061731e+00 -8.09077442e-01 4.24079835e-01
1.61468312e-01 -2.68474281e-01 4.49373722e-01 5.13469100e-01
-7.66106129e-01 4.05741125e-01 -7.71264493e-01 -5.86468041e-01
-7.30583787e-01 7.84087896e-01 6.04283571e-01 -2.26294473e-01
-6.71424925e-01 2.31359914e-01 2.02726528e-01 -6.20317638e-01
-8.78239870e-02 -3.57854575e-01 -3.75227213e-01 -2.99223006e-01
2.23759249e-01 3.98704648e-01 -1.65381327e-01 -8.98634434e-01
-1.69751629e-01 4.01245177e-01 2.46439606e-01 -6.96120203e-01
1.08778155e+00 -1.46800965e-01 5.31937242e-01 3.51541787e-01
8.39754701e-01 -2.15185851e-01 -1.58172345e+00 2.90052772e-01
-2.47593045e-01 -1.84350595e-01 -2.46899262e-01 -6.37537062e-01
-7.20651507e-01 8.09308350e-01 7.11273968e-01 2.35348195e-01
3.74003112e-01 -1.67012885e-01 6.65685177e-01 6.20898008e-01
8.93212914e-01 -9.24099863e-01 8.50462466e-02 8.29701662e-01
8.47438037e-01 -1.15544999e+00 -6.66081786e-01 -5.38496226e-02
-9.74462509e-01 7.68325269e-01 9.95592117e-01 -4.24411923e-01
5.50379574e-01 2.39706308e-01 2.01450437e-01 -7.70720020e-02
-1.27613771e+00 -7.46583939e-01 -2.18096986e-01 1.17573178e+00
-3.07393104e-01 2.87750900e-01 -2.30081216e-01 3.33457232e-01
-5.93696892e-01 -7.05509856e-02 8.56163740e-01 7.05789685e-01
-5.99356651e-01 -1.27119505e+00 -1.13973573e-01 -4.81494702e-02
2.91354835e-01 4.95077193e-01 1.83232471e-01 1.09612644e+00
2.78979689e-01 9.09601092e-01 3.73217404e-01 -5.74130297e-01
4.74066228e-01 -9.26134363e-02 4.59915131e-01 -5.36945879e-01
-4.71820861e-01 -4.89371061e-01 2.79179066e-01 -8.50422919e-01
-2.36717798e-03 -5.64916551e-01 -1.37418056e+00 -5.40326655e-01
3.86855781e-01 2.17515707e-01 1.02904892e+00 7.84370899e-01
4.36590463e-01 7.00609863e-01 5.57323933e-01 -1.26458251e+00
-2.31607288e-01 -6.93176270e-01 -3.03323239e-01 5.18061578e-01
4.69678760e-01 -9.66828704e-01 4.02506180e-02 -3.38201195e-01] | [5.096607685089111, 1.0661993026733398] |
cb3c7c41-3ae2-4b5a-bda4-aa6124494889 | raddet-range-azimuth-doppler-based-radar | 2105.00363 | null | https://arxiv.org/abs/2105.00363v1 | https://arxiv.org/pdf/2105.00363v1.pdf | RADDet: Range-Azimuth-Doppler based Radar Object Detection for Dynamic Road Users | Object detection using automotive radars has not been explored with deep learning models in comparison to the camera based approaches. This can be attributed to the lack of public radar datasets. In this paper, we collect a novel radar dataset that contains radar data in the form of Range-Azimuth-Doppler tensors along with the bounding boxes on the tensor for dynamic road users, category labels, and 2D bounding boxes on the Cartesian Bird-Eye-View range map. To build the dataset, we propose an instance-wise auto-annotation method. Furthermore, a novel Range-Azimuth-Doppler based multi-class object detection deep learning model is proposed. The algorithm is a one-stage anchor-based detector that generates both 3D bounding boxes and 2D bounding boxes on Range-Azimuth-Doppler and Cartesian domains, respectively. Our proposed algorithm achieves 56.3% AP with IOU of 0.3 on 3D bounding box predictions, and 51.6% with IOU of 0.5 on 2D bounding box prediction. Our dataset and the code can be found at https://github.com/ZhangAoCanada/RADDet.git. | ['Robert Laganiere', 'Farzan Erlik Nowruzi', 'Ao Zhang'] | 2021-05-02 | null | null | null | null | ['radar-object-detection'] | ['robots'] | [-2.82718718e-01 -3.72158736e-01 1.13900959e-01 -7.48543441e-01
-9.38642979e-01 -6.66626394e-01 8.04897904e-01 -4.81465161e-01
-2.57209390e-01 3.49312812e-01 3.22460383e-02 -3.63615960e-01
-3.14582139e-01 -9.15533781e-01 -5.73526740e-01 -5.53018808e-01
-1.65577576e-01 6.77909315e-01 3.24676186e-01 -2.08752453e-01
2.78939754e-01 8.75217140e-01 -1.22828722e+00 1.04426675e-01
3.91890377e-01 1.39303637e+00 -2.11445391e-01 8.48675728e-01
3.07344139e-01 5.27529359e-01 -6.81616664e-01 -4.99888510e-01
8.47332656e-01 2.61165828e-01 1.53736621e-01 -3.08514178e-01
1.01699317e+00 -7.34641016e-01 -8.04325163e-01 1.01729417e+00
4.50043499e-01 4.77358252e-02 7.68849909e-01 -1.49180496e+00
-7.04852998e-01 6.25332892e-02 -7.04181850e-01 2.76187778e-01
-1.51385769e-01 4.59896773e-02 9.92036164e-01 -1.05164385e+00
3.51265341e-01 1.29172945e+00 5.47055960e-01 1.93907857e-01
-7.27059007e-01 -1.07107341e+00 1.85647868e-02 4.30721551e-01
-1.56360579e+00 -7.08029419e-02 7.60707021e-01 -7.30419278e-01
6.78464651e-01 2.75728196e-01 4.27023560e-01 1.06862950e+00
4.48563695e-01 4.32099611e-01 1.02030051e+00 2.66398847e-01
-7.40588177e-03 -2.09186807e-01 3.65055680e-01 6.59989715e-01
5.48134685e-01 7.40822852e-01 -2.18314856e-01 5.16843535e-02
7.25122929e-01 2.09901646e-01 2.80096471e-01 -4.82984573e-01
-1.14508605e+00 9.09196913e-01 6.30731404e-01 -3.12278628e-01
-1.77907348e-01 3.32083762e-01 1.22389063e-01 -7.62430504e-02
6.53996587e-01 1.50284201e-01 -3.70181173e-01 8.39698017e-02
-7.19128609e-01 5.20122945e-01 5.79448760e-01 1.17175031e+00
6.35850668e-01 2.92497337e-01 6.19696267e-02 6.31972611e-01
4.31220710e-01 1.19482422e+00 -2.67468363e-01 -6.60360754e-01
3.97160351e-01 4.95192081e-01 2.55413741e-01 -1.19585323e+00
-8.11128736e-01 -6.82568252e-01 -7.23565340e-01 4.66241956e-01
3.03543001e-01 -2.69926816e-01 -1.10986161e+00 1.24004984e+00
5.13074458e-01 1.24082677e-01 6.39940500e-02 1.40203142e+00
8.94203722e-01 6.78162158e-01 -3.12146217e-01 4.21525806e-01
1.75460923e+00 -8.73850405e-01 -5.23918569e-01 -3.28673512e-01
5.23780882e-01 -7.66766846e-01 3.70933115e-01 2.77347118e-01
-3.05463433e-01 -6.74803734e-01 -1.07809746e+00 3.01925689e-02
-5.64569235e-01 5.75128198e-01 4.97715652e-01 7.02891350e-01
-4.16370511e-01 -1.00891083e-01 -5.98285973e-01 9.02294740e-02
4.68962520e-01 -1.33391961e-01 -2.20427126e-01 -2.05406934e-01
-1.13276494e+00 1.02262223e+00 1.59727514e-01 3.17142159e-01
-1.18022394e+00 -8.93754065e-01 -7.28440762e-01 -3.32007438e-01
3.75502497e-01 -2.30655044e-01 1.06025410e+00 -8.15826561e-03
-8.74792039e-01 5.19285798e-01 2.52187997e-01 -6.87033057e-01
7.14895904e-01 -5.99314928e-01 -9.09047127e-01 -9.74950790e-02
1.69343844e-01 6.34852588e-01 9.40695047e-01 -1.18141830e+00
-9.68203127e-01 -5.25405049e-01 4.87802699e-02 -3.81013304e-02
3.22541893e-01 -1.16178855e-01 -2.96820551e-01 -6.92987859e-01
2.46998936e-01 -1.02519929e+00 -5.49021401e-02 -1.03159755e-01
-4.84589100e-01 -8.50910768e-02 1.00048470e+00 -7.53216326e-01
8.99493515e-01 -2.13929057e+00 -3.92835021e-01 1.50978222e-01
4.09602851e-01 2.14835107e-02 2.66919495e-03 1.53041020e-01
4.20838557e-02 -2.00397626e-01 2.38463320e-02 6.93433583e-02
2.88786352e-01 -1.55629620e-01 -7.42761314e-01 8.36553931e-01
4.19323832e-01 5.59598684e-01 -6.32579863e-01 -8.39877203e-02
2.72720724e-01 5.13596177e-01 -3.54699492e-01 4.73104268e-02
-5.39839789e-02 2.64209270e-01 -4.53532904e-01 1.14492130e+00
1.22158635e+00 2.51885682e-01 -3.77305180e-01 -7.43464053e-01
-4.60135221e-01 1.20536558e-01 -1.08503306e+00 8.93803298e-01
-1.38798147e-01 1.05276859e+00 -2.22875491e-01 -6.23618543e-01
1.35531795e+00 -1.28552765e-01 4.17574614e-01 -9.39018846e-01
1.83033586e-01 9.99584123e-02 2.96840221e-01 -1.09294258e-01
7.66417146e-01 -5.43130077e-02 -4.84386057e-01 3.07527065e-01
-2.17901617e-01 7.26386830e-02 9.72524881e-02 5.23963049e-02
1.15240061e+00 5.28134182e-02 -6.60704374e-02 8.72561410e-02
3.55110228e-01 2.65609384e-01 5.48847735e-01 8.43076050e-01
-2.18230322e-01 4.97077674e-01 3.18652630e-01 -9.58951771e-01
-1.21296954e+00 -1.27197433e+00 -5.90118647e-01 9.78486896e-01
1.54053167e-01 -1.52323708e-01 -2.10635364e-01 -5.50240636e-01
3.91806126e-01 6.77116930e-01 -5.90233147e-01 -1.18021786e-01
-6.41094387e-01 -7.49610603e-01 7.07425535e-01 7.10578918e-01
5.64166963e-01 -3.44422817e-01 -8.29951942e-01 -1.17752589e-01
1.19482391e-01 -1.47044015e+00 -2.21786246e-01 -8.28600600e-02
-5.96573055e-01 -1.04228425e+00 -4.34194207e-01 -1.44918501e-01
4.63660091e-01 4.34158593e-01 8.12739849e-01 -5.78323781e-01
-5.63767374e-01 2.17510477e-01 -3.81156594e-01 -8.16500962e-01
2.90234596e-01 -2.57759243e-01 4.25656259e-01 9.48585570e-02
7.41721511e-01 -2.04554096e-01 -6.13603055e-01 6.88933372e-01
-3.33371371e-01 1.16395742e-01 1.01927674e+00 5.05584002e-01
4.71785665e-01 -3.63921791e-01 3.14827681e-01 -4.73667383e-01
1.05533287e-01 -5.53744972e-01 -1.34493887e+00 -2.10997120e-01
-3.31239671e-01 -1.82391778e-01 2.98643738e-01 -2.58786857e-01
-6.81017101e-01 9.68039483e-02 1.47793651e-01 -6.76841140e-01
-3.35620105e-01 2.57565349e-01 -5.60315792e-03 -1.10859647e-01
6.38955593e-01 1.70749560e-01 -3.44899625e-01 -5.49505830e-01
5.81596792e-01 7.44411767e-01 6.10101342e-01 -2.86421299e-01
1.34828639e+00 7.43763924e-01 6.34558424e-02 -6.29423857e-01
-1.17022395e+00 -5.37476480e-01 -7.80300856e-01 -5.50284803e-01
9.18326378e-01 -1.26747417e+00 -6.24965549e-01 2.10655600e-01
-1.02348506e+00 -3.93126048e-02 2.48546585e-01 7.65158713e-01
-3.38683754e-01 1.14835724e-01 -1.19218275e-01 -1.07277179e+00
-3.91126305e-01 -9.00723577e-01 1.29767287e+00 2.92000026e-02
1.93282828e-01 -3.47380728e-01 2.55013146e-02 6.68056726e-01
3.65292966e-01 3.81017774e-01 5.05653679e-01 -9.54892159e-01
-1.13621056e+00 -7.05106020e-01 -7.54541516e-01 2.89029688e-01
-3.31835479e-01 -2.52940714e-01 -7.82031178e-01 -2.93369740e-01
-3.50388199e-01 -7.41806626e-03 8.33707511e-01 3.90561312e-01
9.04209137e-01 -6.74960092e-02 -3.25953752e-01 8.92857492e-01
1.24714243e+00 4.76831943e-01 2.79302359e-01 3.71307701e-01
8.24817955e-01 3.13330889e-01 1.00003481e+00 7.15446472e-01
5.18007994e-01 8.08579445e-01 6.98856771e-01 2.29344413e-01
-7.14938790e-02 2.65230332e-02 3.78595233e-01 2.17270285e-01
-1.10814221e-01 -8.36778879e-02 -1.24668217e+00 5.53393066e-01
-1.55809784e+00 -1.01146007e+00 -3.63609165e-01 1.91158950e+00
9.18435901e-02 4.17959452e-01 9.47122872e-02 -3.40699524e-01
7.16981232e-01 3.72238517e-01 -6.09543264e-01 -7.09853321e-03
5.87823391e-02 -1.80173814e-01 1.06851780e+00 4.65745270e-01
-1.51831794e+00 8.20011079e-01 4.93162441e+00 4.37857747e-01
-1.06703269e+00 9.54671949e-02 8.65242705e-02 -3.42739582e-01
1.23362184e-01 -6.91058040e-02 -1.49548101e+00 4.28978443e-01
1.10001040e+00 1.13707393e-01 9.14911628e-02 1.12494171e+00
-6.63334876e-03 1.49862587e-01 -9.20061350e-01 9.94050801e-01
2.57679492e-01 -1.40201783e+00 -6.69681653e-02 3.53172511e-01
2.95175582e-01 4.61365610e-01 1.44981906e-01 6.08240545e-01
3.71873707e-01 -7.45706499e-01 9.37594116e-01 5.17533302e-01
9.36996460e-01 -7.98377216e-01 7.42159367e-01 1.05101049e-01
-1.31661594e+00 -2.38418356e-01 -5.37862301e-01 6.23920634e-02
2.82321334e-01 6.25189185e-01 -1.06044042e+00 6.41521156e-01
8.94262969e-01 6.50826275e-01 -5.37571371e-01 1.02303863e+00
-4.03355546e-02 4.44239408e-01 -4.18273956e-01 1.27829120e-01
4.77363795e-01 -3.68659735e-01 8.37826550e-01 1.25803387e+00
6.04997039e-01 8.26190114e-02 3.14998239e-01 7.65572608e-01
1.10321730e-01 -5.39190114e-01 -7.16692805e-01 1.42496109e-01
7.45914638e-01 1.65047479e+00 -2.64693141e-01 -2.12672159e-01
-4.26307559e-01 1.71116337e-01 -1.38333455e-01 1.95600316e-01
-1.59892190e+00 -6.74927175e-01 1.02299786e+00 1.29598245e-01
5.92699707e-01 -5.86545229e-01 -3.07141125e-01 -9.18509483e-01
-7.88313821e-02 -4.81665075e-01 2.60164291e-01 -9.72054541e-01
-1.41420734e+00 6.86182976e-01 2.63868362e-01 -1.61852884e+00
7.10272267e-02 -1.15105653e+00 -3.41826260e-01 7.56757557e-01
-1.49089015e+00 -1.45440781e+00 -6.87559247e-01 3.38660330e-01
3.66575748e-01 -7.10667610e-01 3.84614289e-01 6.79644704e-01
-6.57125413e-01 4.33996350e-01 1.88137572e-02 7.02760756e-01
8.22294056e-01 -1.01743793e+00 7.14471102e-01 1.02894318e+00
5.87114505e-02 1.97081760e-01 6.98581874e-01 -7.65668154e-01
-1.60265148e+00 -1.61199439e+00 4.86669898e-01 -8.22176933e-01
1.10233450e+00 -7.41443098e-01 -5.64157367e-01 8.89291525e-01
-2.73455739e-01 3.38613540e-01 7.12174356e-01 3.08280122e-02
-7.25596428e-01 -5.11470199e-01 -8.35582554e-01 3.85095835e-01
1.02269864e+00 -3.72052670e-01 -4.86002058e-01 5.82927644e-01
5.37523448e-01 -7.57445514e-01 -8.50093007e-01 5.65989614e-01
7.55706549e-01 -6.26898706e-01 1.08137035e+00 -5.58078468e-01
1.69496372e-01 -7.38071918e-01 -6.89025462e-01 -9.11703944e-01
-5.55497646e-01 -4.97807004e-02 -3.77688259e-01 7.00042307e-01
3.39776754e-01 -6.06044352e-01 6.72240317e-01 -2.82693282e-02
-6.07439995e-01 -6.24372721e-01 -1.14508653e+00 -1.04634798e+00
-2.29468852e-01 -8.20959985e-01 4.78640795e-01 4.86886263e-01
-7.48243749e-01 3.68701726e-01 -5.87764025e-01 8.62934351e-01
9.56370115e-01 5.14672160e-01 1.13510132e+00 -1.56880772e+00
1.27932772e-01 -2.58865088e-01 -6.55808806e-01 -1.10235929e+00
-5.80960549e-02 -8.65070283e-01 2.63250768e-02 -1.34798932e+00
-1.56691760e-01 -5.83557427e-01 -9.96133015e-02 4.03489202e-01
4.83497173e-01 3.41009617e-01 2.85388678e-01 3.84434372e-01
-5.83718061e-01 4.26979661e-01 8.16258192e-01 -1.65640980e-01
3.90091687e-01 -1.98662002e-02 -4.47044522e-01 6.83716357e-01
9.16725278e-01 -5.66773653e-01 4.42738235e-02 -5.79940379e-01
1.79511249e-01 -6.68780059e-02 8.12555611e-01 -1.25938714e+00
2.38629028e-01 -1.07978322e-01 7.82866240e-01 -1.52613986e+00
7.63001382e-01 -8.65977347e-01 6.92275399e-03 4.45226371e-01
1.72043905e-01 2.73072362e-01 2.05044389e-01 8.57378542e-01
-4.48564254e-02 2.16452628e-01 7.10312784e-01 1.54167101e-01
-1.01675475e+00 5.03138661e-01 -3.34482044e-01 -1.07594423e-01
1.15083158e+00 -2.30923310e-01 -9.13442671e-01 -1.99285462e-01
-3.89430106e-01 3.81495178e-01 -2.47222595e-02 7.33490527e-01
7.77338684e-01 -1.64265335e+00 -9.66881335e-01 3.45166922e-01
4.04270977e-01 -1.40019238e-01 5.16864181e-01 8.77864301e-01
-6.48878694e-01 8.23774099e-01 -4.47622567e-01 -7.63235033e-01
-9.99041975e-01 3.57346892e-01 3.73779178e-01 1.22972861e-01
-6.22805178e-01 8.30193639e-01 3.01476181e-01 -5.62683225e-01
6.55209571e-02 -2.33316228e-01 -1.18005872e-01 1.74710035e-01
6.71202302e-01 3.83647978e-01 -2.28585657e-02 -8.40726912e-01
-7.09050536e-01 6.00860357e-01 -2.44042158e-01 -6.74032867e-02
1.22771215e+00 2.49358878e-01 2.14523017e-01 1.83916196e-01
9.28262234e-01 -5.54097630e-02 -1.47942233e+00 -1.51816964e-01
-2.62644198e-02 -6.05757654e-01 2.37421453e-01 -9.66557920e-01
-9.27641332e-01 1.00421715e+00 9.82349992e-01 7.78694674e-02
7.35790431e-01 1.82529595e-02 7.50443935e-01 5.80795228e-01
2.61626601e-01 -9.98239577e-01 4.40522879e-02 9.93610442e-01
9.63292360e-01 -1.29652464e+00 1.67347670e-01 -2.11071149e-01
-7.20984459e-01 1.11029720e+00 7.58595407e-01 -4.07191634e-01
6.99329078e-01 3.72976780e-01 1.85726687e-01 -4.42814678e-01
-7.85186589e-01 -3.88204932e-01 5.52183211e-01 7.42934406e-01
2.46229574e-01 3.46800268e-01 1.16500050e-01 6.34951174e-01
-3.95893872e-01 -5.82675874e-01 4.47502196e-01 6.21026695e-01
-6.91269040e-01 -4.86335099e-01 -6.36760056e-01 4.61748481e-01
-3.95831019e-02 5.82670681e-02 -1.66710973e-01 9.28215206e-01
2.46752411e-01 7.12229192e-01 5.39782405e-01 -7.71605074e-01
5.91080129e-01 -1.41436443e-01 2.10517004e-01 -4.18531120e-01
1.41927004e-01 2.94371899e-02 2.56948173e-01 -4.87571388e-01
1.53306663e-01 -5.36176741e-01 -1.04436147e+00 -2.82261610e-01
-1.11888692e-01 -1.72637761e-01 9.09067035e-01 6.20849550e-01
4.64992374e-01 4.96765018e-01 6.01257384e-01 -8.38267446e-01
-7.18561828e-01 -1.01724923e+00 -4.24332529e-01 -5.40591516e-02
4.46642011e-01 -1.12276936e+00 -2.98761100e-01 -2.39829823e-01] | [7.864893913269043, -1.4574320316314697] |
7953fa25-e4a1-4202-bcdb-eb0d00067c02 | gercct-an-annotated-corpus-for-mining | null | null | https://aclanthology.org/2022.lrec-1.658 | https://aclanthology.org/2022.lrec-1.658.pdf | GerCCT: An Annotated Corpus for Mining Arguments in German Tweets on Climate Change | While the field of argument mining has grown notably in the last decade, research on the Twitter medium remains relatively understudied. Given the difficulty of mining arguments in tweets, recent work on creating annotated resources mainly utilized simplified annotation schemes that focus on single argument components, i.e., on claim or evidence. In this paper we strive to fill this research gap by presenting GerCCT, a new corpus of German tweets on climate change, which was annotated for a set of different argument components and properties. Additionally, we labelled sarcasm and toxic language to facilitate the development of tools for filtering out non-argumentative content. This, to the best of our knowledge, renders our corpus the first tweet resource annotated for argumentation, sarcasm and toxic language. We show that a comparatively complex annotation scheme can still yield promising inter-annotator agreement. We further present first good supervised classification results yielded by a fine-tuned BERT architecture. | ['Manfred Stede', 'Robin Schaefer'] | null | null | null | null | lrec-2022-6 | ['argument-mining'] | ['natural-language-processing'] | [ 2.19109282e-01 8.55421782e-01 -3.22830290e-01 -2.44418934e-01
-8.18508565e-01 -9.87362742e-01 1.05451119e+00 1.01541221e+00
-5.62828898e-01 1.00026274e+00 6.74898326e-01 -7.47577786e-01
1.11362757e-02 -5.82225740e-01 -3.49086165e-01 -3.20728421e-01
1.84195101e-01 5.04909039e-01 1.65360391e-01 -4.27378893e-01
4.64134216e-01 -1.73500590e-02 -1.59890926e+00 5.56526601e-01
8.21668684e-01 5.91964722e-01 -4.23716873e-01 4.27295268e-01
-4.46523517e-01 1.11249173e+00 -7.58956015e-01 -9.68485773e-01
-3.70438188e-01 -4.17412341e-01 -1.41917038e+00 -3.35818976e-01
1.30035385e-01 2.12772265e-01 4.01790082e-01 8.38560343e-01
4.73290026e-01 -4.96133596e-01 3.62572074e-01 -1.07526600e+00
-1.82067513e-01 1.36239636e+00 -2.69411325e-01 1.82629049e-01
5.06044030e-01 -4.53684896e-01 1.42201674e+00 -6.11090541e-01
1.01979506e+00 1.04437900e+00 8.67104173e-01 3.93079728e-01
-9.00476336e-01 -4.26255792e-01 1.71670988e-01 -9.33096930e-02
-6.77705944e-01 -2.38262385e-01 9.09537911e-01 -6.36507154e-01
8.43488812e-01 5.97035527e-01 7.81081080e-01 1.06026661e+00
-3.45464945e-01 7.40002751e-01 1.47886634e+00 -6.34148836e-01
6.78034127e-02 3.46816599e-01 4.27058309e-01 4.21401978e-01
5.53067386e-01 -6.18469775e-01 -3.00885230e-01 -5.98697662e-01
-8.61391723e-02 -5.75372875e-01 2.39579156e-02 1.07284822e-01
-1.21090806e+00 1.16999578e+00 6.06774315e-02 7.10262716e-01
-1.55029282e-01 -1.80605352e-01 1.01715326e+00 2.75203824e-01
1.10231173e+00 4.36942399e-01 -6.84068799e-01 -4.21467215e-01
-7.43977845e-01 2.73074538e-01 1.40807474e+00 4.19945568e-01
3.81829470e-01 -4.94795352e-01 7.05073625e-02 9.67652321e-01
1.53522462e-01 2.72781104e-01 1.44450277e-01 -7.91491985e-01
7.67656863e-01 8.68161559e-01 2.98262239e-01 -1.09704816e+00
-6.81226134e-01 -2.96105772e-01 -4.27295268e-01 1.79047883e-02
6.47714257e-01 -5.35374403e-01 -1.04907282e-01 1.40671003e+00
5.50850332e-01 -5.26430905e-01 2.48199016e-01 6.99036300e-01
1.17774630e+00 2.30077162e-01 4.27803367e-01 -3.10469747e-01
1.70655680e+00 -5.45898199e-01 -9.66288388e-01 1.51881520e-02
1.09431112e+00 -1.16743004e+00 9.76520836e-01 1.15493201e-01
-1.06162548e+00 9.17993560e-02 -7.98801720e-01 3.68615948e-02
-6.87866628e-01 -1.86463129e-02 8.63360345e-01 9.69722986e-01
-4.73324090e-01 3.48346621e-01 -4.25441712e-01 -3.76134247e-01
5.02691686e-01 9.28313471e-03 -2.98780978e-01 6.16667747e-01
-1.49707448e+00 1.04453170e+00 2.86511302e-01 -2.73090154e-01
5.24486564e-02 -4.45445806e-01 -5.65689504e-01 -3.78208548e-01
5.87867677e-01 -3.24929625e-01 1.40132213e+00 -9.81975555e-01
-1.22271311e+00 1.47740316e+00 7.64602199e-02 -7.90269971e-01
6.04261935e-01 -3.24094325e-01 -4.94479477e-01 -8.78788182e-04
4.03757036e-01 2.90494889e-01 4.27029252e-01 -1.03006947e+00
-7.04206824e-01 -1.30966127e-01 3.67119670e-01 -6.92672841e-03
-5.27751684e-01 7.76683509e-01 3.05260688e-01 -6.51397526e-01
-2.21364215e-01 -9.45327818e-01 -3.37031335e-02 -5.32050550e-01
-4.41690505e-01 -7.64802814e-01 8.33763719e-01 -3.93193275e-01
1.41950250e+00 -1.71489704e+00 -2.48123556e-01 -1.50097189e-02
2.81760573e-01 2.22108111e-01 3.22031915e-01 8.71562898e-01
-1.61316052e-01 7.61888146e-01 -2.11364672e-01 -9.39402878e-02
7.49880299e-02 3.35980296e-01 -4.76739943e-01 5.80694795e-01
1.94924369e-01 8.18672299e-01 -1.26410699e+00 -8.41326416e-01
-2.14634523e-01 2.31059387e-01 -3.79678071e-01 -2.81811029e-01
-2.80832291e-01 2.12052181e-01 -7.62704432e-01 5.35907567e-01
3.59342784e-01 -4.01576787e-01 4.94536430e-01 -1.85099930e-01
-7.22962260e-01 9.39313829e-01 -7.88816929e-01 1.26895475e+00
-2.59790599e-01 7.32122362e-01 4.08202767e-01 -1.02555203e+00
7.45211720e-01 4.93121833e-01 6.86778724e-01 -5.48533618e-01
5.10152996e-01 7.23700225e-01 5.65954559e-02 -3.87762398e-01
6.75554693e-01 -2.55176928e-02 -4.60578024e-01 8.65882099e-01
-6.12577021e-01 -2.24298462e-01 4.35092658e-01 2.84769416e-01
8.26524854e-01 2.29136832e-02 4.59069788e-01 -4.41195339e-01
7.11328387e-01 5.82190394e-01 3.91172580e-02 4.87051398e-01
-9.75923464e-02 2.52713203e-01 6.39348269e-01 -6.22938216e-01
-1.12510717e+00 -2.46009544e-01 -4.91899669e-01 1.16146171e+00
-2.41742656e-01 -8.02377045e-01 -9.10502672e-01 -1.04980958e+00
-1.79261208e-01 4.20413345e-01 -5.96132696e-01 7.35911429e-01
-8.23227346e-01 -1.10290265e+00 8.50654185e-01 -4.82868701e-02
3.93442303e-01 -1.09640825e+00 -1.02503431e+00 4.43573713e-01
-6.27923310e-01 -1.06286216e+00 2.95930177e-01 2.36623108e-01
-4.67436522e-01 -1.51281905e+00 -3.81202936e-01 -5.92384398e-01
2.35613421e-01 -4.20102105e-03 1.43546200e+00 4.42556351e-01
1.60156805e-02 7.51844719e-02 -7.71060109e-01 -9.19446528e-01
-6.20179594e-01 4.25644398e-01 -5.19331455e-01 -4.10149634e-01
3.18982333e-01 -3.09542000e-01 -3.76310468e-01 1.25095814e-01
-7.58875310e-01 4.42639440e-02 1.81946024e-01 7.53241777e-01
8.24899599e-02 -3.02937984e-01 6.71651900e-01 -1.41605830e+00
1.11764526e+00 -7.35445559e-01 -2.23640963e-01 -1.62094563e-01
-7.16209352e-01 -1.14438437e-01 2.53987998e-01 -1.75456777e-01
-9.73606467e-01 -3.92504573e-01 -5.24708807e-01 8.41919363e-01
-1.29535988e-01 9.19984996e-01 5.09129107e-01 3.01520497e-01
8.57830465e-01 -6.01376832e-01 5.38776331e-02 -3.77822101e-01
4.71032768e-01 9.16864157e-01 1.96262911e-01 -7.38551080e-01
5.70792437e-01 7.08080649e-01 -9.77016911e-02 -6.83044434e-01
-1.46823156e+00 -5.80473065e-01 -3.42261732e-01 -2.62310803e-01
6.77370250e-01 -8.18249762e-01 -7.06798553e-01 -1.29958633e-02
-1.33195579e+00 -1.77557200e-01 -4.29425925e-01 1.94464788e-01
-1.40394419e-01 4.49312121e-01 -6.63447022e-01 -9.40802693e-01
-5.56378901e-01 -5.59696257e-01 8.30398619e-01 -2.63349742e-01
-9.27354932e-01 -1.17201781e+00 4.12096620e-01 5.71601748e-01
2.90205210e-01 6.46749735e-01 5.91512918e-01 -9.99044776e-01
2.17282087e-01 -2.92738289e-01 -3.38447332e-01 -6.70393407e-02
-1.22418456e-01 2.05355249e-02 -1.01796663e+00 9.32728574e-02
-1.34783566e-01 -6.52608931e-01 6.88136935e-01 -9.76608098e-02
6.92001581e-01 -5.93188882e-01 -4.02040035e-01 -3.17465961e-01
9.34690952e-01 -2.61750311e-01 5.45959532e-01 1.14965785e+00
3.24873835e-01 1.00365305e+00 8.59268248e-01 5.45480251e-01
4.83650535e-01 5.20668089e-01 5.87368429e-01 -2.31202379e-01
-3.98752987e-02 8.21859688e-02 9.04483870e-02 7.31605351e-01
-3.72140139e-01 -3.59072566e-01 -1.15747535e+00 7.45756269e-01
-1.88073421e+00 -1.20720601e+00 -9.37429547e-01 1.45529222e+00
1.00183499e+00 3.85841727e-01 2.58168876e-01 5.94326615e-01
5.41596115e-01 2.98808396e-01 2.97557712e-01 -6.00256443e-01
-5.59115708e-01 2.78314352e-01 2.80643880e-01 5.12199461e-01
-1.25370812e+00 7.61850536e-01 6.12160635e+00 6.65395021e-01
-7.80391932e-01 3.52965087e-01 5.19759417e-01 3.58122885e-01
-6.30449057e-01 2.37934321e-01 -7.42046714e-01 3.59993309e-01
9.84164059e-01 1.18765563e-01 -2.78821677e-01 7.61427760e-01
2.79641092e-01 -5.86777888e-02 -5.61964631e-01 3.76723140e-01
-1.16013683e-01 -1.61494887e+00 -3.28974694e-01 1.05626635e-01
6.41226172e-01 3.33395332e-01 -2.88164943e-01 1.58074275e-01
4.56070662e-01 -7.98659086e-01 1.12841487e+00 -5.11601344e-02
3.98071289e-01 -3.70115042e-01 1.08747852e+00 2.81378061e-01
-8.32723081e-01 -9.42925140e-02 8.74219611e-02 -6.01134419e-01
3.26220304e-01 8.24240685e-01 -7.87750542e-01 5.51549017e-01
5.62859535e-01 6.84261084e-01 -3.97559583e-01 5.93163967e-01
-3.84677947e-01 1.07599044e+00 -3.75497311e-01 -5.02092719e-01
5.63528359e-01 -3.51575352e-02 6.05867505e-01 1.62667847e+00
7.74833262e-02 1.07252710e-01 2.45343894e-01 3.23126137e-01
-1.59504056e-01 6.46175325e-01 -4.20275450e-01 -1.62637949e-01
4.92996693e-01 1.61160326e+00 -1.07776880e+00 -4.10228908e-01
-4.59107220e-01 3.48061770e-01 4.83060241e-01 -2.37615734e-01
-7.93952107e-01 -1.41649032e-02 2.14619353e-01 3.22934031e-01
1.09501090e-02 -1.09635266e-02 -4.00072545e-01 -1.02088952e+00
-3.37126926e-02 -8.71266007e-01 8.12520087e-01 -3.61150801e-01
-1.20150542e+00 7.21254408e-01 6.34521171e-02 -1.02755868e+00
-3.02216202e-01 -3.69123876e-01 -4.55640167e-01 6.31262302e-01
-1.70560253e+00 -1.36213326e+00 -1.01763874e-01 8.95931497e-02
3.03120285e-01 2.20977053e-01 9.83035445e-01 3.38409275e-01
-2.36284927e-01 6.93465769e-02 -2.97442168e-01 2.27647290e-01
7.75003910e-01 -1.05367196e+00 2.29067042e-01 1.64308816e-01
7.42591769e-02 5.62392652e-01 9.72776353e-01 -6.62311494e-01
-7.20330298e-01 -6.08835876e-01 1.60740638e+00 -7.64355361e-01
1.34243083e+00 -1.32675171e-01 -6.70118690e-01 2.54947692e-01
7.04581618e-01 -3.60110641e-01 9.20318186e-01 5.32606840e-01
-6.11333430e-01 5.15174985e-01 -8.97709131e-01 4.85541165e-01
8.05575669e-01 -5.78090847e-01 -8.77021611e-01 5.23728967e-01
3.94101083e-01 -3.99450630e-01 -7.93197393e-01 2.59398967e-01
5.80413699e-01 -8.54009151e-01 7.71503747e-01 -5.84005535e-01
5.33163905e-01 -1.69920102e-01 8.97562876e-02 -9.11889434e-01
2.49532714e-01 -6.60701513e-01 1.96054146e-01 1.54002964e+00
7.52752244e-01 -5.19013524e-01 8.14201534e-01 2.87168056e-01
-3.68324399e-01 -5.44223487e-01 -7.92650521e-01 -2.50253856e-01
3.20126146e-01 -6.27388597e-01 4.60239470e-01 1.36144733e+00
5.54980338e-01 7.80157506e-01 -1.56373456e-01 -4.89379227e-01
2.61457264e-01 4.62737560e-01 6.40728354e-01 -1.88003588e+00
2.70276636e-01 -6.65988386e-01 1.96467280e-01 -4.38382268e-01
4.01117086e-01 -9.26651478e-01 -2.27747262e-01 -1.52364028e+00
2.27335051e-01 -6.82931304e-01 2.33089715e-01 5.28913319e-01
-5.69507703e-02 5.51504254e-01 -1.66332513e-01 3.92395556e-01
-6.44325078e-01 -1.05733328e-01 9.18907523e-01 -5.08329384e-02
-4.29457389e-02 -1.62752166e-01 -9.05149281e-01 1.08781242e+00
1.19080985e+00 -6.46408260e-01 -1.15700856e-01 -1.75539538e-01
1.14077902e+00 -4.42457497e-01 2.81360030e-01 -3.74411613e-01
-2.18548141e-02 -8.54144320e-02 -2.60738105e-01 -7.10218787e-01
-4.55205590e-02 -6.35384202e-01 4.45670336e-02 4.54074800e-01
-4.95583504e-01 3.80076915e-02 1.38067856e-01 3.67967874e-01
-3.19838673e-01 -4.06945735e-01 3.49034309e-01 -2.28192851e-01
-1.46952227e-01 -3.72089773e-01 -5.77015281e-01 5.20169973e-01
5.88015437e-01 -5.14679365e-02 -7.53327131e-01 -2.60933816e-01
-6.51005566e-01 -3.27472799e-02 3.19328398e-01 1.93120286e-01
-3.03584207e-02 -9.80488598e-01 -1.15797341e+00 -6.60747528e-01
1.39448687e-01 -3.75623107e-01 -2.92983443e-01 1.08255446e+00
-3.52906674e-01 4.41031516e-01 1.98838457e-01 -2.44959950e-01
-1.63003588e+00 2.08845183e-01 -2.17434675e-01 -6.66134596e-01
-5.57607114e-01 4.34956491e-01 -5.30539930e-01 -4.11132246e-01
-1.90544173e-01 -1.26242593e-01 -8.65502715e-01 8.61072361e-01
3.82854879e-01 3.60377401e-01 1.97614521e-01 -9.16759491e-01
-3.22339743e-01 2.20961839e-01 1.07474230e-01 -2.41006643e-01
1.48339212e+00 -2.64987379e-01 -5.64157069e-01 4.89425004e-01
8.34659338e-01 7.39291906e-01 -4.42062736e-01 1.64686531e-01
5.82279444e-01 -5.62126115e-02 -1.41408101e-01 -8.46711278e-01
-3.38147998e-01 3.58605921e-01 -8.08873475e-02 1.17621160e+00
5.18386543e-01 3.37005645e-01 5.40284216e-01 3.10360253e-01
7.36187547e-02 -1.32310104e+00 -1.16477974e-01 8.60804796e-01
8.57432187e-01 -1.16315162e+00 2.66977102e-01 -8.48369300e-01
-3.57201397e-01 1.15487909e+00 -2.73136608e-02 2.94295698e-01
5.90667069e-01 3.85965735e-01 2.62256652e-01 -8.60781729e-01
-5.77246368e-01 -3.92111361e-01 9.45436358e-02 2.31969446e-01
1.24741483e+00 6.19733483e-02 -1.25888574e+00 4.25827920e-01
-5.39337218e-01 -2.93822646e-01 5.85285962e-01 9.68072295e-01
-5.81015646e-01 -1.47773731e+00 -3.85142416e-01 2.53491968e-01
-1.24827111e+00 -2.58815080e-01 -9.94355083e-01 1.23107195e+00
1.91921875e-01 1.16273856e+00 -4.16652448e-02 1.49337500e-01
2.23209485e-01 1.06690697e-01 1.00429356e-01 -5.05861819e-01
-1.20722020e+00 4.76362929e-02 1.40517628e+00 -1.05525807e-01
-1.36274779e+00 -9.24002528e-01 -1.24637902e+00 -3.12031269e-01
-4.32257593e-01 9.48946536e-01 7.78733671e-01 1.05463719e+00
1.89674124e-02 2.31358409e-01 1.88483551e-01 -5.19260645e-01
-1.84433326e-01 -1.00461626e+00 -7.99260139e-02 7.01456785e-01
1.67712182e-01 -4.82538581e-01 -4.34152216e-01 1.28985152e-01] | [9.372610092163086, 9.689542770385742] |
984c8247-911b-4eec-9aab-58723d9e911b | prompting-and-evaluating-large-language | 2305.13626 | null | https://arxiv.org/abs/2305.13626v1 | https://arxiv.org/pdf/2305.13626v1.pdf | Prompting and Evaluating Large Language Models for Proactive Dialogues: Clarification, Target-guided, and Non-collaboration | Conversational systems based on Large Language Models (LLMs), such as ChatGPT, show exceptional proficiency in context understanding and response generation. However, despite their impressive capabilities, they still possess limitations, such as providing randomly-guessed answers to ambiguous queries or failing to refuse users' requests, both of which are considered aspects of a conversational agent's proactivity. This raises the question of whether LLM-based conversational systems are equipped to handle proactive dialogue problems. In this work, we conduct a comprehensive analysis of LLM-based conversational systems, specifically focusing on three aspects of proactive dialogue systems: clarification, target-guided, and non-collaborative dialogues. To trigger the proactivity of LLMs, we propose the Proactive Chain-of-Thought prompting scheme, which augments LLMs with the goal planning capability over descriptive reasoning chains. Empirical findings are discussed to promote future studies on LLM-based proactive dialogue systems. | ['Tat-Seng Chua', 'Lizi Liao', 'Wenqiang Lei', 'Yang Deng'] | 2023-05-23 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 7.37831220e-02 9.35199022e-01 -3.11546717e-02 -4.79567856e-01
-5.33158660e-01 -7.09219813e-01 1.20261145e+00 -6.38001487e-02
-6.91199377e-02 8.98447335e-01 7.44642615e-01 -7.14199781e-01
-5.44641390e-02 -8.16296935e-01 3.70836049e-01 -2.52318591e-01
1.65591121e-01 7.46270180e-01 2.32848659e-01 -9.43564296e-01
8.07455897e-01 1.98061615e-01 -1.10265398e+00 8.36016297e-01
9.68263149e-01 3.12603980e-01 2.53645390e-01 9.70359623e-01
-8.68504822e-01 1.75175154e+00 -7.98128545e-01 -2.34174088e-01
-3.92078251e-01 -6.86343849e-01 -1.68048418e+00 1.24325372e-01
-2.89873838e-01 -5.25126636e-01 -9.04470775e-03 5.41387975e-01
3.76173675e-01 3.26663256e-01 4.28785324e-01 -1.45810652e+00
-3.61269116e-01 1.03930235e+00 4.21406269e-01 -5.93095012e-02
1.23482537e+00 5.60813367e-01 1.04279137e+00 -5.81788659e-01
5.14820933e-01 1.69588077e+00 4.30172354e-01 1.13349676e+00
-9.00637269e-01 -3.58029492e-02 2.39230260e-01 -3.64990421e-02
-7.41108000e-01 -6.19714439e-01 5.54828167e-01 -4.49658930e-01
1.42299044e+00 7.22751856e-01 3.56809199e-01 1.02524924e+00
1.08522981e-01 7.14634478e-01 1.28413486e+00 -6.29809082e-01
2.94859469e-01 4.20085430e-01 4.22438920e-01 2.39915222e-01
-4.07215863e-01 -2.24460647e-01 -7.33633399e-01 -6.40310228e-01
6.55904353e-01 -4.79068041e-01 -6.39845207e-02 4.02117759e-01
-1.08871877e+00 1.06377721e+00 -1.96713343e-01 6.67790473e-01
-6.09304190e-01 -4.52138782e-01 3.30892265e-01 6.59991205e-01
3.03766370e-01 9.95284796e-01 -2.88393050e-01 -7.14612842e-01
-1.40352875e-01 4.32555735e-01 1.88782763e+00 1.04334402e+00
4.30885583e-01 -9.06758532e-02 -6.24307454e-01 1.05984688e+00
5.50968766e-01 1.20850310e-01 3.94122690e-01 -1.20661116e+00
4.86167610e-01 1.14776349e+00 4.45991993e-01 -6.51506126e-01
-5.74718058e-01 3.37332368e-01 -3.64613444e-01 -2.83367962e-01
3.56630236e-01 -4.59649891e-01 9.90246087e-02 1.49064076e+00
3.03773463e-01 -4.80713874e-01 6.95513964e-01 7.56103694e-01
1.08752441e+00 6.66082799e-01 3.44776928e-01 -5.96192181e-01
1.40657723e+00 -1.15902531e+00 -8.58104467e-01 -2.72201061e-01
9.06867683e-01 -7.85507798e-01 1.23717391e+00 1.28675640e-01
-1.17710698e+00 -2.48613626e-01 -3.42984796e-01 9.46029276e-02
-1.42838627e-01 -2.57959783e-01 8.89479637e-01 5.51153779e-01
-1.18594289e+00 3.12866569e-02 -9.59969610e-02 -6.22831583e-01
-5.98071575e-01 1.64830059e-01 6.91788420e-02 2.95608371e-01
-1.59652209e+00 1.02422965e+00 2.43455678e-01 1.53028861e-01
-6.03887737e-01 -1.57522365e-01 -5.71060658e-01 -4.61667329e-02
5.15775025e-01 -5.80081999e-01 2.00468206e+00 -5.78891337e-01
-2.15635347e+00 6.82327569e-01 6.91760704e-02 -3.65407974e-01
5.91307163e-01 -8.14431310e-02 -3.39882821e-01 1.24807633e-01
8.48804116e-02 4.36136544e-01 8.99339989e-02 -1.13406253e+00
-8.28875899e-01 4.28084433e-02 7.59088039e-01 6.53396010e-01
1.03091881e-01 4.53167170e-01 1.56967004e-03 1.23615131e-01
-2.04694867e-01 -8.33311498e-01 -4.97535616e-01 -7.08384752e-01
-4.88307714e-01 -8.95768166e-01 4.39737558e-01 -2.82782793e-01
1.53440857e+00 -1.54335308e+00 -2.30113268e-01 -4.06933613e-02
2.11312816e-01 4.89749193e-01 -2.59908050e-01 1.55746782e+00
5.28346539e-01 2.42768183e-01 2.05490425e-01 -1.27922013e-01
3.13550353e-01 3.38887542e-01 -3.88303041e-01 -4.24822062e-01
1.83542911e-02 7.63391614e-01 -9.63156879e-01 -6.34325504e-01
2.29694560e-01 -2.01936290e-01 -5.24906218e-01 1.08970726e+00
-8.20933223e-01 6.79657578e-01 -7.31251776e-01 3.20023328e-01
2.06949666e-01 -3.38339061e-01 6.07194185e-01 6.05167806e-01
-4.57534879e-01 1.03003120e+00 -6.70320570e-01 1.12436318e+00
-6.99584663e-01 2.82359213e-01 3.87021750e-01 -3.85864049e-01
8.82463098e-01 8.04873705e-01 1.72180325e-01 -5.51576197e-01
-1.20745599e-01 5.85237965e-02 4.10803497e-01 -8.73756409e-01
8.08678925e-01 4.57556136e-02 -2.27008641e-01 1.15689838e+00
-3.15595657e-01 -2.32863560e-01 2.54958540e-01 5.91640234e-01
9.03180957e-01 -2.20670179e-01 5.50389051e-01 -6.41772151e-02
1.07027161e+00 2.74065077e-01 1.72859043e-01 1.17794383e+00
-2.15557411e-01 -1.43109441e-01 8.48088920e-01 -2.92669535e-01
-4.76316810e-01 -2.96859592e-01 4.51077700e-01 1.65424693e+00
-4.82191518e-03 -5.72515428e-01 -1.00882351e+00 -6.36425972e-01
-4.44964379e-01 1.12946117e+00 -1.40237454e-02 1.61524251e-01
-7.65885830e-01 -5.04014373e-01 7.21590757e-01 1.32143363e-01
7.89250910e-01 -1.61079550e+00 -6.62912667e-01 5.37976503e-01
-7.43192971e-01 -1.09999633e+00 -2.67304182e-01 -2.27589741e-01
-5.81933081e-01 -1.19238389e+00 -8.42769071e-03 -5.85741937e-01
3.35692972e-01 4.40231711e-01 1.12630630e+00 6.88525915e-01
4.25289363e-01 8.07701826e-01 -6.71817243e-01 -7.35584348e-02
-1.28328609e+00 1.74751967e-01 -1.58029854e-01 -2.76023209e-01
4.10423905e-01 -3.85826856e-01 -5.07594407e-01 7.95992613e-01
-5.18843830e-01 4.43653584e-01 3.20399463e-01 7.71337926e-01
-4.08365607e-01 -5.53279579e-01 1.05000424e+00 -1.05060279e+00
1.87564087e+00 -5.42796254e-01 -7.32187629e-02 4.53082114e-01
-7.52171338e-01 -1.11138590e-01 6.41093671e-01 -4.45431530e-01
-1.64542735e+00 -5.86993694e-01 -5.20064354e-01 7.92973101e-01
-3.91211689e-01 4.90850359e-01 -3.90106440e-03 -1.84031632e-02
7.77052760e-01 4.31242347e-01 2.83379078e-01 -2.00182736e-01
4.45446253e-01 1.12293899e+00 1.94675401e-02 -1.06666768e+00
1.28949538e-01 -3.53094578e-01 -7.49201179e-01 -9.46488023e-01
-4.26151514e-01 -7.05912828e-01 -2.11660430e-01 -6.24525964e-01
5.63988864e-01 -4.57589954e-01 -1.49882627e+00 2.45247737e-01
-1.42834878e+00 -7.90432513e-01 6.03805073e-02 8.36396739e-02
-8.55947077e-01 5.93492031e-01 -1.09481692e+00 -1.62970161e+00
-7.63732910e-01 -1.02293336e+00 5.77823341e-01 3.98521334e-01
-9.37762201e-01 -1.11365676e+00 4.06001918e-02 8.15763056e-01
9.46955621e-01 -4.65595096e-01 1.23631525e+00 -1.51033270e+00
-3.66134554e-01 -5.18887024e-03 7.93619677e-02 2.90325303e-02
8.84673819e-02 -2.69914985e-01 -8.25310230e-01 1.46545708e-01
2.28208661e-01 -6.22342348e-01 -2.27992624e-01 -2.46667817e-01
3.08513165e-01 -1.09493577e+00 -1.16613857e-01 -4.73204702e-01
6.01117373e-01 7.87423432e-01 6.04801118e-01 1.73068643e-01
-3.23243439e-02 1.34176886e+00 8.78216505e-01 5.56035817e-01
7.98696637e-01 5.99690676e-01 2.57625878e-01 3.79641443e-01
1.13278963e-01 -4.89082485e-01 3.01753014e-01 8.62878144e-01
-2.02330291e-01 -4.24152374e-01 -1.01974344e+00 1.62704945e-01
-2.10900974e+00 -1.05922532e+00 -3.65752131e-01 1.66755080e+00
1.15491247e+00 -6.67228922e-03 9.61412862e-02 -1.46470740e-01
5.38893759e-01 3.81015867e-01 -2.00246453e-01 -8.99362922e-01
2.23785341e-01 -4.07243013e-01 -3.86657327e-01 1.02224517e+00
-3.61969352e-01 1.16340590e+00 6.73990917e+00 3.51832092e-01
-7.72935569e-01 1.15009949e-01 4.07378227e-01 5.81356049e-01
-4.07404363e-01 3.65283996e-01 -7.97875464e-01 2.26392254e-01
1.11323333e+00 -2.34422922e-01 5.60793102e-01 8.98106039e-01
5.91230989e-01 -3.80139887e-01 -1.14485013e+00 4.68650967e-01
-1.40251368e-01 -1.33081019e+00 9.82547849e-02 5.50494194e-02
1.72434047e-01 -4.76541936e-01 -6.30312085e-01 8.41185570e-01
6.45272434e-01 -8.83212626e-01 2.20742613e-01 5.20451903e-01
3.73875909e-02 -2.14193493e-01 7.64252782e-01 1.17565000e+00
-7.15985596e-01 -1.96282834e-01 -6.77671805e-02 -4.99882996e-01
3.56309682e-01 -2.09885031e-01 -1.48848867e+00 2.77971655e-01
-6.21897802e-02 -2.33626664e-01 -1.57887280e-01 3.92735392e-01
-3.12628180e-01 7.05612659e-01 -1.61393639e-02 -6.85506999e-01
2.74451554e-01 -3.39717329e-01 5.46450138e-01 1.23235762e+00
-2.71453440e-01 7.24617004e-01 5.74998498e-01 7.85804749e-01
4.20713633e-01 3.56694460e-01 -4.40977246e-01 -2.00688571e-01
9.24104810e-01 1.10819304e+00 -2.34066308e-01 -5.19835949e-01
-2.23846763e-01 6.02565944e-01 1.35430157e-01 2.37280920e-01
-2.59948999e-01 -2.14524511e-02 4.15492356e-01 5.38764670e-02
-7.78901398e-01 -1.20357417e-01 -1.17790453e-01 -1.01698506e+00
-3.68450403e-01 -1.48305309e+00 2.51615614e-01 -7.55026579e-01
-1.24557078e+00 7.70667315e-01 1.07747525e-01 -6.54753387e-01
-9.90987599e-01 -1.96212500e-01 -1.13658738e+00 8.75823736e-01
-1.04713249e+00 -1.24335361e+00 -1.58867076e-01 5.22148430e-01
1.13335276e+00 -1.95119038e-01 1.27171004e+00 -1.46998122e-01
-3.18974406e-01 1.53641939e-01 -7.56425381e-01 -2.83057671e-02
5.88425279e-01 -1.12563133e+00 2.88307279e-01 2.57370830e-01
-5.10048091e-01 1.05732679e+00 8.63110006e-01 -5.88648617e-01
-1.50445807e+00 -4.55787957e-01 1.45996451e+00 -3.53909165e-01
5.97869039e-01 -2.42159396e-01 -1.07598639e+00 5.10569572e-01
6.82541728e-01 -9.68488872e-01 8.65105391e-01 3.49413037e-01
1.74804032e-01 4.29961622e-01 -1.13067102e+00 9.37370718e-01
9.23671961e-01 -6.37691677e-01 -9.90963519e-01 8.54058623e-01
7.67648041e-01 -4.37912911e-01 -8.58542800e-01 -7.90421441e-02
2.82562345e-01 -1.13908505e+00 6.87728405e-01 -8.35983217e-01
2.25686520e-01 1.34932533e-01 -5.35293147e-02 -1.01792061e+00
4.97899503e-02 -1.33759725e+00 -1.73607282e-03 1.47729492e+00
3.78665298e-01 -1.03428566e+00 4.28308755e-01 1.38087714e+00
-1.69530511e-01 -5.78024089e-01 -5.37992001e-01 -2.76050866e-01
2.82247011e-02 -4.54960346e-01 6.32225156e-01 8.40869963e-01
1.01266336e+00 8.52985322e-01 -5.09585023e-01 -1.25754535e-01
-4.91630398e-02 7.91267976e-02 1.16324627e+00 -1.18696690e+00
-1.67255029e-01 -4.50188309e-01 5.06667674e-01 -1.43888700e+00
2.28069067e-01 -3.29595685e-01 3.38460147e-01 -1.80893600e+00
-2.38603354e-01 -4.87921029e-01 6.90239668e-01 5.03655910e-01
2.24232133e-02 -7.55557477e-01 1.58673733e-01 5.14449477e-01
-7.39635408e-01 3.34885508e-01 1.17244816e+00 2.24228993e-01
-6.46125019e-01 5.45710623e-01 -7.96778679e-01 6.76548064e-01
9.35997486e-01 -1.36931941e-01 -7.36577690e-01 2.28680357e-01
2.33823895e-01 1.10080314e+00 -8.44721869e-03 -3.66860926e-01
6.69779241e-01 -8.92844677e-01 -7.52971232e-01 -3.82669955e-01
2.81319708e-01 -2.99667746e-01 -1.23502478e-01 3.83309603e-01
-1.13587093e+00 5.69661967e-02 -2.57465005e-01 2.06158653e-01
-2.96091288e-01 -4.79299635e-01 2.76419789e-01 -4.77257162e-01
-6.63806498e-01 -3.07682216e-01 -1.54046345e+00 3.54110226e-02
8.83570433e-01 -1.14662357e-01 -7.82426775e-01 -9.77540910e-01
-5.57288527e-01 6.28278255e-01 1.03677109e-01 4.77062017e-01
5.38470805e-01 -7.06822872e-01 -5.21222889e-01 -7.75494874e-02
2.03493893e-01 -2.33330593e-01 2.04670161e-01 8.02243769e-01
-3.24630082e-01 1.02709842e+00 -3.34341603e-04 -2.85057396e-01
-1.24130189e+00 2.42536351e-01 4.35949981e-01 -5.56951106e-01
-4.77102488e-01 4.68847632e-01 8.41659978e-02 -9.12985265e-01
3.87308478e-01 -1.08370081e-01 -7.38298535e-01 -4.70141880e-02
6.28103793e-01 2.10844159e-01 -3.41108948e-01 -3.00328165e-01
-2.10509360e-01 -2.73484170e-01 -5.28001413e-02 -5.02689958e-01
6.48828924e-01 -5.90949655e-01 -5.36641240e-01 3.82455349e-01
2.48365790e-01 1.40761416e-02 -6.39424682e-01 -3.60857606e-01
5.00871956e-01 -2.38251254e-01 -7.06099033e-01 -1.26000726e+00
2.87949592e-02 5.58992267e-01 -2.06984758e-01 1.05072618e+00
4.81291354e-01 -1.20997332e-01 5.52179754e-01 8.25996399e-01
6.26539826e-01 -1.19240284e+00 3.87886941e-01 1.09280157e+00
1.20402515e+00 -1.20492470e+00 -4.74673241e-01 -4.08417225e-01
-1.17890894e+00 1.27384293e+00 1.17555833e+00 5.49357653e-01
1.10508867e-01 1.21890441e-01 4.84276593e-01 -3.04020822e-01
-1.70270491e+00 -1.43575743e-02 -3.12496245e-01 5.36505163e-01
7.86302686e-01 8.22642744e-02 -8.86439264e-01 5.78563750e-01
-2.65209913e-01 -1.20506696e-01 7.63808250e-01 1.10276794e+00
-7.79437780e-01 -1.19217622e+00 -4.46687818e-01 6.51791990e-02
-1.90334871e-01 -1.43378779e-01 -1.03842080e+00 5.98105133e-01
-4.96555328e-01 1.97464907e+00 -3.52576673e-01 -3.13251704e-01
4.36811477e-01 5.11716008e-01 -1.96529001e-01 -8.43161464e-01
-1.34245694e+00 -2.64198095e-01 1.12916887e+00 -3.44942451e-01
-4.96868044e-01 -2.42591307e-01 -1.29932404e+00 -5.45139968e-01
-3.52844477e-01 8.72699201e-01 3.38737488e-01 1.11057842e+00
2.31826738e-01 2.77388431e-02 6.76389396e-01 -3.62600952e-01
-1.00530815e+00 -1.44804907e+00 -1.30695542e-02 1.50234401e-01
-6.51582927e-02 -1.11921027e-01 -3.06031168e-01 -3.12560827e-01] | [12.794414520263672, 8.032617568969727] |
b9e289ed-fcd2-4e4d-9c2d-3766d2494745 | learning-for-open-world-calibration-with | 2305.12039 | null | https://arxiv.org/abs/2305.12039v1 | https://arxiv.org/pdf/2305.12039v1.pdf | Learning for Open-World Calibration with Graph Neural Networks | We tackle the problem of threshold calibration for open-world recognition by incorporating representation compactness measures into clustering. Unlike the open-set recognition which focuses on discovering and rejecting the unknown, open-world recognition learns robust representations that are generalizable to disjoint unknown classes at test time. Our proposed method is based on two key observations: (i) representation structures among neighbouring images in high dimensional visual embedding spaces have strong self-similarity which can be leveraged to encourage transferability to the open world, (ii) intra-class embedding structures can be modeled with the marginalized von Mises-Fisher (vMF) probability, whose correlation with the true positive rate is dataset-invariant. Motivated by these, we design a unified framework centered around a graph neural network (GNN) to jointly predict the pseudo-labels and the vMF concentrations which indicate the representation compactness. These predictions can be converted into statistical estimations for recognition accuracy, allowing more robust calibration of the distance threshold to achieve target utility for the open-world classes. Results on a variety of visual recognition benchmarks demonstrate the superiority of our method over traditional posthoc calibration methods for the open world, especially under distribution shift. | ['Yifan Xing', 'Joseph Tighe', 'Ying Nian Wu', 'Qingming Tang', 'Tong He', 'Tianjun Xiao', 'Dongsheng An', 'Qin Zhang'] | 2023-05-19 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [ 9.40925628e-02 1.40407115e-01 -4.22189087e-01 -4.48664635e-01
-7.79242277e-01 -6.73429906e-01 6.47412896e-01 1.17002062e-01
8.78468379e-02 4.60814536e-01 2.23355159e-01 -1.55162647e-01
-5.66209197e-01 -7.65344262e-01 -6.04595184e-01 -9.80116606e-01
-4.01854783e-01 4.94606465e-01 -2.83923149e-02 3.27522397e-01
1.96739778e-01 4.73701924e-01 -1.61420166e+00 6.86152978e-03
1.23624218e+00 1.22140241e+00 -2.56140321e-01 4.09636170e-01
-8.69180411e-02 4.88652110e-01 -3.12911749e-01 -5.04053056e-01
6.83509588e-01 -1.30698770e-01 -6.82142377e-01 1.41134843e-01
5.63355029e-01 2.99830794e-01 -5.04967153e-01 1.41520691e+00
4.36151296e-01 1.73420683e-01 1.37513018e+00 -1.56096327e+00
-1.24957800e+00 5.29470801e-01 -6.16557598e-01 2.93452173e-01
7.66168833e-02 2.04601675e-01 1.30748403e+00 -7.10793972e-01
5.58948457e-01 1.22972894e+00 5.07067323e-01 7.30287656e-02
-1.45271385e+00 -8.08956861e-01 1.90971226e-01 3.08866978e-01
-1.72723722e+00 -4.73635159e-02 6.75097883e-01 -6.83848381e-01
4.22861069e-01 3.28416675e-01 5.00059307e-01 1.10660255e+00
2.79930562e-01 4.30429935e-01 1.22940159e+00 -2.62477398e-01
4.11119401e-01 2.31989324e-01 4.93693411e-01 6.15334988e-01
6.17170691e-01 1.89500302e-01 -2.17954591e-01 -2.02090487e-01
4.79686230e-01 2.48125523e-01 -4.20591086e-01 -1.01987541e+00
-1.31204093e+00 1.10217500e+00 9.73144770e-01 9.60625112e-02
-6.41678274e-02 -2.22350538e-01 3.72607678e-01 4.69573915e-01
5.61607957e-01 4.96483713e-01 -1.66972518e-01 2.88088262e-01
-5.05278826e-01 -5.31994879e-01 7.91613817e-01 7.33597755e-01
9.15936530e-01 -1.95492342e-01 -3.85145307e-01 9.93466973e-01
5.68157494e-01 4.37581658e-01 7.60096192e-01 -4.22140241e-01
4.72389996e-01 9.55712855e-01 -4.54556108e-01 -1.18870676e+00
-1.43270746e-01 -7.73010373e-01 -1.05966282e+00 -7.17819482e-02
1.72449917e-01 2.06103519e-01 -1.08067000e+00 1.57166564e+00
4.01143640e-01 5.34655631e-01 1.93184033e-01 7.20184803e-01
7.23414361e-01 6.50523365e-01 4.76987921e-02 -2.46641606e-01
1.13258123e+00 -7.08577156e-01 -5.38977265e-01 -1.43081276e-02
8.17729354e-01 -2.85371423e-01 9.20073748e-01 1.23634182e-01
-8.97565261e-02 -4.46385443e-01 -1.31664872e+00 1.66605383e-01
-6.88860238e-01 -2.07233597e-02 6.45014882e-01 8.27465117e-01
-7.47276545e-01 4.02155340e-01 -3.83673191e-01 -3.58025104e-01
7.19622672e-01 4.30105925e-01 -5.55101931e-01 -2.38011345e-01
-1.08091688e+00 6.65780783e-01 5.46263814e-01 1.30511805e-01
-7.49181926e-01 -6.30299985e-01 -8.07561278e-01 1.04221702e-01
2.19332486e-01 -5.40379941e-01 1.40269876e-01 -8.31448555e-01
-1.03466356e+00 1.02077138e+00 3.42760831e-01 -3.76837164e-01
2.69317776e-01 3.22602808e-01 -5.29035628e-01 6.11279532e-02
1.78913772e-01 5.13474882e-01 8.21934998e-01 -1.39495528e+00
-1.53230801e-01 -7.26363420e-01 -2.43578702e-01 2.73920447e-01
-7.57879972e-01 -4.44970071e-01 -2.65152127e-01 -6.14680767e-01
4.76062447e-01 -9.43432510e-01 -2.17962563e-01 2.58733947e-02
-6.48875177e-01 -3.80013406e-01 6.30768239e-01 -2.54136145e-01
1.20880985e+00 -2.32006478e+00 2.58715361e-01 8.56773615e-01
6.14220917e-01 2.59237662e-02 -3.86105955e-01 1.28577441e-01
-3.46701801e-01 1.28149137e-01 -9.55795348e-02 1.90599009e-01
1.76233515e-01 1.48533821e-01 -4.95407015e-01 9.45930481e-01
5.97173758e-02 9.33494747e-01 -7.30097115e-01 -4.40865278e-01
1.81329355e-01 2.60575742e-01 -1.54610887e-01 2.78121293e-01
2.60895431e-01 4.76332707e-03 -2.27844015e-01 5.90411663e-01
8.00994217e-01 -4.88546014e-01 7.78403506e-02 -1.40315324e-01
5.07860720e-01 -1.52130514e-01 -1.38754570e+00 1.23419619e+00
-1.18111365e-01 3.23631614e-01 -2.79864579e-01 -1.12146246e+00
1.35117626e+00 -2.31408149e-01 3.10939729e-01 -3.25303495e-01
3.33343804e-01 3.21930274e-02 1.13489002e-01 -1.45803481e-01
2.28842162e-02 1.33101180e-01 5.04829176e-03 4.86292630e-01
2.91967601e-01 1.01830997e-01 -1.82126462e-01 2.38450810e-01
1.03383327e+00 -5.22451282e-01 2.77295947e-01 -4.76599276e-01
3.46871495e-01 -4.78242993e-01 5.83083510e-01 6.62982643e-01
-4.04280305e-01 4.59807128e-01 5.69787383e-01 -2.94777036e-01
-7.07662106e-01 -1.46682072e+00 -6.03192091e-01 7.84175813e-01
4.67683315e-01 -1.59161374e-01 -4.93414849e-01 -8.84999096e-01
3.37955385e-01 3.26093197e-01 -1.05250478e+00 -6.29622519e-01
3.52521986e-01 -7.29820728e-01 4.13763314e-01 4.48065609e-01
2.40870968e-01 -5.00933230e-01 -9.39889178e-02 -3.94661367e-01
-6.58349972e-03 -9.80903029e-01 -5.25210440e-01 5.81436634e-01
-7.59667933e-01 -1.34874511e+00 -6.45956278e-01 -8.41204643e-01
9.16569591e-01 4.30459350e-01 6.80776536e-01 -1.33436806e-02
-3.50675732e-01 6.25562787e-01 -3.43961269e-01 -1.86603814e-01
-8.65834802e-02 3.49261649e-02 2.91115642e-01 5.59162259e-01
5.85419238e-01 -3.43319863e-01 -4.19502199e-01 5.84435880e-01
-8.06245744e-01 -4.06347305e-01 6.68454051e-01 1.08429718e+00
7.34998882e-01 1.05332121e-01 3.22013021e-01 -8.26694548e-01
5.57780147e-01 -8.26745570e-01 -4.37263727e-01 6.46362960e-01
-9.29545105e-01 1.35284975e-01 5.39855540e-01 -7.76989996e-01
-5.06675303e-01 -1.41678005e-01 6.40782595e-01 -8.50776196e-01
2.26905812e-02 3.54252577e-01 -4.55894202e-01 -1.95535436e-01
8.25102091e-01 6.46942155e-03 4.77609485e-02 1.64319664e-01
6.85888052e-01 9.50194895e-01 4.97590303e-01 -5.07687569e-01
1.17480683e+00 3.99574727e-01 -7.30277672e-02 -7.12007821e-01
-7.49662578e-01 -7.44787574e-01 -9.85185683e-01 -7.26362765e-02
7.56074309e-01 -1.09248948e+00 -4.16682154e-01 8.84038657e-02
-5.43822050e-01 -4.43569897e-03 -4.48942721e-01 5.84822774e-01
-4.56882596e-01 5.85138500e-01 -2.68603176e-01 -7.22878277e-01
-1.45279512e-01 -1.07948780e+00 9.32141185e-01 1.33690938e-01
-8.42641890e-02 -1.29817784e+00 1.11988917e-01 4.57981259e-01
-1.50293544e-01 2.85651952e-01 9.88527000e-01 -1.19391489e+00
-6.21768534e-01 -3.36003900e-01 -4.26205307e-01 3.73700738e-01
9.51364189e-02 1.19420603e-01 -1.12781131e+00 -4.48955059e-01
-3.42947930e-01 -5.27391553e-01 1.02691066e+00 4.23042744e-01
1.25757539e+00 -1.70646563e-01 -5.94578087e-01 9.14810836e-01
1.23217499e+00 -2.23081917e-01 7.50236273e-01 1.91236332e-01
7.54704535e-01 5.52227080e-01 5.12578130e-01 3.42108876e-01
1.50151521e-01 3.45582932e-01 3.55746180e-01 -1.35561705e-01
1.14482574e-01 -4.42026228e-01 3.64313662e-01 8.07923019e-01
3.56962830e-01 -9.20712575e-02 -9.16394234e-01 4.82465863e-01
-1.69762552e+00 -9.23420787e-01 -8.71330574e-02 2.48088384e+00
5.25366008e-01 1.64060686e-02 -1.59414798e-01 1.46994635e-01
1.09671128e+00 1.46493837e-01 -7.16639280e-01 -2.85633981e-01
-4.25866634e-01 -9.25070047e-02 5.72617888e-01 1.55479953e-01
-1.04243410e+00 5.16562223e-01 6.13258410e+00 9.55540657e-01
-9.20322597e-01 -1.87874753e-02 9.47606206e-01 2.49219149e-01
-5.75363159e-01 9.78096947e-02 -6.46201193e-01 3.72570306e-01
9.66236353e-01 -2.19276428e-01 3.50025445e-01 8.26117039e-01
-4.12035465e-01 2.87397921e-01 -1.27290809e+00 1.24121976e+00
3.71189594e-01 -1.14112246e+00 2.69794136e-01 5.21862686e-01
9.58396792e-01 1.13018766e-01 3.46870810e-01 4.58570004e-01
6.67129993e-01 -1.15594149e+00 3.05381835e-01 4.19922173e-01
9.06259477e-01 -6.77456617e-01 6.33497596e-01 1.26884624e-01
-1.21136916e+00 -6.00803852e-01 -8.04640174e-01 3.45540583e-01
-5.77866197e-01 7.75005817e-01 -8.04643214e-01 6.09197140e-01
6.29867315e-01 8.98443162e-01 -9.98262465e-01 1.03399134e+00
-1.24580763e-01 4.16897148e-01 -1.91329196e-01 7.32969642e-02
-8.26816931e-02 -6.09864950e-01 2.77844459e-01 7.43369818e-01
1.35478422e-01 -1.03821300e-01 3.50449085e-01 9.39170957e-01
-2.73252070e-01 3.93836379e-01 -9.62586462e-01 -2.69730389e-01
6.47169590e-01 1.28003454e+00 -1.02627969e+00 -1.38980031e-01
-3.90654534e-01 8.12126994e-01 6.64439440e-01 4.29127336e-01
-8.17912459e-01 -4.12252069e-01 5.97340405e-01 -1.52822047e-01
4.16407138e-01 1.46042362e-01 -3.84210497e-01 -1.41558707e+00
-6.19282722e-02 -7.29992807e-01 9.23357844e-01 -5.64103007e-01
-1.93220723e+00 4.49754238e-01 -1.79846779e-01 -1.47924554e+00
3.43097478e-01 -8.25457275e-01 -6.69423223e-01 5.17986536e-01
-1.43110871e+00 -1.22621202e+00 -2.43649408e-01 6.80780292e-01
5.24314791e-02 -3.30026478e-01 8.21505904e-01 3.56489210e-03
-8.52973402e-01 9.22099113e-01 5.19702494e-01 4.59739566e-01
8.07927072e-01 -1.21960175e+00 -3.91955115e-02 8.41895223e-01
6.62283421e-01 6.92597449e-01 3.18090498e-01 -5.78011751e-01
-1.33943284e+00 -1.39837956e+00 2.48820409e-01 -6.74393594e-01
9.48214114e-01 -4.74594653e-01 -9.79879141e-01 5.54214060e-01
-2.87008345e-01 6.72019124e-01 1.21113563e+00 5.34607649e-01
-1.03346539e+00 -3.09389979e-01 -9.76598382e-01 2.93772608e-01
1.05921686e+00 -8.36909294e-01 -6.22108638e-01 4.21339422e-01
4.76310283e-01 2.29057833e-03 -1.15787613e+00 3.27556282e-01
2.89630175e-01 -6.49641395e-01 9.48300302e-01 -7.89275348e-01
1.49868980e-01 -3.60696137e-01 -4.89993304e-01 -1.44003010e+00
-6.10669851e-01 -2.14992851e-01 -7.06765801e-02 1.25530374e+00
3.20033640e-01 -8.80176008e-01 3.49594712e-01 4.52808380e-01
1.25098839e-01 -6.48335934e-01 -1.29252708e+00 -9.60551262e-01
2.07315147e-01 -1.66869774e-01 4.07134503e-01 1.23747206e+00
7.70590780e-03 4.45896804e-01 -1.08463265e-01 6.12132490e-01
7.77510405e-01 2.02853218e-01 8.39144826e-01 -1.90272772e+00
-2.35529225e-02 -2.85149068e-01 -1.22346199e+00 -7.87057698e-01
6.60476148e-01 -1.54256248e+00 -1.22039288e-01 -1.27129018e+00
6.51272476e-01 -4.63413686e-01 -7.62669981e-01 2.87747473e-01
-1.38547599e-01 2.93493807e-01 -6.68008551e-02 4.65508282e-01
-8.97873998e-01 9.79259253e-01 1.03822362e+00 -5.12705863e-01
-1.00613683e-01 -3.18740904e-01 -9.42517936e-01 3.29720944e-01
2.77820230e-01 -3.24675024e-01 -5.70932031e-01 1.31977141e-01
-6.32842705e-02 -3.22772473e-01 3.56806457e-01 -1.07168233e+00
1.21175349e-01 -1.86980844e-01 6.64655149e-01 -3.21920693e-01
1.10417046e-03 -9.29704964e-01 1.06732948e-02 3.43676537e-01
-4.91463214e-01 -3.93557429e-01 -1.55464783e-01 1.28633881e+00
-1.01560399e-01 1.33062497e-01 8.31664324e-01 4.98384565e-01
-6.07032239e-01 7.43675947e-01 1.53940275e-01 3.04099441e-01
1.60085177e+00 -5.86444259e-01 -6.01201594e-01 -1.48165286e-01
-4.73954022e-01 4.53222930e-01 3.78955871e-01 6.83089674e-01
6.86239719e-01 -1.61947680e+00 -5.07139683e-01 4.02358234e-01
8.12639415e-01 -2.59056181e-01 5.85871488e-02 7.93655932e-01
-8.97527635e-02 2.24084601e-01 -4.35188599e-02 -1.09808755e+00
-1.06812871e+00 9.20243382e-01 3.49924564e-01 -1.36637837e-01
-5.51060498e-01 7.87486851e-01 4.23340112e-01 -7.08648801e-01
4.07808840e-01 -2.24368408e-01 -2.41624638e-01 2.69389331e-01
3.72921020e-01 2.53787756e-01 -4.15592268e-02 -8.19961488e-01
-2.73469388e-01 6.43489301e-01 -4.98353802e-02 5.39930403e-01
1.30435133e+00 -1.76474258e-01 -1.60795674e-01 6.51452780e-01
1.61579001e+00 -2.21643552e-01 -1.14426327e+00 -3.61162215e-01
-2.98506655e-02 -7.15802491e-01 1.37036711e-01 -4.53232914e-01
-9.47889268e-01 7.47317314e-01 1.03094471e+00 1.28235206e-01
7.69279361e-01 2.60600299e-01 1.95936561e-01 4.09623861e-01
1.89625740e-01 -1.16740286e+00 1.54519990e-01 1.73247233e-01
7.34102905e-01 -1.38818872e+00 9.93214250e-02 -5.20181358e-01
-7.22712874e-01 1.15818787e+00 6.34834051e-01 -1.04300760e-01
9.22050714e-01 -1.89819932e-01 1.68620512e-01 -3.74820858e-01
-6.53355181e-01 -1.72547162e-01 6.55350268e-01 8.90891254e-01
-2.60447785e-02 3.21790963e-01 2.35178903e-01 4.55769032e-01
-1.21023260e-01 -5.68961799e-01 3.98474336e-01 1.00907080e-01
-4.11248773e-01 -5.60923159e-01 -4.26737159e-01 8.67970169e-01
2.18462020e-01 2.02982888e-01 -5.59885502e-01 4.10908967e-01
4.80444320e-02 8.23118925e-01 2.01397687e-01 -6.35075688e-01
-5.96743859e-02 1.75971344e-01 1.54234618e-01 -7.07433462e-01
-3.00648827e-02 -2.73090512e-01 -3.95233840e-01 -4.77490842e-01
-2.72951096e-01 -4.62640017e-01 -9.26429987e-01 -8.40046704e-02
-9.53507125e-01 3.39791507e-01 1.91051230e-01 7.01357245e-01
4.20016468e-01 2.18335271e-01 1.08606601e+00 -3.72596651e-01
-1.03720462e+00 -7.32677281e-01 -1.06433988e+00 7.13564217e-01
1.78511426e-01 -1.03302884e+00 -8.81834507e-01 -2.57459253e-01] | [9.595698356628418, 2.8291590213775635] |
b362df8c-520d-450e-b8a8-45565084c88b | generalizable-no-reference-image-quality | null | null | https://ieeexplore.ieee.org/abstract/document/9405680 | https://zhuhancheng.github.io/Hancheng_files/files/2021-TCSVT.pdf | Generalizable No-Reference Image Quality Assessment via Deep Meta-learning | Recently, researchers have shown great interest in using convolutional neural networks (CNNs) for no-reference image quality assessment (NR-IQA). Due to the lack of big training data, the efforts of existing metrics in optimizing CNN-based NR-IQA models remain limited. Furthermore, the diversity of distortions in images result in the generalization problem of NR-IQA models when trained with known distortions and tested on unseen distortions, which is an easy task for human. Hence, we propose a NR-IQA metric via deep meta-learning, which is highly generalizable in the face of unseen distortions. The fundamental idea is to learn the meta-knowledge shared by human when evaluating the quality of images with diversified distortions. Specifically, we define NR-IQA of different distortions as a series of tasks and propose a task selection strategy to build two task sets, which are characterized by synthetic to synthetic and synthetic to authentic distortions, respectively. Based on these two task sets, an optimization-based meta-learning is proposed to learn the generalized NR-IQA model, which can be directly used to evaluate the quality of images with unseen distortions. Extensive experiments demonstrate that our NR-IQA metric outperforms the state-of-the-arts in terms of both evaluation performance and generalization ability. | ['and Guangming Shi', 'Weisheng Dong', 'Jinjian Wu', 'Leida Li', 'Hancheng Zhu'] | 2021-04-15 | null | null | null | ieee-transactions-on-circuits-and-systems-for-4 | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [ 8.29033926e-02 -5.73279500e-01 1.25988409e-01 -3.75805974e-01
-7.29929745e-01 -1.52727976e-01 4.75906283e-01 -4.42448229e-01
-2.53021866e-01 3.30711067e-01 1.24172769e-01 -5.32944314e-02
-3.08130682e-01 -8.67506742e-01 -5.88482976e-01 -5.69129586e-01
2.53465027e-01 -9.77640525e-02 -1.15249880e-01 -3.86382878e-01
1.90721571e-01 2.75476545e-01 -1.54080153e+00 3.54698271e-01
1.01672196e+00 1.26724672e+00 1.03941292e-01 4.53549355e-01
1.92188144e-01 5.37852407e-01 -9.62240100e-01 -8.27357113e-01
4.74515766e-01 -5.68267107e-01 -6.73803270e-01 9.79582891e-02
5.22240043e-01 -5.59199929e-01 -7.72518754e-01 1.33082867e+00
8.63076389e-01 3.00158054e-01 6.37623966e-01 -1.35729086e+00
-1.19121742e+00 3.05743120e-03 -2.95043111e-01 4.11639929e-01
2.74128675e-01 5.15109003e-01 8.11341941e-01 -1.05438375e+00
1.95436060e-01 1.54741812e+00 4.59765345e-01 6.91737235e-01
-8.71165395e-01 -7.12695301e-01 -2.64073219e-02 5.79476297e-01
-1.45932090e+00 -4.72396344e-01 7.02018440e-01 -3.08094293e-01
5.93361974e-01 9.96315256e-02 2.17694268e-01 1.13979375e+00
1.68536052e-01 5.27705610e-01 1.19415140e+00 -8.76226425e-02
1.58118486e-01 -4.86760773e-02 -4.11609739e-01 3.83266807e-01
1.84268415e-01 4.14555997e-01 -3.97679567e-01 3.43711436e-01
8.11992705e-01 3.57618071e-02 -4.47041392e-01 -3.71962115e-02
-1.35606146e+00 4.12072539e-01 5.36468625e-01 1.43019751e-01
-1.67745262e-01 -1.11108966e-01 5.79142213e-01 6.23703420e-01
4.14652467e-01 5.16542077e-01 -2.27968007e-01 -3.72530730e-03
-7.01786160e-01 2.46708989e-01 3.34432960e-01 8.50989759e-01
5.80097616e-01 1.23284362e-01 -4.93149310e-01 1.20442450e+00
6.31257147e-02 5.03640115e-01 7.30325401e-01 -9.48140144e-01
7.05477834e-01 5.18810093e-01 2.38833234e-01 -1.46416473e+00
-1.41483486e-01 -7.55172551e-01 -1.28466487e+00 3.34863633e-01
2.61683673e-01 1.89037949e-01 -7.69046545e-01 1.68865275e+00
1.15922973e-01 1.34048268e-01 1.60031304e-01 1.13888896e+00
9.94714379e-01 7.22890139e-01 -8.28832760e-02 -1.90720424e-01
1.02439868e+00 -1.06787181e+00 -6.79785132e-01 1.14458814e-01
2.24807575e-01 -6.79936111e-01 1.31387162e+00 5.40987670e-01
-1.21625602e+00 -1.36089075e+00 -1.27360868e+00 -4.86227497e-02
-3.24186444e-01 -1.97217893e-02 3.49327140e-02 6.19159997e-01
-1.03101313e+00 7.70183146e-01 -3.12090844e-01 1.06403150e-01
5.67673147e-01 4.07535322e-02 -2.44686335e-01 -4.81445879e-01
-1.33944798e+00 8.41715813e-01 4.36905354e-01 3.39624316e-01
-1.39386582e+00 -6.44224942e-01 -6.01479828e-01 3.04778125e-02
3.61656278e-01 -6.32868350e-01 9.92078185e-01 -1.17046106e+00
-1.45394313e+00 8.25655520e-01 2.07951620e-01 -1.56390145e-02
7.37453997e-01 -4.20063436e-02 -1.09098375e+00 -1.95778999e-02
3.05369105e-02 5.48761904e-01 9.91053998e-01 -1.30109334e+00
-5.34999549e-01 -2.17842013e-01 4.37577903e-01 2.14006037e-01
-4.60281193e-01 -9.91836861e-02 -6.83923781e-01 -8.96483660e-01
-1.03621744e-01 -5.89564383e-01 3.02127376e-02 1.11488856e-01
-3.64724964e-01 -1.08980484e-01 5.54042637e-01 -6.77782595e-01
1.26054108e+00 -2.14580131e+00 9.29672271e-02 5.97000718e-02
3.59513462e-01 6.19530320e-01 -7.14040041e-01 1.39778778e-01
-1.57815456e-01 3.21506530e-01 -8.21800455e-02 -3.59488875e-02
1.18055884e-02 2.41302373e-03 -3.70473266e-02 1.84829190e-01
4.15906698e-01 9.49632287e-01 -1.11548269e+00 -4.69171435e-01
2.11983204e-01 4.48326081e-01 -2.12048277e-01 7.64815271e-01
7.88782835e-02 6.08782232e-01 -3.38784307e-01 6.71429873e-01
9.10298228e-01 -3.15873414e-01 -3.67127538e-01 -6.79116964e-01
1.35513365e-01 -1.69039309e-01 -1.14038861e+00 1.69150388e+00
-6.43559515e-01 3.30981016e-01 -4.08181250e-01 -1.01873052e+00
9.15344417e-01 2.37013102e-01 2.32523814e-01 -1.32789183e+00
1.36199743e-01 3.11678857e-01 1.12648599e-01 -7.61149526e-01
2.72818387e-01 -2.71066912e-02 2.09485754e-01 2.78951079e-01
2.98279166e-01 1.05766036e-01 2.47013733e-01 -1.36398897e-01
9.38756108e-01 -5.69031127e-02 5.43346144e-02 -1.05067166e-02
8.71405542e-01 -5.74056923e-01 7.18799055e-01 6.54657006e-01
-4.74478811e-01 9.02402341e-01 1.47319987e-01 -7.62304246e-01
-1.22140217e+00 -1.15526688e+00 -1.47139549e-01 8.69569659e-01
5.53326428e-01 -1.53675839e-01 -7.73901880e-01 -6.31529212e-01
-3.74153107e-01 2.88837433e-01 -6.71537280e-01 -4.64825481e-01
-5.33671916e-01 -8.63581598e-01 3.03814620e-01 3.20959717e-01
1.16010416e+00 -1.12246692e+00 -2.02233300e-01 6.63884804e-02
-4.04210091e-01 -1.15606880e+00 -5.42478979e-01 -5.03232658e-01
-5.96904755e-01 -1.11723256e+00 -1.11230123e+00 -6.48204327e-01
5.08398592e-01 5.39380193e-01 1.25817406e+00 3.69982034e-01
-1.78752452e-01 1.91404924e-01 -4.70645905e-01 -2.18093589e-01
-5.52507997e-01 -2.70391434e-01 -4.20763418e-02 4.12888050e-01
-4.50683460e-02 -5.84506392e-01 -1.15134859e+00 8.13346207e-01
-1.13851261e+00 6.03375621e-02 8.64812851e-01 8.55000496e-01
6.00222290e-01 3.56892258e-01 8.41520548e-01 -4.26762938e-01
7.65199661e-01 -4.83432859e-01 -3.52490306e-01 4.85404819e-01
-6.82771981e-01 -1.36941522e-01 8.13795447e-01 -5.08800805e-01
-1.01271725e+00 -7.29638457e-01 -8.37760195e-02 -6.55086339e-01
-7.54984021e-02 4.41150516e-01 -7.53074586e-01 -3.33074331e-01
6.26526356e-01 2.71278918e-01 -2.41634384e-01 -3.41274679e-01
3.33808750e-01 6.60147369e-01 8.77955019e-01 -5.12974858e-01
1.02150857e+00 3.09729993e-01 5.57199009e-02 -3.35214555e-01
-1.01306474e+00 -1.62601888e-01 -3.68487000e-01 -4.17410463e-01
7.53024697e-01 -8.99637222e-01 -4.87114102e-01 9.68863785e-01
-1.18827331e+00 -1.16963826e-01 -1.01698905e-01 3.13345939e-01
-5.37144482e-01 4.60427880e-01 -3.58689964e-01 -3.43888551e-01
-3.96334171e-01 -1.53202987e+00 8.54030371e-01 3.33378851e-01
3.80866826e-01 -7.68589914e-01 -4.07559983e-02 4.79226738e-01
7.28927612e-01 3.46389890e-01 9.77788031e-01 -3.20670217e-01
-5.97478926e-01 -8.18666890e-02 -5.44809699e-01 9.52075303e-01
2.23011449e-01 -2.39864588e-01 -9.82787609e-01 -5.37191510e-01
6.38401657e-02 -4.84031826e-01 5.43058991e-01 6.98860884e-02
1.70655680e+00 -4.24901992e-01 2.97186464e-01 9.10006523e-01
1.33508849e+00 3.03021342e-01 1.08875728e+00 5.12892842e-01
5.78806758e-01 3.69480073e-01 6.38886929e-01 2.59993464e-01
3.38393956e-01 8.05310547e-01 6.18293881e-01 -1.39759764e-01
-3.83093297e-01 -1.68722197e-01 1.82620779e-01 1.05975914e+00
-2.10089147e-01 -4.76355582e-01 -6.21541858e-01 5.74905694e-01
-1.45320034e+00 -9.41761672e-01 4.12621200e-01 2.19626713e+00
7.57696748e-01 8.08210149e-02 -4.95854504e-02 3.64317238e-01
8.84669840e-01 2.40916014e-01 -8.12969685e-01 -1.08788840e-01
-2.25970939e-01 9.92006734e-02 -1.03467945e-02 -1.65101975e-01
-1.12128448e+00 4.64124680e-01 5.87634230e+00 1.09569526e+00
-1.16117275e+00 2.25788042e-01 1.02437949e+00 4.01463360e-02
-1.58364922e-01 -4.01776522e-01 -2.67681301e-01 8.29034209e-01
8.64401460e-01 -3.41796339e-01 5.75465918e-01 6.71401143e-01
1.98131740e-01 4.87549514e-01 -1.16049254e+00 1.42762756e+00
2.25113824e-01 -9.52227116e-01 5.85653961e-01 -9.37356949e-02
1.10392821e+00 -2.67213613e-01 6.98989987e-01 4.15014356e-01
-1.19532116e-01 -1.30097365e+00 6.47562206e-01 5.96249580e-01
1.14365077e+00 -7.69304931e-01 8.51720870e-01 1.48284107e-01
-1.01801872e+00 -2.89562464e-01 -7.70654559e-01 2.37971723e-01
-6.43279850e-02 5.41624129e-01 -1.04409404e-01 9.03532147e-01
8.46607149e-01 6.50882483e-01 -7.85694540e-01 1.19243574e+00
-2.68095732e-02 3.01720381e-01 4.98635471e-01 3.91710490e-01
1.01334512e-01 -1.16367184e-01 3.35708499e-01 8.98807526e-01
5.78514576e-01 2.10539132e-01 -1.01623230e-01 9.54071999e-01
-5.44966877e-01 6.36002347e-02 -2.13807911e-01 2.01981172e-01
3.73464912e-01 1.23604023e+00 -1.44351631e-01 -1.76271155e-01
-4.66130555e-01 1.05659664e+00 1.01060815e-01 5.01560271e-01
-7.33095646e-01 -6.03215396e-01 5.59608519e-01 -2.47212350e-02
2.01972388e-02 7.50050023e-02 1.02020837e-01 -1.27131093e+00
2.77015895e-01 -1.34944558e+00 1.55328453e-01 -9.86600041e-01
-1.56304455e+00 9.17443752e-01 -6.02693530e-03 -1.69247127e+00
-4.95166704e-03 -6.01222038e-01 -7.78844774e-01 9.42788720e-01
-1.83037055e+00 -1.05884254e+00 -9.78880465e-01 7.56206453e-01
5.87100267e-01 -3.99751753e-01 4.55836862e-01 6.05315804e-01
-6.15664721e-01 1.09762704e+00 5.30135743e-02 4.20091629e-01
7.28752911e-01 -9.06086087e-01 5.70819676e-01 9.81625557e-01
-5.30415997e-02 4.12676454e-01 3.62206966e-01 -2.60081261e-01
-1.24613893e+00 -1.54005587e+00 3.25120836e-01 -2.23825231e-01
2.14823470e-01 4.27584425e-02 -1.17241919e+00 1.28719434e-02
8.33406076e-02 4.59242880e-01 4.97399777e-01 -3.38616639e-01
-5.99969864e-01 -5.70931852e-01 -1.25730777e+00 4.78159130e-01
1.29269373e+00 -6.17498755e-01 -3.91645581e-01 1.50122613e-01
9.22493100e-01 -3.54303479e-01 -1.14340699e+00 7.58404851e-01
5.21599293e-01 -1.15760434e+00 1.16948652e+00 -5.05544543e-01
7.28910387e-01 -2.83007115e-01 -2.30961055e-01 -1.62077510e+00
-4.39914614e-01 -3.05323869e-01 -3.23722921e-02 1.21243334e+00
1.29627049e-01 -4.92114842e-01 2.69620776e-01 2.64087975e-01
-2.04599559e-01 -7.75936425e-01 -9.06348169e-01 -1.09858370e+00
1.09203249e-01 -2.68852770e-01 1.14199328e+00 8.90566707e-01
-6.46961153e-01 3.05705015e-02 -5.43177187e-01 2.32904315e-01
7.35413194e-01 -1.08713441e-01 8.14074457e-01 -9.98429418e-01
-2.56560773e-01 -4.42324549e-01 -7.68913269e-01 -8.59685659e-01
1.61713790e-02 -6.78331852e-01 -1.08188547e-01 -1.25895107e+00
3.35499048e-01 -3.27144742e-01 -8.44980717e-01 1.29996464e-01
-4.12019849e-01 3.10286909e-01 2.31368005e-01 5.51398695e-01
-8.49030793e-01 8.54012489e-01 1.78619444e+00 -5.49539447e-01
8.86061266e-02 -1.74051121e-01 -6.52926385e-01 4.54428375e-01
5.74929476e-01 -2.44300753e-01 -5.32437742e-01 -8.50568652e-01
3.24194670e-01 9.13617015e-02 5.62852144e-01 -1.43384564e+00
-7.52359442e-03 -2.40859270e-01 5.22501111e-01 -3.73772115e-01
1.00532241e-01 -5.65229774e-01 1.03396073e-01 1.41005382e-01
-4.92057025e-01 2.09678710e-01 -9.23349559e-02 5.32208085e-01
-5.90260863e-01 -5.88642359e-02 9.36072350e-01 -2.66761720e-01
-7.61729479e-01 8.29242647e-01 4.70434040e-01 2.77621448e-01
6.37296736e-01 -2.67366529e-01 -4.08786297e-01 -5.15985906e-01
-3.27637523e-01 -3.81439272e-03 3.95030707e-01 7.18954086e-01
1.02905273e+00 -1.80302131e+00 -1.09237778e+00 -1.78475175e-02
4.45928037e-01 -2.11534929e-02 7.39143431e-01 4.60728496e-01
-3.58486235e-01 8.16625804e-02 -6.59126520e-01 -4.89175022e-01
-8.20331037e-01 8.44709814e-01 5.91600716e-01 -2.34645560e-01
-1.78037599e-01 5.90862393e-01 3.96007538e-01 -2.55544752e-01
1.10362850e-01 -2.14079954e-02 -2.62732208e-01 -3.67456019e-01
8.60188603e-01 5.53508639e-01 3.60095620e-01 -8.28238189e-01
7.45927636e-03 6.22508109e-01 4.20358963e-02 1.04375951e-01
1.13515413e+00 -2.32508078e-01 5.03487289e-02 1.24751471e-01
1.62801027e+00 -5.41440785e-01 -1.20089829e+00 -4.79390293e-01
-3.11575502e-01 -9.53836203e-01 2.99514942e-02 -1.05017269e+00
-1.46597922e+00 1.16155243e+00 1.17593706e+00 -1.10904120e-01
1.50075126e+00 -3.90205801e-01 9.07381952e-01 3.75027239e-01
3.55005056e-01 -9.64631557e-01 6.94270968e-01 1.76810861e-01
1.26890290e+00 -1.48932981e+00 -1.58340439e-01 1.19105436e-01
-4.46703911e-01 1.02761257e+00 8.81414354e-01 4.41196673e-02
3.76141310e-01 -4.79467243e-01 1.51472047e-01 4.45178114e-02
-4.85113293e-01 3.82292345e-02 6.79466784e-01 8.12759042e-01
8.68024975e-02 -1.55302241e-01 -5.33753075e-02 6.27883792e-01
-3.37973461e-02 1.26278177e-01 4.01980937e-01 3.92155409e-01
-1.51698902e-01 -9.15897787e-01 -2.87959784e-01 4.20104623e-01
-4.29550648e-01 2.94711608e-02 3.69132645e-02 5.87003469e-01
4.51702029e-01 1.22638428e+00 -1.82100400e-01 -8.90754879e-01
6.71096623e-01 -5.43215632e-01 3.17338139e-01 -2.60642618e-01
-4.16087270e-01 -3.93112242e-01 -3.57611537e-01 -7.25945950e-01
-5.49198985e-01 -1.67190611e-01 -4.26235616e-01 -3.09918910e-01
-2.73020506e-01 -1.40325382e-01 4.68445241e-01 8.80391717e-01
3.30780178e-01 5.40121734e-01 1.20731151e+00 -7.73821056e-01
-7.92256474e-01 -8.89829934e-01 -3.69061559e-01 9.23008800e-01
3.02088708e-01 -7.33129740e-01 -3.38205069e-01 -1.05982035e-01] | [11.850998878479004, -1.8400912284851074] |
be35f97d-d2c4-4852-b280-3260a6d9d77b | anomaly-detection-in-time-series-with-triadic | 2012.04936 | null | https://arxiv.org/abs/2012.04936v1 | https://arxiv.org/pdf/2012.04936v1.pdf | Anomaly Detection in Time Series with Triadic Motif Fields and Application in Atrial Fibrillation ECG Classification | In the time-series analysis, the time series motifs and the order patterns in time series can reveal general temporal patterns and dynamic features. Triadic Motif Field (TMF) is a simple and effective time-series image encoding method based on triadic time series motifs. Electrocardiography (ECG) signals are time-series data widely used to diagnose various cardiac anomalies. The TMF images contain the features characterizing the normal and Atrial Fibrillation (AF) ECG signals. Considering the quasi-periodic characteristics of ECG signals, the dynamic features can be extracted from the TMF images with the transfer learning pre-trained convolutional neural network (CNN) models. With the extracted features, the simple classifiers, such as the Multi-Layer Perceptron (MLP), the logistic regression, and the random forest, can be applied for accurate anomaly detection. With the test dataset of the PhysioNet Challenge 2017 database, the TMF classification model with the VGG16 transfer learning model and MLP classifier demonstrates the best performance with the 95.50% ROC-AUC and 88.43% F1 score in the AF classification. Besides, the TMF classification model can identify AF patients in the test dataset with high precision. The feature vectors extracted from the TMF images show clear patient-wise clustering with the t-distributed Stochastic Neighbor Embedding technique. Above all, the TMF classification model has very good clinical interpretability. The patterns revealed by symmetrized Gradient-weighted Class Activation Mapping have a clear clinical interpretation at the beat and rhythm levels. | ['Xin Chen', 'Yadong Zhang'] | 2020-12-09 | null | null | null | null | ['ecg-classification', 'atrial-fibrillation-detection', 'electrocardiography-ecg'] | ['medical', 'medical', 'methodology'] | [ 2.49341041e-01 -4.55490708e-01 -4.39523198e-02 -2.62410551e-01
-4.13401663e-01 -3.60378027e-01 9.09678731e-03 1.84762388e-01
-1.24229632e-01 6.48207188e-01 1.55566148e-02 -4.96127605e-01
-6.59909248e-01 -6.13428116e-01 -2.47765258e-01 -9.21024501e-01
-9.60187793e-01 1.71021909e-01 -2.96251833e-01 5.71647920e-02
1.79809198e-01 4.62796986e-01 -1.31682181e+00 6.96300805e-01
9.03442562e-01 1.49104834e+00 -2.83794254e-01 8.27682853e-01
1.59053672e-02 4.43515033e-01 -5.62830687e-01 2.11598024e-01
9.56170354e-03 -6.81410611e-01 -4.31018442e-01 -4.18890685e-01
-2.20638856e-01 1.90244451e-01 -4.38115060e-01 6.06404901e-01
9.18496668e-01 -4.49288726e-01 6.85111284e-01 -1.26774907e+00
-3.17872137e-01 3.46940309e-01 -3.76426637e-01 9.62384522e-01
3.54678661e-01 1.36112683e-02 5.97451568e-01 -8.39418113e-01
3.24027449e-01 6.63626611e-01 9.62403357e-01 1.13111667e-01
-1.07592702e+00 -5.47952116e-01 -3.61753792e-01 8.59983385e-01
-1.32294214e+00 2.11720318e-01 1.09789348e+00 -6.64233267e-01
9.95014012e-01 5.46103716e-01 1.28085005e+00 1.04820764e+00
8.38418663e-01 3.79746437e-01 1.21147609e+00 -3.17481726e-01
-1.25873417e-01 -5.76538622e-01 1.85657680e-01 6.68638468e-01
-4.30073105e-02 2.70708501e-01 -3.95478040e-01 -6.24049187e-01
5.95128119e-01 4.89731371e-01 -6.05371714e-01 6.64186776e-02
-1.74786770e+00 5.94016194e-01 3.77011865e-01 6.97673857e-01
-5.79553246e-01 5.54121062e-02 9.59377348e-01 6.78599417e-01
3.83349240e-01 4.71511155e-01 -5.70546508e-01 -4.74314749e-01
-8.14022422e-01 -1.73185676e-01 3.97143453e-01 2.39474684e-01
2.33153746e-01 3.33461940e-01 -2.87965506e-01 6.42168045e-01
2.06204370e-01 5.71805000e-01 1.04161346e+00 -6.87249780e-01
2.51793295e-01 6.84909046e-01 -4.62025791e-01 -1.42228007e+00
-6.11153781e-01 -5.98550022e-01 -1.46537554e+00 -1.17562234e-01
3.15547317e-01 -4.62099276e-02 -7.15588808e-01 1.35629487e+00
-6.82103410e-02 6.67105138e-01 1.24161758e-01 6.20730221e-01
7.26038933e-01 6.87937975e-01 -1.44236431e-01 -6.23543978e-01
1.28516757e+00 -1.76995099e-01 -9.06892538e-01 2.94492006e-01
9.04468179e-01 -4.09488142e-01 7.98208177e-01 3.97866726e-01
-4.25952017e-01 -5.08235097e-01 -1.18699467e+00 5.52450180e-01
-1.99370027e-01 2.64598280e-01 4.86613870e-01 4.95598763e-01
-7.27686226e-01 1.00418043e+00 -1.05397677e+00 -1.92631170e-01
6.65965378e-01 2.06401989e-01 -4.78809774e-01 7.62828067e-02
-1.34774888e+00 5.53060949e-01 2.11818784e-01 5.11089981e-01
-5.27449131e-01 -6.31610811e-01 -7.26312160e-01 3.16339210e-02
-5.38238704e-01 -4.89433467e-01 3.28202218e-01 -8.44589770e-01
-1.07897365e+00 8.48307967e-01 -9.83387455e-02 -4.97902006e-01
3.04801852e-01 1.43800363e-01 -8.82165492e-01 4.03768450e-01
1.07758962e-01 -1.38428569e-01 1.01033556e+00 -3.62250835e-01
-2.80410916e-01 -5.83787382e-01 -6.65931344e-01 -1.00757375e-01
-3.59782159e-01 -1.63838446e-01 5.96690536e-01 -9.32572842e-01
5.51930308e-01 -8.48004758e-01 -1.04858845e-01 -1.00903004e-01
-2.16175243e-01 -1.26137316e-01 1.16392827e+00 -7.49855638e-01
1.71570086e+00 -2.53183484e+00 2.86526263e-01 4.30844486e-01
5.35154998e-01 1.15560785e-01 2.38947019e-01 4.84007984e-01
-5.65810382e-01 1.42872840e-01 -4.84982163e-01 5.49859881e-01
-5.99985301e-01 2.63315111e-01 -2.38192230e-01 6.29073560e-01
1.25914454e-01 1.15095127e+00 -8.09851706e-01 -4.30537730e-01
1.51913106e-01 3.00618708e-01 -1.03606313e-01 9.10582617e-02
4.40656483e-01 7.56992698e-01 -3.71391058e-01 5.26201367e-01
3.57466072e-01 -4.31309700e-01 5.40468097e-02 -4.14979011e-01
2.64439225e-01 -1.63253739e-01 -6.01399541e-01 1.71839297e+00
-1.19513553e-02 7.38801122e-01 -6.42247438e-01 -1.41993654e+00
1.04713070e+00 7.26638436e-01 8.96619141e-01 -6.80023074e-01
-4.11594659e-02 4.96894509e-01 5.76301873e-01 -1.04791546e+00
-6.75451458e-01 2.84653064e-03 5.65822236e-02 4.97544736e-01
-2.65014786e-02 6.70696020e-01 -2.86196768e-01 -2.81743467e-01
1.21495974e+00 -1.60744801e-01 3.47903103e-01 -3.55485916e-01
5.26690543e-01 -3.24603826e-01 6.51636839e-01 3.83843154e-01
-2.63512433e-01 7.92040527e-01 5.11468649e-01 -1.28047860e+00
-8.02354634e-01 -9.00207341e-01 -6.20553315e-01 2.41521731e-01
-2.88064480e-01 -4.26981390e-01 -2.56764412e-01 -7.18290687e-01
1.30691677e-01 1.93430353e-02 -6.91110909e-01 -4.89959121e-01
-6.98457181e-01 -1.08168209e+00 9.56425846e-01 6.88911378e-01
4.03143674e-01 -1.37785435e+00 -9.36827004e-01 3.85145664e-01
-3.96141887e-01 -5.98393023e-01 -1.47327542e-01 2.05506235e-01
-1.36426449e+00 -1.18171978e+00 -8.89404178e-01 -8.08979154e-01
3.70584786e-01 -3.67010057e-01 6.25615537e-01 1.73199311e-01
-6.91418886e-01 1.20472573e-01 -4.39868927e-01 -3.56120855e-01
-1.24101050e-01 -2.97441125e-01 1.70309007e-01 5.08229434e-01
4.43489343e-01 -1.19948542e+00 -8.59762013e-01 1.59422487e-01
-4.21341687e-01 -2.36568436e-01 5.34490705e-01 1.17278028e+00
7.76598215e-01 -1.63224056e-01 1.00229514e+00 -5.63077569e-01
6.75491810e-01 -5.00971138e-01 1.24025412e-01 3.67907397e-02
-9.24521506e-01 -3.67582649e-01 7.05570698e-01 -4.85786676e-01
-6.50470704e-02 -2.96465635e-01 -4.98378696e-03 -8.34763706e-01
-1.46528289e-01 9.17011857e-01 1.96766719e-01 1.47742003e-01
1.02755713e+00 7.85952866e-01 2.34586537e-01 -2.00327039e-01
-2.25335091e-01 7.12399185e-01 3.34766746e-01 -3.49661410e-01
3.98337930e-01 4.17515397e-01 1.37990013e-01 -8.68936956e-01
-1.63236126e-01 -3.23403180e-01 -6.14271462e-01 -3.10955375e-01
1.00058818e+00 -6.01765811e-01 -5.67405105e-01 6.54093623e-01
-1.11371565e+00 1.35552436e-01 -4.21620578e-01 8.17566335e-01
-5.76247811e-01 6.92454457e-01 -6.22382820e-01 -7.20309258e-01
-7.22751021e-01 -7.87468314e-01 6.83580697e-01 -2.23491326e-01
-3.58272582e-01 -9.61887538e-01 1.53282583e-02 -2.99352974e-01
4.35397714e-01 1.03223383e+00 1.47589850e+00 -7.20335126e-01
-5.48451692e-02 -4.11939800e-01 2.37548336e-01 4.20884311e-01
3.81781250e-01 -6.64836168e-02 -9.25433457e-01 -4.48227763e-01
3.01065892e-01 2.71113336e-01 7.15156078e-01 7.01923668e-01
1.60927904e+00 -2.00639948e-01 -3.37148011e-01 1.06463230e+00
1.07061350e+00 8.51715088e-01 7.87176549e-01 3.12327415e-01
7.24079967e-01 1.96585730e-01 2.45338365e-01 3.46306294e-01
2.68660374e-02 3.07033449e-01 1.45416141e-01 -2.89718211e-02
3.52144480e-01 -1.17405213e-01 1.40442789e-01 1.26143122e+00
-6.72670901e-01 2.31653392e-01 -1.23589087e+00 4.63447034e-01
-1.93031061e+00 -1.29667616e+00 -4.26638275e-01 2.07087922e+00
4.28095669e-01 1.72914825e-02 -5.84894791e-02 9.87498403e-01
7.17867672e-01 1.72757402e-01 -4.94378209e-01 -4.39543277e-01
-3.78838420e-01 2.26823568e-01 -1.24667838e-01 -1.86953038e-01
-1.03892314e+00 -1.23576596e-01 5.89645195e+00 5.47616124e-01
-1.61569047e+00 2.95117032e-02 7.56284297e-01 2.84779191e-01
-9.59090292e-02 -3.01620305e-01 2.11009562e-01 8.43628049e-01
1.24261236e+00 -4.01394725e-01 2.57716298e-01 4.03811157e-01
3.02767962e-01 6.08266711e-01 -1.02501631e+00 1.64535499e+00
-7.40692466e-02 -1.37095308e+00 6.79574115e-03 7.21420795e-02
1.06166184e-01 9.74349864e-03 9.66338888e-02 3.93490046e-02
-1.06308079e+00 -1.45522213e+00 1.23799078e-01 8.33555043e-01
1.46735239e+00 -5.43551862e-01 9.47504997e-01 1.79065526e-01
-1.33844066e+00 -5.31444490e-01 -1.57814950e-01 -1.37297183e-01
5.75629249e-02 7.76531577e-01 -4.94388431e-01 7.33369708e-01
1.16714346e+00 1.53740704e+00 -4.21631664e-01 1.04514575e+00
7.45701492e-02 1.13292575e+00 -2.77020782e-01 1.26683056e-01
-2.33414590e-01 -4.04242545e-01 8.35715830e-01 1.03799009e+00
5.83277822e-01 1.15015022e-01 -1.40364289e-01 5.34381747e-01
5.36049306e-01 2.22560227e-01 -9.06968713e-01 -2.82895058e-01
1.09456711e-01 1.12800062e+00 -7.13956952e-01 -1.36884436e-01
-1.68088332e-01 7.88574398e-01 -3.09186429e-01 5.00574887e-01
-6.16554081e-01 -8.23046982e-01 2.44579747e-01 3.21077883e-01
3.46882306e-02 -8.17964878e-03 -4.80126202e-01 -1.27471232e+00
4.96938795e-01 -8.82752478e-01 6.94325089e-01 -5.92993915e-01
-1.51387286e+00 1.12136936e+00 -1.31538808e-01 -1.90085888e+00
-2.31156930e-01 -5.22539377e-01 -9.88811970e-01 9.44338918e-01
-1.29032445e+00 -7.99523294e-01 -3.98255467e-01 8.20297599e-01
-3.74102779e-02 -6.11199319e-01 1.43599319e+00 5.03283858e-01
-2.86993057e-01 5.06669104e-01 2.31414586e-02 4.02660340e-01
3.36918920e-01 -1.30389440e+00 -3.59416083e-02 4.53219891e-01
7.56018981e-02 5.40600181e-01 1.08383089e-01 -4.36825335e-01
-1.02524292e+00 -1.08864534e+00 9.60675478e-01 -3.56125832e-01
2.67289937e-01 -7.82605261e-02 -1.16848707e+00 4.36021328e-01
-1.30768448e-01 5.81685126e-01 1.01288283e+00 -1.10483997e-01
-1.91674903e-01 -2.91904360e-01 -9.87054288e-01 2.89547890e-02
8.99443865e-01 -7.60353327e-01 -8.36774290e-01 3.91453207e-01
2.60770321e-01 -2.92076230e-01 -1.35817778e+00 7.91130424e-01
9.10890579e-01 -7.59384453e-01 9.49537635e-01 -8.31375301e-01
5.02801776e-01 -3.05748105e-01 -3.17362696e-02 -1.25690031e+00
-4.73749608e-01 -7.30832040e-01 -2.71894634e-01 5.73897839e-01
3.76328498e-01 -1.04530919e+00 5.22578359e-01 -3.68026078e-01
-1.36587664e-01 -1.34149432e+00 -1.43447423e+00 -5.85705638e-01
-1.88689038e-01 -4.41756159e-01 4.94310319e-01 1.18830061e+00
2.69101322e-01 1.56531617e-01 -2.08896950e-01 -1.62904430e-02
3.62820864e-01 2.61978567e-01 6.53975308e-02 -1.47626066e+00
-8.41825530e-02 -1.80515438e-01 -1.18800139e+00 -1.67176783e-01
-1.77774653e-01 -1.48185492e+00 -6.38541102e-01 -1.21742463e+00
-2.23649383e-01 -5.34853339e-01 -1.06059623e+00 3.22814643e-01
-1.03280179e-01 1.24466293e-01 -1.42830193e-01 4.99367386e-01
1.24016823e-02 4.42748427e-01 1.28060687e+00 -2.56119281e-01
-2.74540693e-01 2.91661710e-01 -1.07002355e-01 5.08895338e-01
8.29276145e-01 -5.43477595e-01 -3.97192597e-01 -6.25645295e-02
1.67242810e-01 4.25445259e-01 4.37004268e-01 -1.09059370e+00
4.40766662e-02 3.78502935e-01 8.50090325e-01 -3.78215045e-01
5.04587777e-02 -7.36813545e-01 3.48817438e-01 9.25255001e-01
-5.80110215e-02 8.12300205e-01 1.48213387e-01 7.38748372e-01
-6.62532330e-01 3.82066697e-01 4.07841116e-01 5.50497919e-02
-2.56817997e-01 6.41142130e-01 -5.64981878e-01 7.46041685e-02
1.01904941e+00 -6.34709716e-01 -5.77162579e-02 -1.97281703e-01
-1.12703836e+00 -1.22605912e-01 -3.01740378e-01 3.37946981e-01
1.06724024e+00 -1.78832972e+00 -8.59724820e-01 6.86402261e-01
4.45028186e-01 -1.89021721e-01 5.17706275e-01 1.43860483e+00
-7.33958364e-01 1.38959393e-01 -6.67056918e-01 -1.31925440e+00
-1.03586912e+00 3.56663615e-01 7.57325709e-01 -8.95134136e-02
-1.20965517e+00 3.35635275e-01 -2.78490752e-01 -3.18720676e-02
-7.60731250e-02 -4.67363775e-01 -6.83953345e-01 1.49266675e-01
4.22078282e-01 3.37483555e-01 1.65984586e-01 -5.63021004e-01
-5.69201350e-01 9.81541514e-01 3.24357390e-01 2.11505011e-01
1.38556063e+00 4.24305230e-01 -4.62563574e-01 8.39176357e-01
1.50984073e+00 -3.70354265e-01 -4.07593250e-01 3.18428203e-02
-6.33812547e-02 -3.18408191e-01 -1.96253315e-01 -7.22804189e-01
-1.09350204e+00 1.37090814e+00 1.30812931e+00 4.57928628e-01
1.31337607e+00 -3.47328424e-01 7.87533760e-01 1.58091903e-01
2.62256861e-01 -5.03372550e-01 -1.05601408e-01 1.08655564e-01
9.77468014e-01 -7.99128413e-01 -3.42234373e-01 -9.81435366e-03
-4.91869211e-01 1.64012182e+00 6.51505813e-02 -2.35700086e-01
1.19718432e+00 5.04182093e-02 3.79724622e-01 -3.82989645e-01
-5.38683653e-01 4.88556415e-01 4.24542159e-01 7.84722745e-01
3.27850103e-01 2.21434981e-01 -6.52520597e-01 8.78744841e-01
-1.33377045e-01 8.00804347e-02 1.28207177e-01 6.50223315e-01
-7.06086755e-02 -8.10381711e-01 3.61353159e-02 1.06908762e+00
-6.74509227e-01 1.62650496e-01 7.11427033e-02 2.77852595e-01
1.98709771e-01 6.37662232e-01 7.46628195e-02 -7.46594071e-01
3.16371650e-01 6.04812980e-01 3.09172004e-01 -2.69113153e-01
-6.35459065e-01 -6.98240846e-02 -3.14999223e-01 -6.42048717e-01
-4.08104032e-01 -3.85827482e-01 -1.23856115e+00 1.60062224e-01
-4.96775918e-02 3.30968291e-01 2.82366633e-01 7.63032496e-01
7.29240179e-01 7.06187069e-01 8.15556347e-01 -4.83132839e-01
-2.28519350e-01 -1.11404574e+00 -7.21574247e-01 5.52548110e-01
6.11523926e-01 -4.05579388e-01 -5.93657553e-01 2.64297072e-02] | [14.26470947265625, 3.2451422214508057] |
fce1405c-01aa-4392-9331-91f83256cb7d | modeling-user-behavior-with-interaction | 2207.10767 | null | https://arxiv.org/abs/2207.10767v1 | https://arxiv.org/pdf/2207.10767v1.pdf | Modeling User Behavior With Interaction Networks for Spam Detection | Spam is a serious problem plaguing web-scale digital platforms which facilitate user content creation and distribution. It compromises platform's integrity, performance of services like recommendation and search, and overall business. Spammers engage in a variety of abusive and evasive behavior which are distinct from non-spammers. Users' complex behavior can be well represented by a heterogeneous graph rich with node and edge attributes. Learning to identify spammers in such a graph for a web-scale platform is challenging because of its structural complexity and size. In this paper, we propose SEINE (Spam DEtection using Interaction NEtworks), a spam detection model over a novel graph framework. Our graph simultaneously captures rich users' details and behavior and enables learning on a billion-scale graph. Our model considers neighborhood along with edge types and attributes, allowing it to capture a wide range of spammers. SEINE, trained on a real dataset of tens of millions of nodes and billions of edges, achieves a high performance of 80% recall with 1% false positive rate. SEINE achieves comparable performance to the state-of-the-art techniques on a public dataset while being pragmatic to be used in a large-scale production system. | ['Charles Rosenberg', 'Vishwakarma Singh', 'Manisha Srivastava', 'Prabhat Agarwal'] | 2022-07-21 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [-3.45438868e-01 -1.29793912e-01 -2.34018475e-01 1.43937841e-01
-2.31497899e-01 -1.02957094e+00 7.81629503e-01 2.78911501e-01
9.78011191e-02 2.52628922e-01 -3.02010328e-02 -5.71356416e-01
4.62888293e-02 -1.14509869e+00 -3.65269244e-01 -1.71726003e-01
-4.41084802e-01 4.83410418e-01 9.63285327e-01 -6.49416864e-01
1.57305688e-01 3.75461638e-01 -1.10661519e+00 2.90570587e-01
9.52117264e-01 9.84727323e-01 -2.01925918e-01 4.25325483e-01
-1.91714048e-01 6.06837094e-01 -5.81365764e-01 -9.41223264e-01
3.58333498e-01 8.23321790e-02 -5.51071405e-01 5.41520007e-02
6.86442494e-01 -1.37436360e-01 -6.87251151e-01 1.45425081e+00
2.30355039e-01 -1.31215602e-01 4.74362463e-01 -1.53353608e+00
-5.80954790e-01 5.13332069e-01 -5.62774420e-01 2.14087948e-01
3.66964161e-01 1.89820692e-01 1.27410197e+00 -3.12691629e-01
7.21108556e-01 1.39546978e+00 9.30209696e-01 2.35842139e-01
-1.23863411e+00 -7.37883389e-01 1.84099302e-01 -9.45094004e-02
-9.29692090e-01 -1.31325666e-02 5.84431529e-01 -1.67980045e-01
5.81362963e-01 4.59648967e-01 7.78839588e-01 1.56370854e+00
3.37176979e-01 7.27673888e-01 7.95947850e-01 2.58764744e-01
1.56004384e-01 1.28883168e-01 5.77703118e-01 9.41667020e-01
8.29617858e-01 -2.69236505e-01 -4.12819237e-01 -9.32202041e-01
4.56647128e-01 2.05512092e-01 -3.29083800e-02 -4.37171072e-01
-5.86394608e-01 1.00061679e+00 7.66467035e-01 1.32608965e-01
-7.23294243e-02 2.13682562e-01 5.21964669e-01 7.35139906e-01
5.45977712e-01 4.71696258e-01 -2.96939939e-01 -8.05665776e-02
-4.95988697e-01 1.76767111e-01 1.53034127e+00 9.89895165e-01
5.37018299e-01 3.42396945e-02 1.71892121e-01 8.12343895e-01
2.77331799e-01 5.71340263e-01 6.84802532e-01 -3.59118521e-01
2.66880453e-01 1.32676423e+00 4.38649952e-02 -1.57096875e+00
-2.95742631e-01 -6.48787737e-01 -8.62507343e-01 -1.59825280e-01
5.72889864e-01 1.34754732e-01 -6.75512373e-01 1.34326589e+00
3.08386356e-01 5.14154077e-01 -7.06481457e-01 7.57678270e-01
6.90970421e-01 1.77413687e-01 -6.54841438e-02 1.29977494e-01
1.43676043e+00 -8.88703763e-01 -2.49296337e-01 -6.06453657e-01
6.28258049e-01 -5.72952330e-01 1.25477040e+00 3.91303778e-01
-7.68851399e-01 1.26993090e-01 -7.01548696e-01 2.51086146e-01
-8.96639228e-01 -5.25623620e-01 8.97959292e-01 9.28856850e-01
-1.08265972e+00 7.58628011e-01 -5.54020524e-01 -5.69537580e-01
7.01601923e-01 4.99950320e-01 -3.28061432e-01 8.36502984e-02
-1.23737633e+00 6.07469559e-01 -3.05655658e-01 -5.85273445e-01
-7.09708452e-01 -5.33195794e-01 -4.56235111e-01 1.95857942e-01
6.77533746e-01 -6.03051066e-01 1.21384287e+00 -9.38039720e-01
-8.56010795e-01 7.29355633e-01 1.97863176e-01 -5.67315102e-01
6.68357253e-01 5.89782298e-02 -8.16250503e-01 5.35399206e-02
2.41288409e-01 -4.65136230e-01 1.20773733e+00 -1.20277333e+00
-4.82910097e-01 -7.01246679e-01 -1.10693254e-01 -2.54434764e-01
-8.69723380e-01 -4.03364301e-02 -3.73313904e-01 -6.44184053e-01
1.72598660e-01 -1.09204435e+00 -1.78548247e-01 -2.89895356e-01
-4.78862017e-01 -2.48678803e-01 1.24623752e+00 -5.17545700e-01
1.51351202e+00 -1.78760219e+00 -3.00853401e-01 7.11922050e-01
8.84440839e-01 6.07920766e-01 -3.31958145e-01 8.01959455e-01
3.67855400e-01 6.06999278e-01 2.46106669e-01 -3.20595503e-02
7.43514970e-02 -2.75150001e-01 -2.29426920e-01 6.94636106e-01
-3.09255779e-01 1.09674489e+00 -1.37660408e+00 -1.54172271e-01
-1.05741419e-01 1.17924571e-01 -2.24190772e-01 2.38074232e-02
-1.95515841e-01 -3.25164050e-01 -9.36226130e-01 1.02116728e+00
6.66182101e-01 -7.42794394e-01 5.70903003e-01 2.93022264e-02
5.45119464e-01 3.58882248e-01 -8.94472361e-01 1.00385070e+00
-3.15862298e-01 3.55021209e-01 6.32222831e-01 -6.14369810e-01
7.55673110e-01 -2.46357977e-01 2.78808355e-01 -4.63045239e-01
4.00514871e-01 3.12290221e-01 1.63211510e-01 -1.60244435e-01
2.30687201e-01 4.23468888e-01 -2.88302720e-01 7.70313740e-01
-1.87087851e-03 8.26416314e-02 3.86379451e-01 1.11905003e+00
2.08103776e+00 -7.61566162e-01 2.26958424e-01 -5.30736029e-01
5.42513907e-01 -2.32491270e-01 -4.97146696e-02 1.16955340e+00
-4.73526210e-01 -3.24169099e-02 1.07371330e+00 -3.21092844e-01
-7.33561993e-01 -1.33113039e+00 1.86470494e-01 1.33316696e+00
4.20479715e-01 -7.87703633e-01 -5.83580494e-01 -1.46977377e+00
7.95847237e-01 3.74209315e-01 -4.47416872e-01 -4.79380786e-01
-4.03031826e-01 -8.76889944e-01 4.62171137e-01 -1.58268735e-01
4.36621726e-01 -8.74509692e-01 6.53849602e-01 1.96939036e-01
1.51372835e-01 -1.13826454e+00 -8.22227061e-01 -3.98678958e-01
-8.52751613e-01 -1.68827486e+00 1.30562529e-01 -8.44258964e-01
6.24595344e-01 9.14759219e-01 1.61426568e+00 6.04844093e-01
-3.16146791e-01 3.94288182e-01 -4.22655016e-01 -2.94472482e-02
-6.73046529e-01 4.09765542e-01 2.17669681e-01 2.91741379e-02
5.92815161e-01 -1.03169012e+00 -7.72966266e-01 7.02468097e-01
-8.41609478e-01 -6.63331211e-01 5.41806638e-01 6.25604153e-01
-3.39183718e-01 1.19192280e-01 1.06218410e+00 -1.48377526e+00
1.26081407e+00 -1.06965351e+00 -5.76557875e-01 1.39990062e-01
-8.79252136e-01 -4.63182092e-01 9.55186605e-01 -6.33985281e-01
-5.88298261e-01 -3.49691033e-01 3.07914287e-01 1.98166400e-01
2.28442550e-01 6.06408007e-02 -1.97849236e-02 -6.08233809e-01
9.54649746e-01 2.06273735e-01 4.55814213e-01 -4.88981336e-01
5.16229868e-01 1.07730508e+00 8.55367333e-02 -2.45835587e-01
1.04573405e+00 4.73966837e-01 9.64198075e-03 -7.95676351e-01
-7.07384706e-01 -1.01102114e+00 -2.18814686e-01 -6.85188696e-02
-1.97605416e-01 -6.96408510e-01 -1.15402269e+00 5.70461631e-01
-5.37736714e-01 1.18197333e-02 3.21214557e-01 -2.91249067e-01
7.17919692e-03 8.66890788e-01 -1.01296318e+00 -5.75094938e-01
-6.44496322e-01 -7.08259523e-01 9.37332392e-01 -1.42412573e-01
-2.30714053e-01 -1.26084709e+00 -2.95979708e-01 3.87234360e-01
5.58515072e-01 -7.88239166e-02 6.46544933e-01 -1.29155838e+00
-7.16308594e-01 -9.13360536e-01 -6.51464820e-01 2.54602075e-01
2.92163342e-01 -3.88122410e-01 -2.91029125e-01 -6.07627273e-01
-1.62733153e-01 -1.15382515e-01 8.09624970e-01 -1.89626634e-01
1.10996878e+00 -6.67237222e-01 -5.90909302e-01 1.90932840e-01
1.35827494e+00 -6.20332718e-01 3.18102539e-01 3.72834146e-01
7.69807398e-01 2.60028213e-01 1.84076279e-02 3.51994038e-01
1.95966229e-01 3.62810194e-01 8.88966858e-01 3.92966092e-01
-1.09726317e-01 -4.11342114e-01 4.14101034e-01 5.62624037e-01
4.84816670e-01 -4.42240745e-01 -7.88450778e-01 3.84020239e-01
-2.13153625e+00 -9.98633444e-01 -7.29703188e-01 2.16816640e+00
5.68357348e-01 4.03268307e-01 6.20572150e-01 -4.13097382e-01
8.72634411e-01 2.24552736e-01 -5.74481428e-01 -5.68583384e-02
-4.62299027e-02 -1.25400722e-01 8.54394913e-01 3.70221615e-01
-1.05532336e+00 1.22885144e+00 6.24809742e+00 9.83523846e-01
-6.69981599e-01 1.44916981e-01 2.09228516e-01 -4.53998744e-02
-2.71294534e-01 -8.41871500e-02 -8.22353303e-01 1.04626298e+00
1.10416543e+00 -3.74625742e-01 8.17606628e-01 9.90526855e-01
1.17635570e-01 7.00212121e-02 -5.95707774e-01 7.40031242e-01
2.60661364e-01 -1.42163372e+00 2.08054662e-01 1.95367724e-01
8.60035658e-01 4.16783780e-01 -3.62420902e-02 3.82818282e-01
1.07194316e+00 -7.40135252e-01 1.21828742e-01 8.91956165e-02
1.74351335e-01 -4.94473994e-01 6.13965809e-01 5.12819827e-01
-9.53748882e-01 -4.51821953e-01 -1.71697527e-01 1.34572983e-01
4.71463799e-02 8.42366695e-01 -8.06208611e-01 1.39679402e-01
7.00219452e-01 9.19062197e-01 -1.18271279e+00 9.63041425e-01
1.66589115e-02 8.41311395e-01 -4.29789841e-01 -5.43508172e-01
3.27894449e-01 -5.89109421e-01 8.31277907e-01 1.16247392e+00
1.25429615e-01 -4.36011612e-01 3.24925393e-01 5.08072197e-01
-7.87347734e-01 1.48967043e-01 -9.14271474e-01 -4.32659626e-01
6.01151526e-01 1.77295041e+00 -7.02193081e-01 -3.31974894e-01
-5.51065087e-01 9.47204113e-01 5.90669394e-01 2.67588466e-01
-4.92711186e-01 -1.70076653e-01 8.70792985e-01 1.01444781e+00
-1.54914245e-01 -1.16597831e-01 -6.06850013e-02 -1.36005569e+00
-1.01235807e-01 -1.29861522e+00 4.97535139e-01 -3.15740973e-01
-2.21174550e+00 4.16305006e-01 -6.86603844e-01 -1.10380018e+00
1.28495917e-01 -8.64151001e-01 -7.86877990e-01 6.35806024e-01
-1.04705584e+00 -1.59011436e+00 -3.97543788e-01 5.86258829e-01
1.76522046e-01 -6.81274712e-01 4.97165978e-01 1.74037561e-01
-4.20969337e-01 4.38621283e-01 3.88333410e-01 2.17841983e-01
7.13820219e-01 -1.39521825e+00 9.18784797e-01 5.26185632e-01
-2.90053431e-04 9.14169610e-01 5.20375192e-01 -1.20633900e+00
-1.71728766e+00 -1.24970567e+00 5.88933408e-01 -8.75640631e-01
1.75352108e+00 -7.40604758e-01 -8.23124230e-01 7.97296345e-01
-2.92398393e-01 1.39122233e-01 5.22310793e-01 4.57383275e-01
-7.10258961e-01 -2.42931899e-02 -1.22640038e+00 9.52568769e-01
1.61219907e+00 -5.21182358e-01 -1.29224390e-01 9.31689918e-01
3.60258698e-01 1.70041889e-01 -6.12836599e-01 6.69257641e-02
4.84589010e-01 -1.06523776e+00 1.08980310e+00 -1.07095039e+00
5.35170687e-03 1.49770677e-01 3.15994501e-01 -1.40148318e+00
-5.84977627e-01 -1.05336881e+00 -7.22952366e-01 9.15972054e-01
2.88974047e-01 -1.07211733e+00 1.11529589e+00 3.26632589e-01
2.38587558e-01 -4.46931124e-01 -5.08888304e-01 -9.95276213e-01
-1.40735909e-01 -1.00386500e-01 5.20077646e-01 8.11125219e-01
2.33758405e-01 5.06422639e-01 -2.63727218e-01 -2.44164504e-02
1.01131034e+00 1.32423207e-01 8.89093637e-01 -1.73273587e+00
-3.08086872e-02 -7.85965204e-01 -5.69040418e-01 -8.87255013e-01
3.07877272e-01 -1.22090590e+00 -8.03305268e-01 -1.32464802e+00
2.04169348e-01 -2.95998007e-01 1.29021019e-01 6.54231086e-02
-4.95882817e-02 3.53827685e-01 -1.65896565e-01 4.55376685e-01
-1.02989638e+00 7.06303120e-02 1.28060985e+00 -2.60335863e-01
-1.04566067e-01 4.37481850e-01 -9.84023154e-01 8.82978320e-01
8.63807380e-01 -4.46259648e-01 -2.98875928e-01 2.34202147e-01
5.67091525e-01 -4.41158116e-01 4.78679448e-01 -4.28832531e-01
2.22211316e-01 -1.69538967e-02 3.34425643e-02 -1.37149896e-02
1.31931201e-01 -7.88851380e-01 -2.10039258e-01 5.87317109e-01
6.34960681e-02 -1.61793176e-02 -3.54348838e-01 1.19794083e+00
1.64525643e-01 -2.21443817e-01 5.93623161e-01 -4.65708435e-01
-5.75164139e-01 6.36578918e-01 -2.48326764e-01 2.20991090e-01
9.49265897e-01 1.60046130e-01 -9.84443963e-01 -7.61546969e-01
-6.97073936e-01 3.31095785e-01 7.56770551e-01 8.55569839e-01
2.08750769e-01 -1.17025459e+00 -5.08531153e-01 2.48989120e-01
1.26176536e-01 -8.98955047e-01 -1.26964197e-01 5.46857238e-01
-5.25153935e-01 -8.27381685e-02 1.90263987e-01 -2.89564252e-01
-1.24278820e+00 3.79091024e-01 2.49469951e-01 -4.90406007e-01
-4.92507428e-01 3.77360255e-01 -3.47564012e-01 -5.76428115e-01
1.26579776e-01 5.14319718e-01 8.05431902e-02 3.51616889e-02
4.37817335e-01 7.45005429e-01 4.78777885e-02 -6.02165818e-01
-2.11830989e-01 -1.95340827e-01 -3.37044209e-01 5.65410316e-01
1.07651818e+00 -9.43602175e-02 -6.04224324e-01 -2.46783122e-02
1.22134578e+00 3.38031113e-01 -5.12270331e-01 -2.85009086e-01
1.79133475e-01 -9.29440200e-01 -6.64567202e-02 -8.72067809e-01
-9.28324699e-01 2.84952104e-01 1.28523156e-01 1.33713019e+00
3.35146427e-01 -1.05457939e-02 1.21232891e+00 4.77780461e-01
5.39144039e-01 -1.01875710e+00 5.49797833e-01 4.67483073e-01
6.51868701e-01 -1.56148279e+00 5.91415726e-02 -1.00201082e+00
-3.42891902e-01 7.11686611e-01 8.08027923e-01 -5.45664132e-01
1.06697249e+00 4.11943011e-02 -2.46363118e-01 -4.36177969e-01
-6.64993942e-01 -2.02486277e-01 3.54573458e-01 6.35001838e-01
1.98329702e-01 1.67442039e-01 -4.87303555e-01 6.45339608e-01
-5.27686663e-02 -5.00559926e-01 5.42281806e-01 5.24619460e-01
-7.12131739e-01 -1.20703983e+00 1.36381358e-01 1.33345532e+00
-5.39507806e-01 -9.51497555e-02 -9.62660313e-01 7.00136721e-01
-4.99491155e-01 1.14620364e+00 -2.58106917e-01 -4.09124076e-01
3.08767974e-01 -1.00641958e-01 -1.72674078e-02 -7.56017268e-01
-1.07378030e+00 -3.21502298e-01 3.96212399e-01 -7.82557607e-01
4.65853989e-01 -4.28115904e-01 -7.59341776e-01 -9.85715389e-01
-4.58284140e-01 1.40794545e-01 5.69442391e-01 3.42898101e-01
7.17509985e-01 2.45018564e-02 8.03856194e-01 -3.32885504e-01
-1.26227427e+00 -9.85697985e-01 -1.18627429e+00 1.18261695e+00
-5.93406186e-02 -3.68526816e-01 -9.06858087e-01 -5.78445554e-01] | [7.870263576507568, 10.04216480255127] |
9a4793c9-2ead-47aa-97cc-10ca2b7923e7 | accelerating-inexact-hypergradient-descent | 2307.00126 | null | https://arxiv.org/abs/2307.00126v1 | https://arxiv.org/pdf/2307.00126v1.pdf | Accelerating Inexact HyperGradient Descent for Bilevel Optimization | We present a method for solving general nonconvex-strongly-convex bilevel optimization problems. Our method -- the \emph{Restarted Accelerated HyperGradient Descent} (\texttt{RAHGD}) method -- finds an $\epsilon$-first-order stationary point of the objective with $\tilde{\mathcal{O}}(\kappa^{3.25}\epsilon^{-1.75})$ oracle complexity, where $\kappa$ is the condition number of the lower-level objective and $\epsilon$ is the desired accuracy. We also propose a perturbed variant of \texttt{RAHGD} for finding an $\big(\epsilon,\mathcal{O}(\kappa^{2.5}\sqrt{\epsilon}\,)\big)$-second-order stationary point within the same order of oracle complexity. Our results achieve the best-known theoretical guarantees for finding stationary points in bilevel optimization and also improve upon the existing upper complexity bound for finding second-order stationary points in nonconvex-strongly-concave minimax optimization problems, setting a new state-of-the-art benchmark. Empirical studies are conducted to validate the theoretical results in this paper. | ['Michael I. Jordan', 'Chris Junchi Li', 'Luo Luo', 'Haikuo Yang'] | 2023-06-30 | null | null | null | null | ['bilevel-optimization'] | ['methodology'] | [-1.70613244e-01 2.12335065e-01 -1.86669603e-01 -2.26711124e-01
-1.31444597e+00 -6.45841599e-01 -5.47863960e-01 6.69819638e-02
-5.83498955e-01 1.09801495e+00 -2.54829675e-01 -8.85229230e-01
-8.26013863e-01 -5.72279274e-01 -9.82125938e-01 -1.00072384e+00
-5.72352350e-01 3.93816710e-01 -2.70435542e-01 -4.84697461e-01
3.59771878e-01 2.12220345e-02 -1.10374582e+00 -2.49404877e-01
1.38439405e+00 1.42177570e+00 5.07622212e-02 5.31370759e-01
3.12631696e-01 2.06123665e-01 -3.58177543e-01 -5.01534343e-01
6.84414923e-01 -4.80163246e-01 -6.34193897e-01 -9.43864658e-02
3.99492115e-01 -9.45060104e-02 -6.28403351e-02 1.41457736e+00
3.41100544e-01 4.28418934e-01 3.29320192e-01 -9.67336059e-01
-4.40265745e-01 2.47484297e-01 -9.37357485e-01 3.03658247e-01
1.00153267e-01 2.16202766e-01 1.10606182e+00 -1.01383603e+00
2.36522272e-01 7.85922766e-01 8.76268744e-01 1.07382521e-01
-8.36449206e-01 -5.38932860e-01 3.52365583e-01 -2.82005399e-01
-1.60929012e+00 -5.99940121e-02 2.35216498e-01 -2.16199264e-01
6.52939439e-01 9.20989454e-01 6.29038692e-01 -1.16965674e-01
1.73160315e-01 3.95949244e-01 1.22200036e+00 -2.18012318e-01
1.47598252e-01 5.73629253e-02 5.51928207e-02 1.22572589e+00
6.04787230e-01 7.16426894e-02 -1.74575806e-01 -2.61741906e-01
6.55646265e-01 -2.46480763e-01 -4.51606482e-01 1.63556799e-01
-8.23416471e-01 1.10161996e+00 4.27137375e-01 2.97112688e-02
-2.94727206e-01 4.77144927e-01 -6.94246888e-02 2.18371987e-01
6.84094310e-01 6.54296875e-01 -6.45281553e-01 -5.10195911e-01
-1.00356007e+00 1.91057771e-01 9.45668936e-01 1.13499153e+00
5.81143022e-01 1.25921354e-01 -2.95839369e-01 6.40038610e-01
4.12820190e-01 8.77340138e-01 -2.69729167e-01 -1.17722082e+00
9.77130294e-01 5.87680757e-01 3.89283180e-01 -9.50032711e-01
-4.13256109e-01 -7.15739012e-01 -7.00566113e-01 7.87907615e-02
6.51972234e-01 -4.90272790e-01 -5.32906532e-01 1.32205915e+00
4.51339036e-01 -1.66787550e-01 -3.31596643e-01 1.13784909e+00
3.85067999e-01 8.12633157e-01 -4.74707693e-01 -8.61114979e-01
1.23893571e+00 -9.41168308e-01 -2.91719943e-01 -5.63791431e-02
6.57664180e-01 -9.57872570e-01 1.43274379e+00 3.34717482e-01
-1.63998353e+00 6.10072426e-02 -8.72842371e-01 2.31665090e-01
7.56122991e-02 1.22875087e-01 7.70668328e-01 8.15521896e-01
-8.47319663e-01 3.93999785e-01 -4.65048462e-01 6.79716468e-01
4.18849468e-01 7.80602038e-01 2.00738423e-02 -3.11603341e-02
-5.22816777e-01 3.34363401e-01 1.50116399e-01 5.16267419e-01
-7.58660913e-01 -1.08267117e+00 -6.95712268e-01 -2.16970649e-02
8.17076862e-01 -2.23784089e-01 8.56131196e-01 -3.73633713e-01
-1.35677707e+00 7.67420352e-01 -2.87831366e-01 -7.14998767e-02
5.78310668e-01 8.96802247e-02 1.40640959e-01 -8.77310485e-02
1.85983226e-01 6.88588023e-02 3.86579126e-01 -1.27177989e+00
-6.88237667e-01 -6.63837731e-01 3.34169298e-01 2.99822241e-01
-2.62772262e-01 -3.11063360e-02 -5.57692170e-01 -5.63634694e-01
2.54161000e-01 -1.18890965e+00 -4.79214370e-01 -3.31471294e-01
-4.52548742e-01 -9.40831378e-02 2.92843312e-01 -8.12115908e-01
1.71513569e+00 -1.90050364e+00 1.53905049e-01 8.25349152e-01
2.08751068e-01 3.26543987e-01 3.85159701e-01 -3.07079498e-02
3.11659247e-01 3.52722496e-01 -4.09195215e-01 -2.48757794e-01
1.06424280e-02 1.77543193e-01 1.93065092e-01 8.14740539e-01
-7.14095771e-01 7.42384136e-01 -9.93968010e-01 -5.64159080e-02
-5.08227460e-02 2.01724708e-01 -6.54497087e-01 -2.82333761e-01
2.04946343e-02 2.55583078e-01 -7.24717855e-01 7.54022241e-01
9.23413157e-01 -4.19232428e-01 5.16991690e-02 7.42956027e-02
-4.19490635e-01 -1.67205930e-01 -1.60322797e+00 1.32187724e+00
-4.70969081e-01 3.04748148e-01 5.50856113e-01 -1.18176603e+00
7.47193038e-01 -1.47297427e-01 9.18677509e-01 -7.81245470e-01
2.58440703e-01 6.14112258e-01 -3.58308464e-01 -4.91891176e-01
1.95839465e-01 -2.96334624e-01 -2.30034208e-03 1.95344195e-01
-5.76394200e-01 -1.74153715e-01 5.03456593e-01 -3.23092192e-01
8.42055678e-01 -3.43776524e-01 -2.47405823e-02 -9.58071768e-01
6.51687980e-01 6.24765344e-02 6.76507771e-01 7.80157864e-01
-7.86467940e-02 3.59368742e-01 9.67929959e-01 -4.97680545e-01
-1.09900057e+00 -8.39003086e-01 -3.35391432e-01 1.20598400e+00
4.80279058e-01 -3.18718463e-01 -6.93331659e-01 -6.05235755e-01
7.49376565e-02 8.19869459e-01 -6.16937339e-01 1.25161037e-01
-8.23057115e-01 -1.30811787e+00 4.04086709e-01 4.20700490e-01
4.45859671e-01 -4.06642109e-01 -5.51942408e-01 1.21111713e-01
-9.17613059e-02 -6.97463632e-01 -1.02010083e+00 3.17869663e-01
-8.73137176e-01 -1.02061641e+00 -7.21699297e-01 -7.87275851e-01
1.15524638e+00 -1.22739999e-02 1.11630237e+00 3.09034497e-01
-3.39362055e-01 2.60132968e-01 -4.71810475e-02 -4.56803054e-01
3.01262170e-01 -2.76224613e-01 9.95675940e-03 -1.66971236e-01
-9.12106633e-02 -1.57105669e-01 -8.33207130e-01 5.10028362e-01
-6.48247242e-01 -3.70708913e-01 2.10686654e-01 9.08366799e-01
1.35014474e+00 9.95506570e-02 2.84410477e-01 -5.71886539e-01
7.93290019e-01 -2.68827289e-01 -1.38859105e+00 2.57163852e-01
-9.33171809e-01 1.09315298e-01 9.18771327e-01 -2.71924287e-01
-5.66339493e-01 -2.09996492e-01 -1.79496258e-01 -4.94017422e-01
5.99804521e-01 7.25225806e-01 2.44385049e-01 -5.44706821e-01
4.68660295e-01 3.09797406e-01 -2.03261435e-01 -3.16161990e-01
1.97502628e-01 3.27921838e-01 4.46471184e-01 -9.33997750e-01
6.80366635e-01 6.18757844e-01 4.65605915e-01 -5.34222662e-01
-1.10391927e+00 -4.87665385e-01 1.36215240e-01 -2.02477816e-02
6.09943926e-01 -4.78629470e-01 -1.52906430e+00 -1.67544618e-01
-5.95628440e-01 -4.82863367e-01 -5.69906890e-01 5.48342288e-01
-6.30482793e-01 3.27240169e-01 -1.56352773e-01 -1.10701263e+00
-5.91534019e-01 -1.38675594e+00 8.60802710e-01 3.67819935e-01
3.67137313e-01 -1.13301766e+00 -2.41702855e-01 8.20599675e-01
4.00854409e-01 3.70710969e-01 6.34318769e-01 -1.35216238e-02
-4.61814255e-01 -2.04178140e-01 -2.67906904e-01 5.38040578e-01
-3.61172438e-01 -2.97772914e-01 5.74288033e-02 -6.54236972e-01
1.46554694e-01 -7.84255862e-02 4.75581884e-01 7.80943334e-01
1.49427688e+00 -1.14501810e+00 -2.16289207e-01 1.19963896e+00
1.68138289e+00 3.99586469e-01 6.11836791e-01 3.58028531e-01
2.73861706e-01 -2.26468697e-01 7.72711515e-01 9.27122831e-01
5.76033831e-01 5.45792401e-01 7.95818686e-01 -3.18780154e-01
3.57257426e-01 2.58481652e-01 6.62759766e-02 5.61796248e-01
-4.92596388e-01 -1.41264573e-01 -7.91618168e-01 7.71326005e-01
-1.63887227e+00 -6.28638685e-01 -5.04517257e-01 2.48549628e+00
9.93520141e-01 -1.52523601e-02 -6.45624772e-02 1.33798225e-03
4.40091848e-01 -1.57095283e-01 -4.53088850e-01 -7.49399006e-01
-2.84688324e-02 6.39507592e-01 9.30329144e-01 8.32174361e-01
-9.60074544e-01 6.64184511e-01 4.55681896e+00 1.10146415e+00
-9.83038664e-01 9.39294100e-02 7.98701882e-01 -4.79286790e-01
-3.44453067e-01 9.95325074e-02 -8.47369075e-01 8.47816885e-01
3.52208287e-01 -2.15131909e-01 8.79890740e-01 9.54270482e-01
3.91075790e-01 -3.74155879e-01 -6.58761084e-01 1.00676239e+00
2.52717137e-02 -1.41330242e+00 -6.33129895e-01 4.41679716e-01
1.37362885e+00 -2.87785262e-01 4.81031626e-01 5.10289408e-02
2.53238857e-01 -1.32773423e+00 4.76420432e-01 1.46364719e-01
1.08305812e+00 -9.66058373e-01 7.78181136e-01 3.78329188e-01
-1.42511463e+00 -3.99436951e-01 -2.14570701e-01 1.16292417e-01
2.94816256e-01 9.38183844e-01 -4.86142308e-01 6.26162052e-01
1.18831706e+00 -1.02259926e-01 1.80165678e-01 1.13079333e+00
-1.84139445e-01 4.27812129e-01 -8.97740960e-01 -2.90605456e-01
7.21328378e-01 -6.23077929e-01 7.03710675e-01 9.71240044e-01
4.82675940e-01 7.79814184e-01 4.48449165e-01 6.40410900e-01
-1.67493641e-01 4.33291584e-01 1.56327248e-01 3.01706344e-01
3.37962419e-01 9.49982345e-01 -6.44537926e-01 -2.04901546e-01
4.76171412e-02 4.70208883e-01 1.17019735e-01 3.51317227e-01
-1.24207342e+00 -6.90952718e-01 6.90714419e-01 1.38274848e-01
4.45204705e-01 -3.92551422e-01 -9.36306298e-01 -1.00757337e+00
6.34608865e-01 -7.51812637e-01 5.65074980e-01 -1.85003072e-01
-8.06050122e-01 3.07663828e-01 8.43675360e-02 -8.60881925e-01
3.16604197e-01 -5.45000792e-01 -4.11964566e-01 8.23646128e-01
-1.35871542e+00 -5.17663062e-01 -3.31323296e-01 7.91373372e-01
3.99828106e-01 -1.37969807e-01 3.80409449e-01 4.98735160e-01
-4.06361490e-01 8.99863839e-01 6.17157102e-01 -2.71910012e-01
-2.03782097e-01 -9.90139544e-01 -4.88391221e-01 6.60274446e-01
-4.34635967e-01 6.51249886e-01 7.19418705e-01 -4.59222615e-01
-1.78488779e+00 -7.05038726e-01 7.48083234e-01 -5.01094908e-02
4.67989236e-01 5.82279488e-02 -4.29008245e-01 5.92660189e-01
-1.25841752e-01 1.90577194e-01 5.38450480e-01 -1.56169057e-01
3.92105132e-01 -3.88879985e-01 -1.40976465e+00 5.90525031e-01
1.02053010e+00 1.19818181e-01 -2.51694210e-02 6.99548662e-01
3.60140741e-01 -9.84313071e-01 -1.34703875e+00 7.52683580e-01
3.88022542e-01 -8.43907297e-01 1.16641879e+00 -4.53097552e-01
8.82989690e-02 -2.02933893e-01 -2.97853023e-01 -9.55461800e-01
7.23162815e-02 -1.14971697e+00 -3.35775673e-01 3.66200119e-01
7.89627373e-01 -7.26620138e-01 9.63061094e-01 6.84181929e-01
-3.99639964e-01 -1.66210163e+00 -1.57885671e+00 -7.67623544e-01
3.17838341e-01 -2.30761915e-01 1.99120775e-01 8.18226993e-01
-6.39307201e-02 -3.11478913e-01 -5.53579628e-01 2.81608850e-01
7.63974786e-01 -5.69076613e-02 5.52635252e-01 -7.93210447e-01
-5.02990186e-01 -4.58131969e-01 5.13056666e-02 -1.22004366e+00
-4.25938219e-01 -7.58145690e-01 -1.91926919e-02 -1.54305792e+00
5.20002432e-02 -1.07377672e+00 -2.54140943e-01 5.77182472e-01
-1.84363276e-01 1.55329436e-01 1.58776507e-01 -1.42157882e-01
-6.33306801e-01 4.33031857e-01 1.53806520e+00 -9.09799039e-02
-4.99500692e-01 1.06488451e-01 -9.43821609e-01 7.53184140e-01
5.77110529e-01 -5.34172952e-01 -4.30253625e-01 -8.59563828e-01
7.56579161e-01 4.46911752e-01 9.16045234e-02 -5.26721299e-01
7.29338005e-02 -6.44865394e-01 4.89997352e-03 -4.17345911e-01
3.68152231e-01 -6.11467481e-01 3.71573167e-03 5.09374976e-01
4.86404262e-02 4.38935012e-01 1.16674118e-01 1.47279859e-01
2.01143324e-01 -4.16439295e-01 9.70570683e-01 -1.79708496e-01
-7.76057839e-02 2.43614793e-01 2.44542927e-01 4.65212554e-01
1.42172003e+00 -3.70116740e-01 -3.85499567e-01 -3.72857571e-01
-5.83008170e-01 6.33067012e-01 1.91841647e-02 -1.90307111e-01
5.76623797e-01 -9.23074067e-01 -7.66136169e-01 9.58324149e-02
-4.14184332e-01 4.25523072e-01 7.39613324e-02 1.49073064e+00
-9.46840048e-01 4.58779633e-01 5.32109201e-01 -5.17554462e-01
-8.65989745e-01 2.84147143e-01 5.62011957e-01 -6.08405590e-01
-1.54399246e-01 1.39411247e+00 -2.77593642e-01 -3.52654994e-01
1.73516929e-01 -4.11697835e-01 6.50257647e-01 -3.51275653e-01
4.22639586e-02 1.19761086e+00 1.12861924e-01 -3.23589206e-01
-4.52377856e-01 5.81756055e-01 3.42860758e-01 -8.77075791e-02
1.30833089e+00 -1.01123631e-01 -5.89730918e-01 -2.91969895e-01
1.52346945e+00 1.06964268e-01 -1.27990615e+00 3.49312931e-01
-4.56957757e-01 -7.88139641e-01 3.34837921e-02 -8.39289665e-01
-1.30835962e+00 4.04712498e-01 5.31424284e-01 1.38567626e-01
1.04311287e+00 -5.06916568e-02 8.20365727e-01 3.34223717e-01
4.37257141e-01 -1.68806815e+00 3.12467348e-02 6.43509328e-01
8.10838640e-01 -1.21581900e+00 4.52556282e-01 -4.36682433e-01
-5.23913383e-01 1.10889482e+00 5.26800394e-01 -2.05559701e-01
8.12848628e-01 1.91802025e-01 -3.05273741e-01 -2.82813340e-01
-3.68678011e-02 8.38077739e-02 5.58291018e-01 -2.99789459e-01
2.64290750e-01 2.42920458e-01 -1.04933858e+00 6.45900905e-01
-5.48148453e-01 -2.15162545e-01 1.86473131e-01 1.00521660e+00
-3.80575985e-01 -6.62478030e-01 -4.31039363e-01 6.25123441e-01
-6.21059537e-01 -2.71323115e-01 2.74327338e-01 7.27834344e-01
3.21152329e-01 1.10809469e+00 -2.81071514e-01 1.30697966e-01
3.34182680e-01 -4.34677452e-01 5.37603259e-01 -1.06060944e-01
-5.01811802e-01 2.76518434e-01 -1.29858861e-02 -5.34558654e-01
1.34019867e-01 -4.46986020e-01 -1.66703391e+00 -6.19772434e-01
-3.22713435e-01 4.93241429e-01 7.59298503e-01 9.24149096e-01
2.21598551e-01 1.50724724e-01 8.37655842e-01 -5.19873142e-01
-8.23087335e-01 -3.89379621e-01 -4.59451824e-01 1.30433738e-01
2.36046717e-01 -5.04034281e-01 -4.22075570e-01 -3.43525529e-01] | [6.5103559494018555, 4.535711288452148] |
12e95028-4250-45f5-a24e-0da5e11074e6 | an-accurate-car-counting-in-aerial-images | null | null | https://link.springer.com/article/10.1007%2Fs12652-021-03377-5 | https://link.springer.com/article/10.1007%2Fs12652-021-03377-5 | An Accurate Car Counting in Aerial Images Based on Convolutional Neural Networks | This paper proposes a simple and effective single-shot detector model to detect and
count cars in aerial images. The proposed model, called heatmap learner convolutional
neural network (HLCNN), is used to predict the heatmap of target car instances. In
order to learn the heatmap of the target cars, we have improved CNN architecture by
adding three convolutional layers as adaptation layers instead of fully connected
layers. The VGG-16 has been used as a backbone convolutional neural network in the
proposed model. The proposed method successfully determines the number of cars
and precisely detects the center of target cars. Experiments on the two different car
datasets (PUCPR+ and CARPK) show the state-of-the-art counting and localizing
performance of the proposed method in comparison with existing methods. Also,
experiments have been conducted to examine the effect of data augmentation and
batch normalization on the success of the proposed method. The code and data will be
made available here [https://www.github.com/ekilic/Heatmap-Learner-CNN-for-Object-Counting]. | ['Serkan Öztürk', 'Ersin Kılıç'] | 2021-07-13 | null | null | null | journal-of-ambient-intelligence-and-humanized-1 | ['object-counting'] | ['computer-vision'] | [-1.92940027e-01 -2.89844602e-01 1.02508068e-01 -3.19110900e-01
-3.27904135e-01 -3.46827894e-01 6.47723079e-01 -1.26879737e-01
-5.32690287e-01 3.09149772e-01 -3.27214479e-01 -1.18256196e-01
4.30077463e-01 -1.04872572e+00 -6.70517564e-01 -5.43839097e-01
1.69557557e-02 2.57372111e-01 7.95889080e-01 2.60360856e-02
2.30379120e-01 5.56641698e-01 -1.51532388e+00 2.28350997e-01
4.11561280e-01 1.04794300e+00 5.03824890e-01 9.93660569e-01
-2.29745716e-01 1.28067219e+00 -5.19414842e-01 -3.89941841e-01
3.61019880e-01 -1.00126095e-01 -6.98410809e-01 -2.78775170e-02
5.50865769e-01 -6.57209992e-01 -5.78294396e-01 1.14257014e+00
4.02173579e-01 2.94530764e-02 5.92858434e-01 -1.22322750e+00
-5.07176578e-01 2.90440708e-01 -8.90925825e-01 6.07862651e-01
-2.65543938e-01 4.63409200e-02 5.28707385e-01 -8.74521971e-01
2.44193763e-01 9.99209940e-01 7.39186883e-01 4.11724746e-01
-4.75904733e-01 -1.13613355e+00 -2.41382733e-01 4.51018035e-01
-1.72347677e+00 -1.83026955e-01 4.94877517e-01 -4.48307872e-01
8.13904464e-01 3.80585864e-02 6.43302083e-01 6.17147386e-01
7.75580779e-02 8.50183487e-01 7.04256833e-01 -5.45204580e-01
1.10678084e-01 1.59851849e-01 3.03290129e-01 1.06702328e+00
4.92677838e-01 6.46359921e-02 -3.16800661e-02 1.10001996e-01
7.48431563e-01 2.95805901e-01 3.11833531e-01 -1.63360700e-01
-9.24965262e-01 1.00191927e+00 1.01009190e+00 3.04707408e-01
-2.84431458e-01 4.77850318e-01 5.66398263e-01 -4.83508945e-01
5.57985902e-01 1.14002533e-01 -2.79844016e-01 2.31790468e-01
-1.03258848e+00 1.71027079e-01 4.55606461e-01 1.19587219e+00
6.25206649e-01 1.10228911e-01 -4.97203380e-01 6.52658284e-01
8.43989253e-02 6.39309525e-01 2.03706846e-01 -7.84872770e-01
4.03826654e-01 9.06360209e-01 1.96036294e-01 -9.62213099e-01
-3.81140262e-01 -1.48677498e-01 -7.87807703e-01 3.75865310e-01
2.83340156e-01 -3.96538764e-01 -1.23200560e+00 1.13762259e+00
3.69063497e-01 6.15579307e-01 -1.64634734e-01 1.02821589e+00
1.13557792e+00 8.73115599e-01 3.88355672e-01 4.17406023e-01
1.36109293e+00 -1.09712064e+00 -3.99051249e-01 -2.45705098e-01
6.72481179e-01 -6.35443747e-01 7.98968792e-01 -1.36064485e-01
-7.88844645e-01 -7.36793697e-01 -1.19406223e+00 -4.08541821e-02
-8.69040668e-01 9.31825042e-01 7.27463365e-01 6.26116574e-01
-6.16792917e-01 4.02503788e-01 -9.63731706e-01 -6.60314679e-01
9.56640542e-01 2.30210438e-01 -1.44110337e-01 -2.42515251e-01
-9.56695855e-01 8.49437475e-01 7.78280199e-01 1.39051497e-01
-1.18333447e+00 -3.97751242e-01 -6.42688990e-01 2.77994037e-01
2.25974649e-01 -3.62172127e-01 1.22141445e+00 -7.14448452e-01
-1.04551566e+00 8.93255949e-01 1.25142917e-01 -5.29390693e-01
4.01699185e-01 -8.36678073e-02 -2.65459746e-01 9.55614820e-02
2.14230701e-01 1.05667150e+00 5.54717243e-01 -9.71316755e-01
-1.18188870e+00 -2.39832476e-01 1.10060513e-01 -1.68691859e-01
-2.33294368e-01 2.94690639e-01 -4.56715196e-01 -2.51019239e-01
-4.50592577e-01 -8.86617064e-01 1.21602947e-02 -1.22491606e-01
-4.24942970e-01 -2.74781764e-01 1.19358885e+00 -4.29905981e-01
1.04007256e+00 -1.86048281e+00 -5.19979119e-01 -2.91099012e-01
1.02519870e-01 8.04989755e-01 -1.06711090e-01 5.64065874e-02
-7.38311484e-02 -1.80798411e-01 -2.75232911e-01 -2.79215515e-01
-3.79181832e-01 -4.15514261e-02 4.92059737e-02 5.94069421e-01
5.67671418e-01 8.20264280e-01 -7.35334337e-01 -6.39633775e-01
7.32182503e-01 6.85228288e-01 -4.85999882e-02 1.79153368e-01
1.98992323e-02 1.48162812e-01 -2.73654640e-01 8.02250445e-01
1.10521758e+00 -2.46474948e-02 -2.68204331e-01 -1.18035458e-01
-3.19114596e-01 -5.15183985e-01 -1.12059200e+00 1.21104610e+00
-3.47438693e-01 7.01036990e-01 -2.54911095e-01 -7.29586601e-01
1.08528996e+00 2.29658242e-02 1.88404068e-01 -5.91511905e-01
6.73854649e-01 -1.15839936e-01 -1.24746136e-01 -4.19791400e-01
6.32830858e-01 1.91041544e-01 -3.22854035e-02 -1.85240805e-01
4.01442498e-01 1.62686661e-01 5.46755075e-01 2.35748976e-01
9.84404683e-01 -8.75906274e-03 2.73625493e-01 -3.34195405e-01
7.33553052e-01 5.57878435e-01 3.97629470e-01 5.89638412e-01
-4.67032611e-01 5.97299635e-01 4.47644383e-01 -8.33732545e-01
-1.20814574e+00 -6.13205850e-01 -1.25923127e-01 1.21380973e+00
1.72561988e-01 -1.38061807e-01 -9.35993493e-01 -4.78019267e-01
-1.47480622e-01 8.25736403e-01 -7.89600313e-01 7.67983273e-02
-4.65565979e-01 -8.77635896e-01 7.92728424e-01 8.69951069e-01
1.07417059e+00 -1.12981009e+00 -9.96798098e-01 -7.79839829e-02
-3.08489595e-02 -1.38893163e+00 -7.00005144e-02 1.07079118e-01
-5.94712555e-01 -1.47342360e+00 -6.55288100e-01 -8.37921798e-01
6.34908855e-01 5.63175917e-01 9.28020716e-01 3.85464162e-01
-7.94034362e-01 -3.66405696e-02 -4.41556036e-01 -9.18454409e-01
-7.20533207e-02 2.93521941e-01 -3.81733656e-01 -1.61760971e-01
7.81395257e-01 1.31875798e-01 -6.96020961e-01 1.63315132e-01
-6.98613763e-01 -3.04608364e-02 5.64523280e-01 5.28301954e-01
3.48298043e-01 9.43929330e-02 2.29388759e-01 -1.07969451e+00
3.79946321e-01 -4.02737528e-01 -1.04791522e+00 2.00085938e-01
-1.95508435e-01 -3.20404530e-01 4.93760705e-01 -2.30727091e-01
-9.19051230e-01 6.91292107e-01 1.64451934e-02 -6.11448169e-01
-4.30575758e-01 -9.44354981e-02 2.04887688e-01 -8.20923895e-02
5.20746589e-01 3.39797847e-02 -3.79549563e-01 -1.52116448e-01
3.56608480e-01 7.31909096e-01 6.84007227e-01 -6.92922398e-02
6.27639234e-01 5.52622676e-01 2.20163707e-02 -7.82439113e-01
-8.41629207e-01 -8.77720714e-01 -1.02376187e+00 -4.97190982e-01
1.30021274e+00 -1.29810619e+00 -5.34656763e-01 6.75171494e-01
-1.30106723e+00 -3.71513575e-01 1.76363420e-02 3.62789333e-01
-2.19995126e-01 -1.45692199e-01 -6.16841733e-01 -9.91649568e-01
-7.13003576e-01 -9.70367312e-01 1.08695316e+00 5.69772661e-01
3.63507658e-01 -6.27671361e-01 -2.59678578e-03 1.88054755e-01
4.77973759e-01 4.58429188e-01 3.88655275e-01 -5.69474757e-01
-5.25208473e-01 -9.16600823e-01 -7.01359332e-01 4.30898070e-01
-3.41826767e-01 3.25884908e-01 -1.15148282e+00 -1.26260906e-01
-6.35310411e-01 -3.13689291e-01 1.08122921e+00 6.13389492e-01
1.21022010e+00 5.32150008e-02 -4.21851873e-01 4.93978858e-01
1.79648018e+00 2.05202703e-03 8.77417862e-01 2.65775561e-01
7.54537940e-01 1.55762032e-01 6.44601524e-01 4.70104963e-01
1.64317653e-01 2.48784706e-01 7.16850519e-01 -2.98980653e-01
-1.35611817e-01 -1.98091239e-01 -6.26965240e-02 5.48054099e-01
-4.03055310e-01 -1.79126307e-01 -1.09782302e+00 6.01264298e-01
-1.77475870e+00 -1.03796089e+00 -4.21517253e-01 1.87915397e+00
1.88797768e-02 1.60094023e-01 3.14734668e-01 2.16508508e-02
1.04226756e+00 1.00768596e-01 -2.09957585e-01 -2.91721135e-01
2.68072158e-01 1.33591250e-01 1.23037982e+00 3.06230158e-01
-1.59127533e+00 1.26863706e+00 5.55679083e+00 7.60317147e-01
-1.08554447e+00 4.48189855e-01 6.20564282e-01 3.24910991e-02
9.76786315e-01 -1.54152244e-01 -1.08093727e+00 5.13079286e-01
8.49451602e-01 1.37110978e-01 2.49276087e-01 1.27997470e+00
4.12701219e-02 -2.22227126e-01 -3.65301937e-01 8.93058062e-01
7.52635002e-02 -1.25298023e+00 -1.70426324e-01 -3.34514290e-01
6.43141150e-01 3.76406342e-01 -3.62458169e-01 6.82621658e-01
4.79558825e-01 -8.02860022e-01 6.30941808e-01 4.95766759e-01
9.42529917e-01 -1.06594706e+00 1.29655838e+00 3.88975501e-01
-1.59380770e+00 -3.72799039e-01 -1.10908663e+00 -1.35511801e-01
-1.26610741e-01 1.74280137e-01 -9.76144910e-01 1.08157657e-01
9.03154314e-01 4.61548597e-01 -1.15978754e+00 1.30099273e+00
-2.30674639e-01 6.08590603e-01 -1.52936488e-01 -2.18430579e-01
4.11787093e-01 1.11958645e-01 4.15192991e-02 1.66561925e+00
4.46028203e-01 2.93767482e-01 1.13043360e-01 1.00956285e+00
-2.02179715e-01 -9.29531381e-02 -6.65001214e-01 -1.69373769e-02
5.26630163e-01 1.89171660e+00 -1.19723535e+00 -5.45008540e-01
-3.85449171e-01 7.07867682e-01 3.05941731e-01 -1.16776399e-01
-1.14617896e+00 -7.97818661e-01 2.36520082e-01 2.77896702e-01
6.51379228e-01 -1.49071425e-01 -6.88105375e-02 -7.28926718e-01
-5.05298793e-01 -1.03327863e-01 3.71031374e-01 -9.12398338e-01
-9.41746175e-01 5.75085759e-01 2.73294747e-01 -1.11661172e+00
1.39832333e-01 -9.32701349e-01 -1.23652732e+00 6.20831907e-01
-1.30120277e+00 -1.63223422e+00 -9.90086317e-01 5.41343033e-01
7.18836784e-01 -3.53754669e-01 6.37952864e-01 3.37572575e-01
-7.25249410e-01 3.08348984e-01 -7.13270605e-02 6.89103842e-01
2.85322994e-01 -1.05671430e+00 5.51261008e-01 1.08732045e+00
-1.76018342e-01 1.57031324e-02 2.89524674e-01 -5.47510505e-01
-8.85289192e-01 -1.64873886e+00 5.29833436e-01 -5.23636043e-01
4.57768530e-01 -5.72569549e-01 -6.13759935e-01 6.32121325e-01
2.32233956e-01 4.65179026e-01 1.06954366e-01 -3.13404888e-01
-1.33609042e-01 -1.20406188e-01 -1.25466025e+00 -9.58677940e-03
5.77357292e-01 -1.09215938e-01 -1.37147024e-01 4.73754048e-01
6.02895856e-01 -3.15775067e-01 -4.54874456e-01 3.85508835e-01
4.90474612e-01 -8.29364181e-01 9.00125027e-01 -3.20666671e-01
6.22768223e-01 -4.54664081e-01 -1.35955960e-01 -8.80440891e-01
-7.14146078e-01 4.56338525e-01 -1.20116882e-01 1.18246067e+00
2.61371285e-01 -2.01165587e-01 8.56020510e-01 2.42064163e-01
-1.80404056e-02 -5.79414427e-01 -7.39472985e-01 -5.44278562e-01
1.45719320e-01 -1.71032444e-01 4.91285145e-01 7.84079850e-01
-4.77116615e-01 3.30092639e-01 -3.22059333e-01 3.48471165e-01
6.07189655e-01 -3.85173500e-01 8.66902232e-01 -1.03146601e+00
3.86832923e-01 -1.56639546e-01 -8.79786789e-01 -3.40751946e-01
-1.60142794e-01 -7.59022653e-01 4.79698516e-02 -1.59908390e+00
6.07893050e-01 -1.05616249e-01 -2.29893059e-01 4.60206598e-01
-1.67532951e-01 4.81949925e-01 5.63889861e-01 4.74752188e-02
-8.40078592e-01 1.54350117e-01 8.83634567e-01 -3.05330664e-01
3.40916999e-02 1.12313837e-01 -2.13449836e-01 6.89617813e-01
1.29649043e+00 -6.85787439e-01 -9.95754078e-02 -4.78810340e-01
-2.91161925e-01 -3.61981273e-01 7.05911398e-01 -1.62319362e+00
5.16062498e-01 2.68853098e-01 9.52716351e-01 -1.07447898e+00
2.40243345e-01 -8.65435004e-01 -1.40427262e-01 6.51757479e-01
-1.25142604e-01 1.82219848e-01 4.24216986e-01 5.37615299e-01
-2.15546396e-02 -4.65226769e-01 1.17592466e+00 -5.06197333e-01
-1.04487693e+00 3.12115461e-01 -1.46986663e-01 -2.96868980e-01
1.29751599e+00 -1.14384510e-01 -4.86446440e-01 1.13180257e-01
-3.05725873e-01 2.60090262e-01 1.85480565e-01 4.39302236e-01
5.22809088e-01 -1.34053302e+00 -8.32742870e-01 8.69575813e-02
3.23896050e-01 2.48900498e-03 1.86734214e-01 6.00094020e-01
-1.07236111e+00 6.43154860e-01 -5.05744636e-01 -5.36891639e-01
-1.23265707e+00 6.40409052e-01 4.31225091e-01 -9.87550989e-02
-5.34151614e-01 8.32729042e-01 -8.97814631e-02 -5.56037784e-01
2.19181255e-01 -3.44015092e-01 -3.57095480e-01 -1.18447341e-01
7.40454853e-01 7.15132833e-01 -2.87272111e-02 -6.79354250e-01
-4.87740040e-01 2.53201127e-01 -2.99063455e-02 4.27428722e-01
1.31592059e+00 1.19888581e-01 1.74243703e-01 2.84537673e-01
1.06836331e+00 -7.55755603e-01 -1.18034256e+00 2.75060665e-02
-1.72154993e-01 -4.90680575e-01 3.76935542e-01 -5.90891004e-01
-1.36329138e+00 1.01141047e+00 1.22587609e+00 -1.48345634e-01
8.79470944e-01 -1.20819844e-01 6.43928587e-01 3.26535553e-01
2.06411973e-01 -1.16273451e+00 -7.35585168e-02 5.48496664e-01
4.49299961e-01 -1.64205623e+00 8.63050967e-02 -1.63833290e-01
-7.42228270e-01 1.28402889e+00 1.01470470e+00 -6.50385857e-01
6.64069593e-01 4.54216719e-01 -1.49092972e-01 -4.21057463e-01
-4.03242588e-01 -6.10383153e-01 -1.12613022e-01 6.29364967e-01
4.25554961e-01 3.78507614e-01 1.70191806e-02 2.39689991e-01
1.49025619e-01 4.30300206e-01 5.53720713e-01 9.50302064e-01
-7.41751552e-01 -2.49001041e-01 -4.88615841e-01 5.89121521e-01
-4.01636034e-01 7.73758534e-03 -4.81428504e-01 1.05053091e+00
5.14339864e-01 9.07044291e-01 3.94362867e-01 -5.14273345e-01
4.37467843e-01 -5.86098656e-02 2.52626538e-01 -4.75100338e-01
-3.06635439e-01 -4.48949575e-01 -2.07134783e-01 -3.98292720e-01
-5.09853959e-01 -2.26137206e-01 -1.05147552e+00 -6.53603792e-01
-7.78169751e-01 -3.31773430e-01 7.91756868e-01 5.07737100e-01
3.27360444e-02 7.29033768e-01 3.77507448e-01 -1.04098272e+00
-2.19758689e-01 -1.41302824e+00 -7.05432177e-01 2.09223762e-01
-1.36560336e-01 -7.51265883e-01 -1.07716031e-01 1.50237888e-01] | [8.721612930297852, -0.22463208436965942] |
68d74fc3-31b8-4fac-95ee-4127f1cd82d1 | a-comprehensive-review-of-yolo-from-yolov1-to | 2304.00501 | null | https://arxiv.org/abs/2304.00501v3 | https://arxiv.org/pdf/2304.00501v3.pdf | A Comprehensive Review of YOLO: From YOLOv1 and Beyond | YOLO has become a central real-time object detection system for robotics, driverless cars, and video monitoring applications. We present a comprehensive analysis of YOLO's evolution, examining the innovations and contributions in each iteration from the original YOLO to YOLOv8 and YOLO-NAS. We start by describing the standard metrics and postprocessing; then, we discuss the major changes in network architecture and training tricks for each model. Finally, we summarize the essential lessons from YOLO's development and provide a perspective on its future, highlighting potential research directions to enhance real-time object detection systems. | ['Diana Cordova-Esparza', 'Juan Terven'] | 2023-04-02 | null | null | null | null | ['real-time-object-detection'] | ['computer-vision'] | [-3.15505594e-01 -3.10918599e-01 -4.87436056e-01 -1.49905950e-01
-1.32692814e-01 -1.75015941e-01 8.54742602e-02 -4.72692847e-01
-3.15673262e-01 3.38568509e-01 -5.60513258e-01 -3.58524740e-01
6.64657652e-02 -4.50428605e-01 -4.68626648e-01 -5.87060392e-01
-1.98953778e-01 -7.95225352e-02 6.40408576e-01 -1.92269355e-01
-1.62167922e-01 7.78259635e-01 -1.92025828e+00 -2.30120756e-02
2.79185086e-01 1.40676439e+00 1.59777299e-01 1.21928871e+00
2.52852559e-01 1.15927541e+00 -8.75110209e-01 -4.31725867e-02
3.96591008e-01 -8.60268176e-02 -1.18356489e-01 1.45827666e-01
9.96860325e-01 -3.42881352e-01 -9.44727838e-01 1.06295502e+00
5.79408944e-01 2.77369738e-01 3.58309239e-01 -1.93075097e+00
-7.10473478e-01 5.35323247e-02 -2.45984167e-01 8.06662083e-01
-3.32689464e-01 4.73023206e-01 7.59380341e-01 -1.18018317e+00
5.31193197e-01 1.16911769e+00 1.00826418e+00 6.49456441e-01
-5.87894440e-01 -7.63231874e-01 2.55358487e-01 7.70744383e-01
-1.20628238e+00 -7.19550490e-01 4.15036112e-01 -5.07922292e-01
1.26726878e+00 -1.32176891e-01 6.84015810e-01 5.87485671e-01
4.12085563e-01 1.22067058e+00 4.37969148e-01 -4.89079446e-01
-2.07877487e-01 1.37901053e-01 5.34444153e-01 9.41243589e-01
5.15875638e-01 5.46520352e-01 -5.72641790e-01 4.61686730e-01
4.88998294e-01 -2.96806902e-01 4.23629522e-01 -6.38883471e-01
-7.04694092e-01 8.80117118e-01 5.10922194e-01 -4.90267947e-02
4.02501486e-02 6.65393591e-01 6.32671237e-01 2.29723379e-01
9.62264612e-02 3.99775535e-01 -2.16249824e-01 4.78525162e-02
-6.95110083e-01 3.59717220e-01 4.31784689e-01 1.54966724e+00
5.30157864e-01 7.62144804e-01 3.52518484e-02 6.51650667e-01
3.10410470e-01 5.69447219e-01 2.34068036e-01 -1.20907712e+00
1.41240373e-01 5.18526956e-02 1.78084727e-02 -8.00772786e-01
-6.79902852e-01 -7.16932416e-01 -1.37709275e-01 5.23818076e-01
1.48536161e-01 -3.24784547e-01 -8.28759491e-01 9.76561189e-01
2.04541236e-01 1.39281094e-01 -1.51351139e-01 7.11702406e-01
1.58790863e+00 3.32301199e-01 1.66292056e-01 1.35357991e-01
1.48426807e+00 -1.68337321e+00 -7.91714072e-01 -5.76544166e-01
8.21770847e-01 -6.98205948e-01 7.12770998e-01 4.93695408e-01
-7.17966020e-01 -1.07522607e+00 -1.44565713e+00 -2.81170100e-01
-5.14501691e-01 6.73352420e-01 8.14732075e-01 1.10850477e+00
-8.47226977e-01 -6.75984919e-02 -7.82025397e-01 -4.26812679e-01
7.03773022e-01 2.19739005e-01 -3.37290199e-04 3.28825379e-04
-1.03388560e+00 1.12090802e+00 4.43673551e-01 3.13398004e-01
-1.07213593e+00 -5.65574288e-01 -9.87945557e-01 -7.16105103e-02
8.02651823e-01 -5.81391215e-01 1.81697679e+00 -3.23138416e-01
-1.39252424e+00 7.32921243e-01 -2.63552755e-01 -1.12504280e+00
3.46489519e-01 -5.53921580e-01 -7.89936304e-01 2.91699111e-01
1.35696590e-01 1.12174368e+00 7.80422747e-01 -8.98654342e-01
-1.39647257e+00 2.26377741e-01 6.76456168e-02 -3.97340693e-02
2.87050810e-02 2.28953630e-01 -7.55061924e-01 -3.15668643e-01
-3.70658576e-01 -8.65579545e-01 -2.04070240e-01 2.78262347e-01
-1.34150103e-01 -7.25994170e-01 1.13882446e+00 -4.89850044e-02
1.19648004e+00 -2.23921633e+00 -6.67529047e-01 -3.08959693e-01
8.82441938e-01 6.56974196e-01 -7.37955496e-02 1.85802236e-01
-4.18771096e-02 -4.85509336e-01 3.05249304e-01 -5.12823164e-01
-6.18899800e-02 2.79511116e-03 -2.07814693e-01 6.66550815e-01
3.64112169e-01 1.34174061e+00 -9.16990995e-01 -5.65366149e-01
7.90230691e-01 2.23837048e-01 -1.97739199e-01 -1.14859693e-01
4.53671850e-02 -1.80193558e-01 -1.85716152e-01 1.13844717e+00
3.61214817e-01 3.77856791e-02 -5.94053686e-01 -1.78600103e-01
-3.74224842e-01 3.52352977e-01 -7.72009790e-01 8.75783741e-01
-2.56287456e-01 1.85149050e+00 5.83258346e-02 -9.40487504e-01
8.42303634e-01 3.36482599e-02 3.75742614e-01 -7.91681528e-01
5.69749415e-01 -1.02949344e-01 2.48983707e-02 -6.44278586e-01
9.23067033e-01 1.02082670e-01 1.43317431e-01 -1.20242015e-01
1.65181711e-01 -7.67690614e-02 5.98927438e-01 7.96983168e-02
7.74072111e-01 -1.35320693e-01 5.40541887e-01 2.08815575e-01
3.77430528e-01 3.59992355e-01 4.48617250e-01 1.31825531e+00
-1.16389322e+00 3.53422821e-01 3.79120559e-01 -4.12641883e-01
-8.94576669e-01 -9.24685776e-01 -2.85532653e-01 1.36984634e+00
2.49623179e-01 -3.26617390e-01 -5.62960088e-01 -6.87694550e-01
2.44883463e-01 5.46321571e-01 -6.26657248e-01 -2.03132167e-01
-6.31362081e-01 -7.65728593e-01 8.64937127e-01 9.83947396e-01
6.01463199e-01 -1.25981653e+00 -1.07938886e+00 1.36453182e-01
3.04049820e-01 -1.54990315e+00 -1.88359872e-01 4.14620072e-01
-7.88683832e-01 -1.10777056e+00 -1.15309298e-01 -9.45101440e-01
2.58021802e-01 9.32004273e-01 1.00069439e+00 -8.22234154e-02
-9.05578732e-01 5.63449025e-01 -2.50674516e-01 -9.25871372e-01
-8.80678967e-02 -2.87892390e-02 4.32834327e-01 -5.37440538e-01
8.37917686e-01 3.16015989e-01 -3.76415968e-01 3.03708732e-01
-5.06048501e-01 -4.91999179e-01 5.94965398e-01 6.98130190e-01
1.17496699e-01 7.02978075e-02 5.98018885e-01 -6.70042276e-01
3.98574591e-01 -2.37888947e-01 -1.06455612e+00 -1.89410314e-01
-7.87780523e-01 -5.71732104e-01 2.76188344e-01 -2.41021231e-01
-8.27554762e-01 1.33628875e-01 -1.87504277e-01 -5.21882713e-01
-2.32910365e-01 -4.73672189e-02 1.23282395e-01 -4.58958179e-01
7.55959570e-01 -1.10831544e-01 -4.30194475e-02 -3.41511399e-01
8.16298604e-01 6.42888963e-01 8.95020604e-01 4.88762232e-03
6.76968634e-01 8.19072962e-01 -1.91082776e-01 -1.28415585e+00
-9.49107289e-01 -9.26772177e-01 -7.32319474e-01 -4.46086347e-01
7.61659205e-01 -9.29937184e-01 -1.09365559e+00 7.08221138e-01
-9.98545110e-01 -3.16378772e-01 -5.84682047e-01 6.36420488e-01
-4.93998230e-01 1.76139548e-01 -5.44234991e-01 -6.32665873e-01
6.25203252e-02 -1.22884619e+00 7.29063630e-01 5.87164760e-01
2.53115624e-01 -8.02784204e-01 -9.46698487e-02 3.68331343e-01
2.42514402e-01 -3.56991976e-01 1.37247205e-01 -1.28943548e-01
-6.18435442e-01 -5.80820143e-01 -7.88566887e-01 6.84522688e-01
-9.92525071e-02 5.04559018e-02 -1.26551294e+00 -3.10188860e-01
4.82314788e-02 -2.81628817e-01 1.20416725e+00 7.32792199e-01
7.92211413e-01 3.52906615e-01 -7.08947361e-01 8.01007450e-01
8.94380391e-01 5.95423937e-01 2.51423627e-01 6.25833809e-01
7.98531175e-01 4.44424361e-01 9.36829448e-01 -1.37374088e-01
4.46986854e-01 5.52389026e-01 4.74481463e-01 -8.05122927e-02
-8.80197465e-01 1.64439790e-02 7.44950235e-01 7.39309371e-01
2.60814518e-01 -3.92749548e-01 -4.91738528e-01 6.04047775e-01
-1.68957376e+00 -9.34843957e-01 -1.70919597e-01 1.52784741e+00
-1.48690507e-01 4.57105219e-01 4.06993449e-01 -1.23994827e-01
6.47726953e-01 2.09991887e-01 -9.15911496e-01 -3.29855084e-01
-2.65543282e-01 -1.77000359e-01 1.08934367e+00 3.13238502e-01
-1.59889448e+00 1.58271790e+00 9.06756401e+00 7.95513272e-01
-1.08850253e+00 2.32065380e-01 3.06918085e-01 -3.97727668e-01
9.26290870e-01 -2.21866563e-01 -1.62632036e+00 -1.46539256e-01
8.19732666e-01 1.10653341e-02 -7.26913959e-02 1.60820031e+00
2.15434879e-01 -3.25229734e-01 -8.95038426e-01 1.00505269e+00
5.09033918e-01 -1.54725420e+00 -4.56509262e-01 -1.91215172e-01
5.89334905e-01 7.60712147e-01 1.93012148e-01 6.26122892e-01
4.12686795e-01 -9.88833368e-01 8.79495025e-01 9.32024941e-02
7.63483584e-01 -5.11983752e-01 6.57069147e-01 -7.67558962e-02
-1.50114465e+00 -3.65261137e-01 -7.51664698e-01 -3.00892293e-01
2.40491062e-01 3.01053971e-01 -9.51645017e-01 3.94158363e-02
9.39966857e-01 1.09542358e+00 -7.12933838e-01 1.32948267e+00
-2.80964404e-01 6.89875066e-01 -2.43204206e-01 -1.26043692e-01
4.67761666e-01 2.64090568e-01 8.06884170e-01 1.40647006e+00
-1.63749829e-01 -2.37623349e-01 1.48735031e-01 6.17063820e-01
4.52784598e-02 -6.57752991e-01 -6.58544958e-01 7.15329349e-02
6.55393362e-01 1.33160102e+00 -8.33323240e-01 -4.98100877e-01
-8.96417737e-01 2.95203775e-01 4.70716506e-02 3.08077037e-01
-1.04570425e+00 -8.54880869e-01 1.01226115e+00 -1.50058955e-01
6.48504674e-01 -8.09923768e-01 -3.52413386e-01 -6.44716680e-01
-1.78390250e-01 -5.15423656e-01 2.22994000e-01 -8.68204176e-01
-9.03955519e-01 5.19558549e-01 -8.59565474e-03 -1.45110488e+00
1.85400806e-02 -1.31183171e+00 -3.93722892e-01 9.53126103e-02
-1.91934192e+00 -8.72168660e-01 -3.59349847e-01 -1.50580734e-01
8.86080861e-01 -5.23959517e-01 -9.65850502e-02 6.27962708e-01
-9.80281353e-01 8.55915785e-01 6.25956953e-02 3.23291630e-01
7.06789792e-01 -8.72451901e-01 8.58311415e-01 1.09687638e+00
1.70892358e-01 3.21958899e-01 7.63335466e-01 -2.94960737e-01
-1.30999303e+00 -1.15208983e+00 4.71131086e-01 -8.72917533e-01
8.96774054e-01 -2.18933657e-01 -4.68372673e-01 1.05870867e+00
1.63999215e-01 3.47914934e-01 1.44114435e-01 -2.71421015e-01
-2.79231463e-02 -3.50605011e-01 -6.95481598e-01 5.26881397e-01
1.16689372e+00 -2.71204442e-01 -4.71199036e-01 1.89723849e-01
7.78631985e-01 -6.57899797e-01 -3.66686970e-01 1.67510286e-01
8.00318658e-01 -1.04482532e+00 1.06949079e+00 -7.42194057e-01
-2.13006109e-01 -6.34056151e-01 2.70987332e-01 -7.61134446e-01
-4.73054767e-01 -5.54849267e-01 -1.89309344e-01 6.85520291e-01
-3.08913980e-02 -8.10506999e-01 1.01342738e+00 -6.10805564e-02
-8.32489669e-01 -6.54657543e-01 -1.01965582e+00 -1.14703143e+00
-3.19775194e-01 -1.20054793e+00 1.45620316e-01 2.29934901e-01
-2.20515460e-01 3.51859063e-01 -5.10955513e-01 8.75374451e-02
4.78391945e-01 -2.68895477e-01 1.09025621e+00 -8.35916042e-01
8.49370956e-02 -7.09211171e-01 -8.33703101e-01 -1.77362466e+00
-2.29773581e-01 -6.13550961e-01 4.13027436e-01 -1.14510465e+00
-2.20233306e-01 -3.42655480e-01 -3.51793736e-01 3.41189981e-01
-9.74789076e-03 8.34294021e-01 4.60004508e-01 2.01200858e-01
-9.73296106e-01 4.15753454e-01 1.03967381e+00 -2.55082417e-02
-2.80586481e-01 1.86752260e-01 -6.33621633e-01 9.09237027e-01
5.12336671e-01 -5.68489969e-01 -1.74743354e-01 -3.23817104e-01
1.54948607e-01 -6.03748322e-01 5.62672555e-01 -1.48770928e+00
5.10733187e-01 6.30284101e-02 2.89268255e-01 -1.24318767e+00
4.89796847e-01 -5.41260242e-01 -6.92757130e-01 6.52498603e-01
2.87626032e-02 2.10010186e-01 4.73933369e-01 4.97281611e-01
-2.55279303e-01 -1.75574735e-01 1.13387978e+00 1.52680144e-01
-1.51666355e+00 2.40058377e-01 -1.02045226e+00 -1.57475173e-01
1.34247279e+00 -6.30713940e-01 -7.28980899e-01 -1.13095455e-01
-7.89340794e-01 7.76209295e-01 -7.16719627e-02 8.93335342e-01
7.06257582e-01 -1.24085104e+00 -4.21398759e-01 2.61332959e-01
4.37106788e-01 -6.94188252e-02 3.32982875e-02 9.11736012e-01
-7.93346763e-01 1.22451115e+00 -2.42748678e-01 -7.58140147e-01
-1.36651123e+00 5.81798792e-01 6.25703692e-01 1.79077640e-01
-7.91496277e-01 1.12197387e+00 5.34437478e-01 -6.41599372e-02
3.99545550e-01 -5.13859153e-01 -4.16922539e-01 -9.41014886e-02
4.44094211e-01 9.55068886e-01 7.83424079e-02 -6.70003355e-01
-4.63768750e-01 5.23197532e-01 4.59727198e-02 4.39054668e-01
9.49867487e-01 -5.11966407e-01 2.97004789e-01 6.37091458e-01
6.94774330e-01 -2.86752015e-01 -1.48328519e+00 -1.64038405e-01
1.49317458e-02 -1.39040696e-02 2.31912553e-01 -1.86303079e-01
-1.09349263e+00 1.12749875e+00 7.81461060e-01 5.16141243e-02
7.39710867e-01 2.67283052e-01 9.39100385e-01 8.21363389e-01
9.69482213e-02 -1.17409956e+00 5.56825250e-02 7.51052320e-01
4.31745797e-01 -1.30703878e+00 1.52582556e-01 -6.00971758e-01
-5.06573975e-01 1.20655406e+00 9.56079543e-01 -3.81221265e-01
7.47462988e-01 2.36112699e-01 4.53992277e-01 -4.86498922e-01
-7.20784068e-01 -6.11290574e-01 3.83084208e-01 1.00659382e+00
2.04940125e-01 -3.31037790e-02 2.84790099e-01 1.80173114e-01
-3.78728211e-01 6.89542294e-02 4.21536237e-01 7.90719151e-01
-8.82214606e-01 -5.70010006e-01 -5.48256278e-01 5.38907647e-01
1.69271246e-01 1.36401097e-03 -2.60437936e-01 1.21467149e+00
5.26633620e-01 1.01660514e+00 3.18235964e-01 -3.82228553e-01
3.40422451e-01 -1.24119602e-01 3.63243371e-01 -8.49918365e-01
-9.23933685e-02 -5.31387851e-02 4.71898355e-02 -5.88674247e-01
-1.89082220e-01 -5.66185594e-01 -1.03797877e+00 -1.67691082e-01
-7.65184939e-01 -2.72966266e-01 5.06466091e-01 1.02275491e+00
4.36087757e-01 9.18103635e-01 2.03068852e-01 -9.11997914e-01
-1.18099920e-01 -8.55115056e-01 -4.06377971e-01 -6.53452277e-01
8.73756647e-01 -1.12830079e+00 -1.74157843e-01 1.83033615e-01] | [8.263554573059082, -0.9097298979759216] |
be9e5555-dc98-4899-b44c-ee88186e4900 | sgram-improving-scene-graph-parsing-via | 2210.08675 | null | https://arxiv.org/abs/2210.08675v1 | https://arxiv.org/pdf/2210.08675v1.pdf | SGRAM: Improving Scene Graph Parsing via Abstract Meaning Representation | Scene graph is structured semantic representation that can be modeled as a form of graph from images and texts. Image-based scene graph generation research has been actively conducted until recently, whereas text-based scene graph generation research has not. In this paper, we focus on the problem of scene graph parsing from textual description of a visual scene. The core idea is to use abstract meaning representation (AMR) instead of the dependency parsing mainly used in previous studies. AMR is a graph-based semantic formalism of natural language which abstracts concepts of words in a sentence contrary to the dependency parsing which considers dependency relationships on all words in a sentence. To this end, we design a simple yet effective two-stage scene graph parsing framework utilizing abstract meaning representation, SGRAM (Scene GRaph parsing via Abstract Meaning representation): 1) transforming a textual description of an image into an AMR graph (Text-to-AMR) and 2) encoding the AMR graph into a Transformer-based language model to generate a scene graph (AMR-to-SG). Experimental results show the scene graphs generated by our framework outperforms the dependency parsing-based model by 11.61\% and the previous state-of-the-art model using a pre-trained Transformer language model by 3.78\%. Furthermore, we apply SGRAM to image retrieval task which is one of downstream tasks for scene graph, and confirm the effectiveness of scene graphs generated by our framework. | ['Byoung-Tak Zhang', 'Yu-Jung Heo', 'Woo Suk Choi'] | 2022-10-17 | null | null | null | null | ['scene-graph-generation', 'dependency-parsing'] | ['computer-vision', 'natural-language-processing'] | [ 8.52510393e-01 3.94787341e-01 2.24628061e-01 -6.43361092e-01
-5.89970648e-01 -3.83997083e-01 8.13138664e-01 1.15342043e-01
-7.92391449e-02 2.95270443e-01 4.26924139e-01 -5.02123952e-01
3.81354928e-01 -1.35695481e+00 -8.31149995e-01 -4.18033242e-01
4.83533382e-01 3.09366345e-01 4.66588974e-01 -4.00037050e-01
2.71958232e-01 2.85972178e-01 -1.34821308e+00 5.23659527e-01
3.02431256e-01 6.19085312e-01 8.79189134e-01 7.38975048e-01
-7.86609054e-01 1.21246397e+00 -7.60152102e-01 -3.77743155e-01
-4.46047820e-02 -8.48836362e-01 -9.29682910e-01 5.21370947e-01
2.93909550e-01 -2.36394763e-01 -3.77119452e-01 1.21365631e+00
1.03913032e-01 2.07037672e-01 6.53276503e-01 -1.12410927e+00
-1.03191853e+00 6.12867892e-01 -5.36285818e-01 -9.67174172e-02
5.98264277e-01 -2.69322932e-01 9.31021750e-01 -6.91625774e-01
8.68503332e-01 1.83597779e+00 -2.76591443e-02 6.00703657e-01
-7.98999488e-01 -3.72504860e-01 4.62891251e-01 2.73937266e-03
-1.23945308e+00 2.70740744e-02 8.19299400e-01 -2.99544036e-01
1.33725727e+00 3.69130634e-02 6.68634593e-01 6.87891006e-01
4.79970038e-01 5.61561346e-01 1.06206596e+00 -6.26291215e-01
1.29182726e-01 -1.02290951e-01 2.60573924e-01 1.02969480e+00
2.52260834e-01 -5.13235748e-01 -3.44683439e-01 3.16171944e-01
1.06181633e+00 -1.56349353e-02 -1.74396187e-02 2.07837317e-02
-8.10592175e-01 1.09872413e+00 5.69641471e-01 1.16492145e-01
-2.26969302e-01 5.22323787e-01 3.52587461e-01 1.53814077e-01
5.16624272e-01 -7.96112642e-02 1.15497112e-01 4.51635242e-01
-4.39255685e-01 1.53630286e-01 5.93547642e-01 1.44859922e+00
7.93181837e-01 3.74949336e-01 -3.30130160e-01 7.06037998e-01
7.30861604e-01 7.47155428e-01 4.54305679e-01 -4.87008393e-01
6.02187395e-01 9.70800281e-01 -5.55204928e-01 -1.05913746e+00
7.70070916e-03 1.49909006e-02 -6.40155077e-01 -2.07993314e-01
-1.51559621e-01 1.86606586e-01 -1.49177754e+00 1.44506919e+00
1.75361499e-01 -7.16715083e-02 5.58870792e-01 8.70605469e-01
1.65679371e+00 1.15175378e+00 5.20689905e-01 5.46626449e-02
1.79713261e+00 -1.17125821e+00 -7.45759189e-01 -6.11964762e-01
8.59804630e-01 -7.75480747e-01 1.17196965e+00 -2.62597427e-02
-8.94320548e-01 -6.51998222e-01 -9.06675816e-01 -4.24637735e-01
-6.46133244e-01 -1.54525444e-01 7.32245624e-01 3.99175972e-01
-1.36577153e+00 -1.41762361e-01 -2.96581656e-01 -7.44327486e-01
2.21550032e-01 -1.24101220e-02 -5.67500174e-01 -6.43342197e-01
-9.83138859e-01 4.84488398e-01 8.28702807e-01 -3.66564058e-02
-9.81662869e-01 -6.97265938e-02 -1.48720467e+00 9.43904892e-02
4.58296210e-01 -1.03660846e+00 9.80977476e-01 -8.26786816e-01
-1.15118825e+00 1.20121169e+00 -3.38316053e-01 -4.20561433e-01
-2.65174299e-01 -3.04369181e-02 -2.19717681e-01 5.21366477e-01
4.48025644e-01 8.58749866e-01 8.06022525e-01 -1.36398995e+00
-3.33312422e-01 -3.82581860e-01 5.10435224e-01 5.47053397e-01
2.39031985e-01 2.96795070e-01 -8.59858871e-01 -6.92034841e-01
2.29403853e-01 -7.76142061e-01 -3.90589982e-01 -6.03620231e-01
-4.56718862e-01 -2.85043210e-01 9.70228553e-01 -7.58848965e-01
1.03368270e+00 -1.88207483e+00 -1.23794451e-01 -9.32974219e-02
-9.21431556e-03 8.55164677e-02 -4.41027373e-01 8.18294168e-01
-2.59798557e-01 3.72578830e-01 -3.84062797e-01 -3.30782235e-01
-1.57197356e-01 5.30443966e-01 -7.26766706e-01 -1.64590795e-02
2.95973927e-01 1.29385519e+00 -9.30612564e-01 -7.71591723e-01
6.54549122e-01 3.30415100e-01 -5.00016451e-01 4.25482512e-01
-5.73770404e-01 2.00754493e-01 -8.59655678e-01 4.85888630e-01
6.29458487e-01 -3.82271469e-01 2.34967649e-01 -2.69939750e-01
2.41957113e-01 1.22680195e-01 -6.88259602e-01 1.94346380e+00
-4.85019654e-01 5.58020055e-01 -5.52602112e-01 -1.12212968e+00
1.30256486e+00 7.35905245e-02 -2.93182284e-02 -8.88112605e-01
1.44552380e-01 -2.97588170e-01 -3.92602891e-01 -4.98712331e-01
6.75529659e-01 -4.38048601e-01 -5.66773534e-01 2.44905144e-01
6.72118291e-02 -7.68020332e-01 2.60783941e-01 7.30703712e-01
1.01941168e+00 4.01506633e-01 4.21673983e-01 -1.65769421e-02
7.09566355e-01 3.94008547e-01 3.28464806e-02 8.11932802e-01
3.20876181e-01 8.30029428e-01 4.82944608e-01 -2.84904182e-01
-7.78141737e-01 -1.09912777e+00 3.52964967e-01 7.68417656e-01
4.31640446e-01 -8.45328093e-01 -9.52026367e-01 -6.79271162e-01
-5.45797348e-01 1.08390653e+00 -3.36900562e-01 -2.68007107e-02
-5.93340576e-01 -5.63260257e-01 2.80839324e-01 4.83639479e-01
9.80189323e-01 -1.53909278e+00 -5.75602412e-01 9.94916856e-02
-2.13838488e-01 -1.81868899e+00 -2.50455052e-01 -2.62131423e-01
-7.16350555e-01 -9.11599696e-01 -3.96262854e-01 -1.21136510e+00
1.00367212e+00 5.23244739e-01 1.32105339e+00 2.09439814e-01
-4.45493221e-01 8.27812850e-01 -9.23437238e-01 -4.08592314e-01
-5.83919227e-01 -5.58012009e-01 -7.14941502e-01 -9.55161918e-03
2.68400639e-01 -1.99978217e-01 -3.69214207e-01 -1.25585988e-01
-1.43581843e+00 7.52290606e-01 6.14683926e-01 4.55619723e-01
8.93685400e-01 2.05631673e-01 1.95191458e-01 -1.22369134e+00
5.98699331e-01 -2.19032779e-01 -5.12161791e-01 4.14567083e-01
-1.79050192e-01 1.77622795e-01 6.71118319e-01 -1.56444963e-02
-1.34131753e+00 5.70938457e-03 -1.85696617e-01 -2.99888819e-01
-3.19136798e-01 5.48521161e-01 -3.01474363e-01 2.70005047e-01
1.73635259e-01 6.21425271e-01 -3.58421981e-01 -2.23471783e-02
8.23792875e-01 5.57193995e-01 5.08960664e-01 -5.15994847e-01
7.06309855e-01 5.34598827e-01 3.34143400e-01 -1.16214347e+00
-9.59813356e-01 -4.58575100e-01 -4.40298587e-01 -2.47656628e-01
1.51165771e+00 -1.16198230e+00 -1.45449489e-02 2.74986416e-01
-1.47674501e+00 -4.49645162e-01 -2.97967911e-01 6.83718622e-02
-6.26141906e-01 5.76363027e-01 -4.27304983e-01 -6.83548272e-01
-5.30889928e-01 -1.00583398e+00 1.79793298e+00 2.37501338e-01
1.90189615e-01 -1.16307425e+00 -3.93063307e-01 6.26542807e-01
-5.64106070e-02 6.47127271e-01 1.11810434e+00 -3.98880184e-01
-8.31956506e-01 -1.08814761e-01 -6.39607430e-01 1.34215683e-01
1.45939082e-01 -3.89824629e-01 -7.84950078e-01 -1.15037307e-01
9.06497017e-02 -1.58936739e-01 7.24833131e-01 4.72655118e-01
9.81167257e-01 -1.61466345e-01 -3.16098362e-01 4.38341200e-01
1.91588676e+00 4.43812102e-01 1.01266360e+00 1.90656394e-01
1.08197868e+00 6.43206894e-01 7.06410110e-01 -1.40634589e-02
6.96287513e-01 5.69670916e-01 3.87967676e-01 -3.59965056e-01
-7.32756853e-01 -7.77173281e-01 4.23361599e-01 9.44950044e-01
2.33405754e-01 -8.03218186e-01 -7.43862867e-01 5.00205100e-01
-1.94125426e+00 -7.03867376e-01 -6.73417151e-01 1.71974957e+00
1.12745322e-01 9.27055851e-02 -4.71012205e-01 -3.42443615e-01
7.92772710e-01 3.87027562e-01 -1.02580763e-01 -7.90099084e-01
-2.18372658e-01 2.54611939e-01 3.58760089e-01 5.16434431e-01
-7.24877417e-01 1.79418671e+00 5.45674229e+00 7.81793177e-01
-8.17852497e-01 6.52956814e-02 5.02557874e-01 7.01089680e-01
-4.42031026e-01 5.51263869e-01 -7.17779875e-01 -7.92245641e-02
5.63301265e-01 -2.28082210e-01 1.43419415e-01 8.30016077e-01
2.37106949e-01 -2.63433814e-01 -8.81766498e-01 1.28336513e+00
6.73074424e-01 -1.23703051e+00 1.20621955e+00 -1.42123327e-01
4.80407089e-01 -3.39503288e-01 -4.38584775e-01 3.74387830e-01
4.18370962e-01 -1.10466623e+00 7.76571214e-01 2.45450675e-01
8.63964260e-01 -3.54895234e-01 6.81774855e-01 9.54165831e-02
-1.89298379e+00 2.53244162e-01 -7.30933249e-01 -1.10743918e-01
7.01096952e-01 4.03392792e-01 -9.81477797e-01 1.07446158e+00
4.59097505e-01 7.33897924e-01 -9.13897753e-01 2.82674879e-01
-7.79617310e-01 5.24189651e-01 1.67259619e-01 -5.33560552e-02
4.58420426e-01 -4.67971742e-01 3.69518906e-01 1.36445701e+00
1.61416039e-01 4.09792334e-01 3.02685946e-01 1.10374987e+00
6.07639626e-02 3.74411434e-01 -1.29535103e+00 -2.51932532e-01
2.74054892e-02 1.32166278e+00 -1.28153443e+00 -7.01275289e-01
-5.72347641e-01 1.41578078e+00 8.75866562e-02 4.94128525e-01
-7.85496652e-01 -3.99186492e-01 -1.46011442e-01 1.67457134e-01
2.27611825e-01 -2.71652699e-01 1.33365035e-01 -1.05895662e+00
-3.19034681e-02 -4.27828133e-01 5.40419519e-01 -1.71777821e+00
-9.51389670e-01 8.43801916e-01 5.70824862e-01 -9.25905406e-01
-3.92053992e-01 -6.43895566e-01 -6.82521522e-01 8.14635515e-01
-1.58068585e+00 -1.72344387e+00 -5.53437710e-01 7.61621237e-01
1.19866717e+00 -7.46173561e-02 8.28852534e-01 -1.82337567e-01
-2.40357071e-01 -5.43549992e-02 -8.13645422e-01 3.26946765e-01
1.26297936e-01 -1.12652719e+00 8.10340941e-01 1.20760047e+00
4.15527493e-01 5.24180949e-01 4.59339589e-01 -9.61784184e-01
-1.58641875e+00 -1.59026170e+00 9.19007778e-01 -3.23806047e-01
4.02898788e-01 -6.54115975e-01 -5.90060771e-01 8.51007581e-01
2.93622255e-01 -1.62489414e-01 3.27716261e-01 -7.35276401e-01
-3.27468157e-01 2.39897743e-01 -9.23462331e-01 9.29458022e-01
1.37456083e+00 -5.23624599e-01 -8.04797351e-01 3.82505208e-01
1.28936672e+00 -3.16837907e-01 -5.39855421e-01 4.45908517e-01
5.61952591e-02 -7.82741785e-01 1.03305459e+00 -3.99729580e-01
5.65259755e-01 -4.26912397e-01 -6.82865798e-01 -8.29249799e-01
-1.40086681e-01 -2.11319700e-01 5.29736459e-01 1.33744061e+00
3.98483733e-03 -6.04594707e-01 5.56970596e-01 3.36193830e-01
-3.68885964e-01 -3.10226738e-01 -3.79131019e-01 -6.28274083e-01
-1.73242375e-01 -6.05395615e-01 4.57423717e-01 6.04088962e-01
-4.17099267e-01 1.04507649e+00 8.18974078e-02 2.22252980e-01
6.77129567e-01 4.15579349e-01 1.02698195e+00 -7.87343860e-01
-1.68925509e-01 -7.05816299e-02 -6.45715117e-01 -1.22895515e+00
4.51073468e-01 -1.19800878e+00 4.37177271e-02 -2.52501082e+00
4.09825891e-01 5.17111532e-02 2.60140210e-01 4.88048941e-01
-1.51009098e-01 2.03890875e-01 5.49213946e-01 -9.07377750e-02
-7.21814334e-01 5.50413489e-01 1.46981418e+00 -2.58202314e-01
2.27862541e-02 -7.51211047e-01 -7.92728841e-01 7.94344246e-01
6.08040690e-01 -4.36804891e-01 -9.62091565e-01 -4.41499144e-01
1.03642479e-01 3.31569284e-01 5.34675002e-01 -6.86431825e-01
-1.08321561e-02 -1.93553686e-01 3.58394235e-02 -6.26227617e-01
3.94985050e-01 -5.73530793e-01 3.06496590e-01 3.53866339e-01
3.95660773e-02 3.25735778e-01 1.70694619e-01 6.33060753e-01
-3.95626187e-01 -4.17049766e-01 3.77570778e-01 -6.50971174e-01
-1.35807979e+00 1.26451969e-01 -3.68236840e-01 1.40766561e-01
8.61256063e-01 -4.86098468e-01 -4.67440337e-01 -6.59051001e-01
-5.03176808e-01 -4.32942137e-02 3.07483554e-01 6.64265096e-01
1.17540693e+00 -1.20758259e+00 -6.62221730e-01 1.06869705e-01
4.02365744e-01 3.61596018e-01 2.39644423e-01 3.21576148e-02
-7.75361001e-01 5.28374493e-01 6.30619302e-02 -7.29464114e-01
-1.54885554e+00 5.25194407e-01 -4.75085936e-02 -3.90602201e-01
-1.04617131e+00 4.75524753e-01 1.16977632e+00 -1.96110025e-01
-3.28761011e-01 -5.04510880e-01 -4.26139265e-01 -5.03605187e-01
2.13843450e-01 -3.43339652e-01 -3.07352453e-01 -1.17744064e+00
-1.78901002e-01 1.09857571e+00 2.69852936e-01 -2.07028553e-01
1.16898191e+00 -1.99830249e-01 -3.42129439e-01 2.45817795e-01
1.24766421e+00 -3.37272018e-01 -5.88996410e-01 8.90458003e-03
-2.68714219e-01 -2.97413141e-01 -5.70948981e-02 -4.26348388e-01
-9.30275142e-01 8.52466345e-01 3.38517755e-01 2.34196171e-01
1.26803696e+00 4.88733768e-01 7.87247717e-01 2.39697829e-01
4.76731479e-01 -5.50916493e-01 4.91900295e-01 5.11021197e-01
1.21804059e+00 -9.67390180e-01 5.13012297e-02 -1.05742121e+00
-8.10599267e-01 1.10418677e+00 5.05000949e-01 -2.67928779e-01
4.23122615e-01 -7.01679140e-02 -1.19133085e-01 -7.81591892e-01
-4.81701344e-01 -6.63937807e-01 3.51861656e-01 7.97884285e-01
2.81029493e-01 1.37475014e-01 -2.40577519e-01 4.49808955e-01
-3.55334729e-01 -2.21732214e-01 5.73075354e-01 9.80911136e-01
-4.19178665e-01 -1.12431693e+00 -2.46932566e-01 2.43788630e-01
-3.56767625e-01 -5.14742076e-01 -5.68537712e-01 8.91090214e-01
-1.61345631e-01 1.28352451e+00 7.40605220e-02 -2.70046145e-01
4.38837469e-01 -3.66946869e-02 6.16612673e-01 -1.26354480e+00
-1.73585743e-01 6.73103482e-02 2.60726213e-01 -4.63144481e-01
-6.23677790e-01 -2.30625495e-02 -1.76306939e+00 2.73239225e-01
-4.65818383e-02 -1.08376227e-01 7.31824696e-01 9.95184660e-01
5.84356673e-02 7.19537139e-01 2.97528833e-01 -5.96566617e-01
3.56309742e-01 -7.24788785e-01 -4.60767537e-01 9.32227314e-01
-1.97115719e-01 -3.00861001e-01 -8.70518163e-02 5.72305322e-01] | [10.49283218383789, 1.577022671699524] |
dd4e1e8b-78c8-486e-9ebf-b89641e16316 | alexa-teacher-model-pretraining-and | 2206.07808 | null | https://arxiv.org/abs/2206.07808v1 | https://arxiv.org/pdf/2206.07808v1.pdf | Alexa Teacher Model: Pretraining and Distilling Multi-Billion-Parameter Encoders for Natural Language Understanding Systems | We present results from a large-scale experiment on pretraining encoders with non-embedding parameter counts ranging from 700M to 9.3B, their subsequent distillation into smaller models ranging from 17M-170M parameters, and their application to the Natural Language Understanding (NLU) component of a virtual assistant system. Though we train using 70% spoken-form data, our teacher models perform comparably to XLM-R and mT5 when evaluated on the written-form Cross-lingual Natural Language Inference (XNLI) corpus. We perform a second stage of pretraining on our teacher models using in-domain data from our system, improving error rates by 3.86% relative for intent classification and 7.01% relative for slot filling. We find that even a 170M-parameter model distilled from our Stage 2 teacher model has 2.88% better intent classification and 7.69% better slot filling error rates when compared to the 2.3B-parameter teacher trained only on public data (Stage 1), emphasizing the importance of in-domain data for pretraining. When evaluated offline using labeled NLU data, our 17M-parameter Stage 2 distilled model outperforms both XLM-R Base (85M params) and DistillBERT (42M params) by 4.23% to 6.14%, respectively. Finally, we present results from a full virtual assistant experimentation platform, where we find that models trained using our pretraining and distillation pipeline outperform models distilled from 85M-parameter teachers by 3.74%-4.91% on an automatic measurement of full-system user dissatisfaction. | ['Prem Natarajan', 'Gokhan Tur', 'Shuai Zheng', 'Haiyang Yu', 'Pan Wei', 'Fabian Triefenbach', 'Liz Tan', 'Mukund Harakere Sridhar', 'Saleh Soltan', 'Anjali Shenoy', 'Andy Rosenbaum', 'Stephen Rawls', 'Chandana Satya Prakash', 'Charith Peris', 'Enrico Palumbo', 'Gokmen Oz', 'Karolina Owczarzak', 'Pradeep Natarajan', 'Alessandro Manzotti', 'Jianhua Lu', 'Beiye Liu', 'Haidar Khan', 'Kevin Martin Jose', 'Jonathan Hueser', 'Wael Hamza', 'Dilek Hakkani-Tur', 'Thomas Gueudre', 'Karthik Gopalakrishnan', 'Turan Gojayev', 'Satyam Dwivedi', 'Anurag Dwarakanath', 'Luoxin Chen', 'Amit Chauhan', 'Rakesh Chada', 'Jin Cao', 'Claudio Delli Bovi', 'Abhishek Bhagia', 'Davide Bernardi', 'Konstantine Arkoudas', 'Shankar Ananthakrishnan', 'Jack FitzGerald'] | 2022-06-15 | null | null | null | null | ['cross-lingual-natural-language-inference', 'xlm-r', 'slot-filling'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 5.85867427e-02 9.28379118e-01 -4.84033048e-01 -7.27073491e-01
-1.10825193e+00 -5.92612803e-01 5.79929948e-01 -3.73432562e-02
-9.78932202e-01 7.65156806e-01 3.23178411e-01 -1.00168943e+00
2.84394845e-02 -4.10301626e-01 -6.70578241e-01 9.18373913e-02
1.62482336e-01 1.21084225e+00 -2.15308711e-01 -6.11690521e-01
-1.55100927e-01 -8.86515602e-02 -1.26343918e+00 3.20838630e-01
9.67929363e-01 7.22495914e-01 1.78988934e-01 8.20695400e-01
-9.92643535e-02 7.14896441e-01 -8.34169447e-01 -1.89180642e-01
5.31034060e-02 6.86705038e-02 -1.17515194e+00 -1.24589205e-01
7.85687625e-01 -9.93214071e-01 -2.24445015e-01 5.40953815e-01
6.31274819e-01 1.28130063e-01 5.93203783e-01 -1.22618365e+00
-6.72410488e-01 9.71516490e-01 -4.64301221e-02 -1.39270917e-01
3.53943974e-01 1.98550105e-01 1.09928048e+00 -8.19567323e-01
5.08225501e-01 1.41542017e+00 5.89729011e-01 7.63643146e-01
-1.47285044e+00 -7.85497308e-01 -1.00796007e-01 -9.08448175e-02
-1.33230925e+00 -7.37407029e-01 -1.33399367e-01 -1.10424697e-01
1.83016396e+00 3.46748121e-02 1.11597136e-01 1.12894225e+00
-4.72135469e-02 1.11976600e+00 1.02297103e+00 -5.87720990e-01
-4.06039469e-02 7.50422895e-01 4.58274871e-01 9.46013510e-01
-4.65307496e-02 -4.99136895e-02 -4.41488117e-01 -1.70868516e-01
5.61980009e-01 -4.77121502e-01 -1.43474951e-01 2.22519457e-01
-7.58678019e-01 1.09285903e+00 2.38471061e-01 1.38016298e-01
6.73697442e-02 5.29721193e-02 4.07579690e-01 6.67039931e-01
4.32911396e-01 8.16843450e-01 -1.01208425e+00 -7.17354000e-01
-8.36475611e-01 1.89989150e-01 1.07679868e+00 1.33531296e+00
7.47005284e-01 1.28140852e-01 -1.11737981e-01 1.12161314e+00
4.97544825e-01 5.79339027e-01 8.46474290e-01 -1.03343105e+00
4.22331572e-01 4.92816955e-01 -1.76795404e-02 -1.97794586e-01
-6.36581123e-01 -3.20924491e-01 -2.05207959e-01 -6.46760613e-02
4.72131610e-01 -4.31796759e-01 -1.21096396e+00 1.92332447e+00
-9.75710601e-02 -1.58080056e-01 3.71456981e-01 5.52757025e-01
1.00730896e+00 9.38215852e-01 3.18582922e-01 9.84353051e-02
1.56576836e+00 -9.78245556e-01 -7.35849619e-01 -5.61710596e-01
1.37164176e+00 -7.74066389e-01 1.49643528e+00 4.99223381e-01
-1.05446577e+00 -7.52433419e-01 -8.93076360e-01 -5.51328599e-01
-4.54082280e-01 3.47743541e-01 6.20954156e-01 7.70755291e-01
-1.46826363e+00 1.79687202e-01 -5.53436279e-01 -7.40314901e-01
8.96595344e-02 5.81129491e-01 -2.51500785e-01 -1.84414774e-01
-1.29329300e+00 1.32365692e+00 5.08363724e-01 -3.48606586e-01
-9.72046196e-01 -9.00945902e-01 -1.06621063e+00 1.70493484e-01
3.77749950e-01 -3.41625541e-01 1.96149206e+00 -1.97665393e-01
-1.54162562e+00 7.60793090e-01 -1.16834417e-01 -6.59021735e-01
2.01112613e-01 -5.56456566e-01 -2.11290717e-01 -1.87966436e-01
1.13051645e-01 1.12187397e+00 2.99476802e-01 -9.23065364e-01
-7.83414364e-01 -7.93637782e-02 2.98540652e-01 4.91088241e-01
-6.08950198e-01 -1.91894755e-01 -2.91725487e-01 -1.03329100e-01
-1.65969849e-01 -9.29783642e-01 -8.93815514e-03 -3.87708038e-01
-2.54344523e-01 -5.98930895e-01 8.97268891e-01 -9.22688901e-01
1.34704864e+00 -2.02187896e+00 -1.78781003e-01 7.19651347e-03
2.76584295e-03 5.30636728e-01 -4.80749846e-01 3.80927563e-01
-1.53469220e-02 1.94349334e-01 6.89189434e-02 -8.22316766e-01
3.71117741e-01 5.26901543e-01 -3.33304852e-01 -1.19551912e-01
2.74463028e-01 9.32848990e-01 -8.17898095e-01 -5.14761090e-01
4.18318510e-01 3.20207328e-01 -1.10734856e+00 4.80338156e-01
-2.16001108e-01 -5.39475158e-02 4.31987643e-02 5.26361585e-01
2.77159423e-01 -2.31294826e-01 3.12565267e-01 -1.77685842e-01
1.22228540e-01 9.78214622e-01 -6.77545965e-01 2.02120018e+00
-1.05917263e+00 7.92459369e-01 1.67306051e-01 -7.77641833e-01
7.62826622e-01 5.43508470e-01 2.57382214e-01 -9.60121393e-01
8.50939080e-02 2.58761346e-01 1.62801951e-01 -2.15331256e-01
1.12782192e+00 -3.45717706e-02 -5.12162268e-01 5.60422838e-01
7.13744938e-01 -3.34082276e-01 6.71797385e-03 5.94039500e-01
1.16159534e+00 -1.28543645e-01 -7.09386822e-03 -4.66564447e-01
1.94523960e-01 4.03541587e-02 1.17356159e-01 9.03796732e-01
-2.91257322e-01 2.07905844e-02 3.22482169e-01 -1.81912884e-01
-9.48075473e-01 -8.93915296e-01 -4.46840107e-01 1.57595241e+00
-2.23757446e-01 -8.83213460e-01 -7.39149988e-01 -6.31658077e-01
2.02937916e-01 1.33937633e+00 -1.20076381e-01 -3.58602017e-01
-4.53273416e-01 -4.72443253e-01 8.70489657e-01 5.82562745e-01
4.24265772e-01 -9.49620306e-01 -3.44080478e-01 3.67806852e-01
-2.77995884e-01 -1.28341091e+00 -2.50574738e-01 6.80408895e-01
-7.14877427e-01 -5.79734683e-01 -1.03910349e-01 -7.67701030e-01
4.55028057e-01 -2.13223055e-01 1.35865486e+00 1.95003182e-01
-1.94903135e-01 5.52425861e-01 -2.25595787e-01 -2.65315533e-01
-7.45385051e-01 6.07141256e-01 5.29469490e-01 -9.79035437e-01
9.05958295e-01 -2.69499511e-01 -8.99369866e-02 3.12525779e-01
-6.50692165e-01 2.43583024e-01 7.32728958e-01 1.26349568e+00
-8.50434825e-02 -9.77793261e-02 5.01968980e-01 -1.00799191e+00
7.49884605e-01 -3.05577725e-01 -4.66588646e-01 1.87010720e-01
-1.12081921e+00 3.38904709e-01 5.56047440e-01 -3.18721145e-01
-1.18003953e+00 -2.88455188e-01 -5.13088882e-01 -7.11356923e-02
-3.05574983e-01 3.74335855e-01 -2.52199657e-02 5.20726800e-01
8.48848879e-01 -4.00926583e-02 1.64001375e-01 -4.54554707e-01
7.41017520e-01 1.23817253e+00 5.45424759e-01 -8.77909958e-01
6.05057776e-01 -2.40519628e-01 -1.03362644e+00 -9.24295962e-01
-8.20519328e-01 -4.77084190e-01 -3.81087840e-01 3.26235205e-01
7.70821333e-01 -1.08658516e+00 -8.83977771e-01 6.06643558e-02
-1.02115297e+00 -9.97843325e-01 -2.83673286e-01 2.76409298e-01
-5.99601686e-01 1.96573734e-02 -9.67390954e-01 -9.16697085e-01
-4.42704558e-01 -1.40080106e+00 1.17705119e+00 8.80812630e-02
-7.23644197e-01 -1.02768600e+00 -2.39215478e-01 7.50367224e-01
5.65639615e-01 -9.20314908e-01 1.08167756e+00 -1.07385147e+00
-5.34895733e-02 -2.31707445e-03 -2.97908634e-01 4.87971902e-01
1.06974706e-01 -4.08244818e-01 -1.10335195e+00 -5.42979181e-01
-3.66358012e-01 -1.04342937e+00 4.23007220e-01 1.66075811e-01
7.21957803e-01 -3.43465745e-01 -2.71847516e-01 3.69888365e-01
1.09660673e+00 1.57164618e-01 3.60175699e-01 3.49729896e-01
4.19162571e-01 4.05198693e-01 8.11776876e-01 3.59872162e-01
5.77058434e-01 6.82796240e-01 1.31380841e-01 -8.72845054e-02
8.39678757e-03 -6.56258702e-01 7.31806159e-01 8.22085798e-01
4.25892591e-01 -2.23776340e-01 -9.33664918e-01 4.21255320e-01
-1.61045873e+00 -3.39443505e-01 3.43836427e-01 1.93994617e+00
1.25991881e+00 4.45000350e-01 -3.80668230e-02 -2.90482253e-01
4.40466739e-02 -1.20137326e-01 -5.11521339e-01 -9.46599782e-01
4.59970862e-01 5.33305645e-01 6.36905313e-01 1.11366844e+00
-8.42098057e-01 1.51643527e+00 6.49429655e+00 8.19182873e-01
-7.47262359e-01 4.46694531e-02 5.83149076e-01 -1.00854054e-01
-3.03908080e-01 -8.31390321e-02 -1.43580043e+00 3.13380778e-01
1.80907333e+00 7.47002512e-02 5.90960026e-01 1.15316546e+00
-1.26380563e-01 -2.71571368e-01 -1.40972018e+00 8.78154635e-01
-6.21580929e-02 -1.01198685e+00 -5.22760786e-02 1.39993414e-01
5.26040614e-01 2.71527976e-01 1.59036994e-01 1.43131983e+00
8.22780371e-01 -1.32547712e+00 3.93199474e-01 -1.18204698e-01
1.05352509e+00 -7.05658734e-01 7.65075803e-01 6.92576587e-01
-7.81729400e-01 -8.45369548e-02 -4.50661689e-01 -9.18832943e-02
1.62626132e-01 -2.04439573e-02 -1.71764815e+00 3.83200683e-02
6.26110673e-01 2.50506133e-01 -5.67644715e-01 -7.51617318e-03
-1.17614552e-01 7.73254931e-01 -6.54418111e-01 -2.72738915e-02
3.58879149e-01 1.28578404e-02 -9.29886252e-02 1.27375555e+00
-1.24710672e-01 2.06502587e-01 2.40873247e-01 7.54019618e-01
-2.17782587e-01 -1.64505169e-01 -5.29405117e-01 -3.28992575e-01
6.98951721e-01 1.17367077e+00 7.57034719e-02 -7.81210363e-01
-4.52004045e-01 1.11469960e+00 5.55415034e-01 3.13304603e-01
-8.01328599e-01 -2.71649301e-01 8.48593891e-01 -9.35883448e-02
-1.89097375e-01 -3.29012781e-01 -3.34930280e-03 -1.06697536e+00
-4.56604093e-01 -1.29890597e+00 3.07327002e-01 -8.04415107e-01
-9.63561893e-01 6.58131778e-01 1.96781471e-01 -5.83768964e-01
-9.69590843e-01 -1.03630257e+00 -1.45685270e-01 1.13167238e+00
-1.42078030e+00 -9.74162579e-01 4.20996472e-02 2.25492060e-01
7.04683185e-01 -4.11112338e-01 1.47936237e+00 3.96817088e-01
-5.45385897e-01 1.10655344e+00 -8.06677062e-03 1.35034099e-01
1.00525582e+00 -1.44138420e+00 6.24144852e-01 2.71360725e-01
1.03657097e-01 1.05105686e+00 6.25181556e-01 -3.77541691e-01
-1.31690013e+00 -8.57821405e-01 1.15093207e+00 -9.39697921e-01
6.56780958e-01 -5.23076952e-01 -7.52566516e-01 1.21167409e+00
4.39943135e-01 -4.92189437e-01 8.85531902e-01 7.05851316e-01
-2.43361384e-01 2.61136711e-01 -1.26097953e+00 6.36370718e-01
9.09145296e-01 -8.56922746e-01 -8.04832220e-01 3.67219537e-01
1.02995348e+00 -5.49136221e-01 -1.19681144e+00 2.97325164e-01
5.77591777e-01 -5.88258624e-01 7.23958910e-01 -8.64922345e-01
3.07250828e-01 2.72811502e-01 -2.97256231e-01 -1.35427523e+00
-9.10810828e-02 -5.14465332e-01 -1.17537409e-01 1.05681884e+00
7.66908467e-01 -6.55201018e-01 8.49524975e-01 9.04110312e-01
-3.85993481e-01 -8.16656172e-01 -6.17509842e-01 -4.41391051e-01
2.85318792e-01 -6.42354012e-01 1.39596149e-01 9.15595472e-01
3.44397336e-01 7.83106565e-01 -1.25273511e-01 -3.37721631e-02
2.08988070e-01 -3.95990670e-01 1.02720630e+00 -8.94214869e-01
-4.62333530e-01 -1.98120713e-01 3.46856005e-02 -1.70495772e+00
4.66007680e-01 -9.50353086e-01 1.69959337e-01 -1.47432101e+00
3.86782102e-02 -7.07944155e-01 -1.92648489e-02 9.44147229e-01
-8.69238302e-02 4.86208014e-02 -5.18165827e-02 -1.83330372e-01
-2.71472186e-01 5.29801846e-01 7.17081189e-01 -2.14340255e-01
-2.59755403e-01 -4.79953855e-01 -7.86190569e-01 4.78269756e-01
7.17260718e-01 -2.73954514e-02 -8.40423048e-01 -6.12994492e-01
-2.68859625e-01 4.52649482e-02 -6.55637309e-02 -7.24397182e-01
-2.17471924e-02 1.78376213e-01 2.01427087e-01 -5.08412004e-01
8.35779727e-01 -6.98609829e-01 -6.47045135e-01 4.55987751e-01
-4.99487400e-01 -3.21733430e-02 7.85244703e-01 -2.37694308e-01
1.77606061e-01 -3.80312771e-01 4.73754793e-01 3.26860473e-02
-1.03515577e+00 -3.54845673e-02 -6.22812867e-01 1.11533917e-01
3.66044790e-01 7.78464228e-02 -4.72390473e-01 -7.63512850e-01
-6.69273555e-01 5.51314592e-01 1.48235649e-01 6.42715573e-01
3.78495365e-01 -1.10354531e+00 -4.26652670e-01 5.86310625e-01
2.00334549e-01 1.45309448e-01 -1.91277727e-01 4.61715281e-01
-2.70053089e-01 9.45083022e-01 8.15595593e-03 -7.49033749e-01
-1.13483810e+00 1.78223327e-02 2.85561264e-01 -2.93441236e-01
-3.60089183e-01 1.14029777e+00 -7.75139555e-02 -1.22265422e+00
5.48723698e-01 -6.97531581e-01 2.25891992e-01 -1.25815779e-01
3.66260141e-01 7.01957494e-02 8.27829689e-02 -3.68522644e-01
-1.39816135e-01 -1.07036524e-01 -5.57005405e-01 -5.26199520e-01
1.09793425e+00 1.36979565e-01 3.00976038e-01 3.60400975e-01
1.45219421e+00 -2.03552902e-01 -8.38686049e-01 -3.45704466e-01
6.87157735e-02 8.30490589e-02 1.51174784e-01 -1.09882987e+00
-3.74386936e-01 7.55187511e-01 6.40245020e-01 -1.67191803e-01
6.92484140e-01 -8.46426748e-03 8.15256000e-01 9.08427000e-01
5.07061183e-01 -1.34198773e+00 1.66992415e-02 8.39600682e-01
5.86634576e-01 -1.48035574e+00 -1.58110604e-01 5.35558462e-02
-5.47725797e-01 7.00790346e-01 1.12967801e+00 3.46768200e-01
4.36954021e-01 4.55173701e-01 5.18182337e-01 -3.31939943e-02
-9.73436475e-01 -2.90915202e-02 2.70594247e-02 3.93190831e-01
1.03005517e+00 2.37138554e-01 -7.12648109e-02 5.16902626e-01
-8.96758497e-01 -2.87743419e-01 3.43251497e-01 8.94935429e-01
-5.64750731e-01 -1.36244321e+00 -1.20046973e-01 5.65464795e-01
1.21006496e-01 -4.08211499e-01 -6.50897697e-02 9.83976662e-01
-1.59243807e-01 1.04149508e+00 3.31706613e-01 -5.94898403e-01
3.54721278e-01 6.61718369e-01 3.39995891e-01 -1.03959274e+00
-5.75841427e-01 -1.07383579e-01 5.79219639e-01 -6.04237378e-01
2.00485274e-01 -3.59403551e-01 -1.57417798e+00 -4.73557651e-01
-5.79114497e-01 4.28226471e-01 8.34977627e-01 8.97428334e-01
3.37428451e-01 6.28543854e-01 2.01925755e-01 -3.99878532e-01
-1.10633695e+00 -1.59682834e+00 -2.82546133e-01 1.62897840e-01
2.03341514e-01 -5.48933864e-01 -2.76004523e-01 -2.34071091e-01] | [11.0014066696167, 8.360823631286621] |
8abad33f-69ad-4098-82e0-72e0626ae18c | mesh-interest-point-detection-based-on | 1604.08806 | null | http://arxiv.org/abs/1604.08806v3 | http://arxiv.org/pdf/1604.08806v3.pdf | Mesh Interest Point Detection Based on Geometric Measures and Sparse Refinement | Three dimensional (3D) interest point detection plays a fundamental role in
3D computer vision and graphics. In this paper, we introduce a new method for
detecting mesh interest points based on geometric measures and sparse
refinement (GMSR). The key point of our approach is to calculate the 3D
interest point response function using two intuitive and effective geometric
properties of the local surface on a 3D mesh model, namely Euclidean distances
between the neighborhood vertices to the tangent plane of a vertex and the
angles of normal vectors of them. The response function is defined in
multi-scale space and can be utilized to effectively distinguish 3D interest
points from edges and flat areas. Those points with local maximal 3D interest
point response value are selected as the candidates of 3D interest points.
Finally, we utilize an $\ell_0$ norm based optimization method to refine the
candidates of 3D interest points by constraining its quality and quantity.
Numerical experiments demonstrate that our proposed GMSR based 3D interest
point detector outperforms current several state-of-the-art methods for
different kinds of 3D mesh models. | ['Yipeng Liu', 'Ce Zhu', 'Xinyu Lin'] | 2016-04-29 | null | null | null | null | ['interest-point-detection'] | ['computer-vision'] | [-1.39481887e-01 -2.64694184e-01 -1.14158466e-02 1.24395952e-01
-6.43281817e-01 -4.03975584e-02 4.64448512e-01 4.02774572e-01
-1.09475657e-01 7.59943053e-02 -2.04804912e-01 6.29575849e-02
-5.98102845e-02 -9.48359549e-01 -4.22514617e-01 -3.60980898e-01
-3.19890350e-01 6.64927721e-01 6.71273470e-01 -2.05410331e-01
8.72201622e-01 1.21267319e+00 -1.47072983e+00 -3.49311590e-01
5.73309481e-01 1.12137139e+00 -6.91044256e-02 3.39453161e-01
-1.16527818e-01 -2.97015846e-01 -3.82733554e-01 4.75102156e-01
4.83371377e-01 -1.77401051e-01 -5.24365127e-01 1.82582542e-01
3.08643728e-01 -7.72809237e-02 2.79914558e-01 1.12126589e+00
5.20322621e-01 2.10944057e-01 1.03755903e+00 -1.00449228e+00
-1.06102884e-01 -4.27845389e-01 -1.14211607e+00 7.22055584e-02
5.70102453e-01 -3.63892317e-01 5.97577393e-01 -1.57329559e+00
9.11432862e-01 1.34840643e+00 6.98476136e-01 1.05313368e-01
-1.05394685e+00 -6.07240438e-01 -1.86664835e-01 -1.47153288e-01
-1.87547171e+00 -1.64991945e-01 1.35579431e+00 -5.42735815e-01
5.35166323e-01 5.04440367e-01 5.65011621e-01 3.62178087e-02
4.66928065e-01 2.72328377e-01 7.70968020e-01 -6.16051733e-01
3.56426597e-01 -6.89730272e-02 1.21433623e-01 1.01226091e+00
1.73098579e-01 -2.55955569e-02 -2.16493666e-01 -6.62225902e-01
1.62693906e+00 2.31878832e-02 -2.65697092e-01 -9.07035470e-01
-1.08152890e+00 8.01211238e-01 3.89304608e-01 2.52354681e-01
-4.46473241e-01 -2.81445920e-01 -5.14968717e-03 -7.92842209e-02
7.58507729e-01 2.60058969e-01 -1.03579000e-01 2.24459797e-01
-7.01923728e-01 2.04677388e-01 5.99563539e-01 9.45010483e-01
9.76822853e-01 -2.86526382e-01 1.31052598e-01 8.03425312e-01
6.09642744e-01 5.51707625e-01 -1.87161386e-01 -9.65825379e-01
3.71387415e-02 1.18368435e+00 5.40009066e-02 -1.54934990e+00
-4.37110305e-01 -2.37197187e-02 -8.17927420e-01 7.37392664e-01
2.59414852e-01 2.03832150e-01 -7.25821257e-01 8.44463348e-01
1.10358167e+00 3.32854748e-01 -5.36576867e-01 1.14255059e+00
7.93247402e-01 6.06374085e-01 -5.54121912e-01 -3.25985581e-01
1.18120182e+00 -1.56111494e-01 -1.01081245e-01 3.56308043e-01
3.38130444e-01 -1.01747143e+00 7.20270813e-01 -8.18002373e-02
-1.13946915e+00 -4.21659678e-01 -9.21308994e-01 1.50017217e-01
5.54342270e-02 -1.44380461e-02 7.43023977e-02 1.18020467e-01
-5.82203567e-01 5.10748565e-01 -7.16316402e-01 -1.56761289e-01
2.28161484e-01 4.25578475e-01 -1.99830383e-01 2.23529905e-01
-7.18470633e-01 7.44970381e-01 -1.75392672e-01 -5.32318726e-02
-4.28149194e-01 -8.12748492e-01 -8.38565826e-01 -2.32841358e-01
3.54369700e-01 -5.42574763e-01 7.30435252e-01 -2.50962436e-01
-1.27385497e+00 1.36947262e+00 -3.35985690e-01 2.12743402e-01
3.08012336e-01 5.83006721e-03 -1.58482000e-01 2.56917179e-01
3.39136362e-01 -3.17530744e-02 8.59452307e-01 -1.54342675e+00
-6.08500600e-01 -6.56854391e-01 -3.24499130e-01 3.60859394e-01
3.35037738e-01 6.38276637e-02 -4.81699675e-01 -4.00122046e-01
9.59388852e-01 -6.22882664e-01 -3.54124039e-01 4.16487247e-01
-4.05932188e-01 -6.08852804e-01 1.25870371e+00 -2.65541762e-01
1.04698730e+00 -2.16730523e+00 1.36484392e-02 9.83680844e-01
5.03943443e-01 1.76875945e-02 2.97960550e-01 2.21125975e-01
2.51327872e-01 1.06860697e-01 -2.06043988e-01 1.38292387e-01
-3.67043763e-01 -3.50925952e-01 3.31597388e-01 8.66192758e-01
2.26719201e-01 2.90149748e-01 -8.19179177e-01 -6.45971715e-01
4.58932996e-01 6.81003094e-01 -3.24940234e-01 1.92810893e-02
2.12913066e-01 2.56499618e-01 -1.17375958e+00 8.21251750e-01
9.32736635e-01 -1.49735004e-01 -3.80068541e-01 -4.16292697e-01
-4.93221104e-01 -2.70921737e-02 -1.81467354e+00 1.25007343e+00
-4.72372696e-02 1.64011568e-01 3.46442580e-01 -7.17245162e-01
1.54094625e+00 2.25666791e-01 6.89018428e-01 -3.42924505e-01
3.22748244e-01 4.17789638e-01 -4.31483895e-01 -2.25614314e-03
2.26981893e-01 -2.90303417e-02 1.48386229e-02 2.84766704e-01
-5.65347254e-01 -6.18305624e-01 -3.11283886e-01 -1.62670389e-01
7.12906301e-01 -2.04481408e-01 6.79113448e-01 -6.87804043e-01
9.14476335e-01 5.53763993e-02 5.96491277e-01 3.88046920e-01
6.51117563e-02 7.81404376e-01 4.03866529e-01 -4.43528056e-01
-1.10034585e+00 -9.34318841e-01 -6.10739470e-01 2.49153614e-01
8.21319997e-01 -1.40375808e-01 -5.39764047e-01 -4.56079155e-01
2.85445482e-01 1.83335260e-01 -5.91508210e-01 1.70936450e-01
-9.90368187e-01 -2.02132374e-01 -3.37458998e-01 1.73802540e-01
2.56776005e-01 -7.06705630e-01 -8.53941202e-01 4.35319431e-02
3.83970380e-01 -6.28510773e-01 -6.11710548e-01 -2.58681715e-01
-1.21028411e+00 -1.38953340e+00 -8.81211281e-01 -1.00191772e+00
1.22894561e+00 5.13623595e-01 1.04853606e+00 3.05085957e-01
-3.54646027e-01 3.57657135e-01 -2.40708604e-01 -1.95685789e-01
-2.53642768e-01 -3.05302560e-01 1.01363242e-01 2.69445553e-02
1.53138474e-01 -3.20381969e-01 -7.39988327e-01 8.24113607e-01
-3.31736237e-01 4.64034267e-02 1.79197922e-01 4.84450787e-01
1.23756599e+00 -1.70695223e-02 1.00857474e-01 -5.25653839e-01
4.16010082e-01 -4.22289163e-01 -8.22136521e-01 -1.53268293e-01
-1.82407022e-01 -1.14355125e-01 2.25751847e-01 -1.83993280e-01
-3.86774123e-01 8.46677423e-02 1.11708213e-02 -7.08381414e-01
-1.74936101e-01 2.82384127e-01 -6.05229139e-02 -4.53853101e-01
4.00991797e-01 -4.98241978e-04 -5.00439219e-02 -6.92760348e-01
-7.65315071e-02 4.91659492e-01 1.68788761e-01 -5.66478014e-01
8.44386816e-01 8.05802226e-01 6.73564076e-01 -1.40734923e+00
-2.60732114e-01 -8.79216671e-01 -8.01055372e-01 -4.78157431e-01
5.78480363e-01 -4.70375091e-01 -8.28228414e-01 3.47429127e-01
-1.19751120e+00 2.89926857e-01 -2.15265572e-01 4.14486706e-01
-3.90597939e-01 4.51033920e-01 -3.48894447e-01 -9.43728626e-01
-4.98587459e-01 -1.13138294e+00 1.34101665e+00 1.86540842e-01
-2.58703917e-01 -8.35442722e-01 3.91018949e-02 -1.48339450e-01
3.61223519e-02 7.59859502e-01 1.21973288e+00 -2.19763160e-01
-4.74723995e-01 -5.18591464e-01 -1.37951761e-01 -1.62851527e-01
3.46115589e-01 2.55683631e-01 -2.97504097e-01 -1.85496584e-01
3.76296073e-01 4.73936051e-01 9.44204554e-02 7.90908635e-01
7.05076277e-01 6.82455748e-02 -6.41905904e-01 5.72749853e-01
1.52315831e+00 1.65182397e-01 3.88006210e-01 2.44526610e-01
6.55095041e-01 3.33321273e-01 8.94340336e-01 5.85070014e-01
3.72316651e-02 8.72359812e-01 5.34774303e-01 -2.84503102e-01
1.04238294e-01 -1.01822019e-01 -1.24824919e-01 8.03572893e-01
-4.11649495e-01 2.95976669e-01 -8.75270784e-01 3.71164382e-01
-1.60909712e+00 -4.80494171e-01 -6.40591323e-01 2.58799601e+00
4.81527984e-01 2.33375579e-01 9.67137888e-02 2.93353617e-01
1.11047268e+00 -1.61494434e-01 -5.97565293e-01 -1.91478640e-01
1.50477171e-01 3.24101031e-01 3.28386962e-01 6.07149124e-01
-9.91408587e-01 6.29304647e-01 5.31243801e+00 8.55289519e-01
-1.02766991e+00 -3.49284321e-01 3.28610063e-01 3.76346350e-01
-8.88458192e-02 -1.03582799e-01 -9.75733817e-01 2.80793965e-01
2.80925781e-02 -2.17141241e-01 -3.00423682e-01 8.77590179e-01
3.94014627e-01 -2.61405230e-01 -8.83186877e-01 1.07231390e+00
1.26844319e-02 -1.40099299e+00 1.17660560e-01 1.18652999e-01
6.47615850e-01 -2.92621851e-01 -1.72600448e-01 -3.18857193e-01
-2.75470942e-01 -7.61433542e-01 4.88337278e-01 4.85102445e-01
8.52278352e-01 -9.14416075e-01 4.02275950e-01 3.76578629e-01
-1.75252342e+00 6.22116029e-01 -4.14605409e-01 1.88589483e-01
1.98289752e-01 6.51896119e-01 -7.91190743e-01 4.05506641e-01
5.28915107e-01 6.25853240e-01 -5.26350103e-02 1.26270831e+00
1.40183389e-01 1.32812902e-01 -6.35553837e-01 -1.54751599e-01
1.09217979e-01 -5.58245897e-01 1.13065565e+00 6.60923004e-01
5.31202137e-01 5.56284130e-01 2.31912732e-01 9.50501084e-01
8.21918696e-02 5.22964418e-01 -6.97712362e-01 7.37294734e-01
7.79117763e-01 1.07969570e+00 -1.18759453e+00 -1.14470720e-01
-2.02881858e-01 5.70115268e-01 -1.53416947e-01 4.32958640e-02
-4.05949622e-01 -4.30725515e-01 5.83529055e-01 8.18486691e-01
1.43211886e-01 -5.25090516e-01 -5.18227696e-01 -7.06011355e-01
-3.32021490e-02 -5.31745732e-01 8.15613195e-02 -4.29754943e-01
-1.08377957e+00 2.95761853e-01 4.38011996e-02 -1.48796487e+00
1.19208589e-01 -5.42371809e-01 -7.32545674e-01 9.89478588e-01
-1.06181729e+00 -7.98567057e-01 -1.59305930e-01 6.67925537e-01
6.10634089e-01 2.54588187e-01 6.76490605e-01 -7.22109079e-02
-2.40092240e-02 1.68432593e-01 9.07880962e-02 1.90457880e-01
1.12836778e-01 -8.42064738e-01 5.64001858e-01 5.22517622e-01
-2.23306015e-01 5.63909769e-01 4.80387688e-01 -1.00560713e+00
-1.60961628e+00 -7.01532423e-01 7.12435782e-01 -2.08457470e-01
1.57407150e-01 -2.32734650e-01 -9.74474311e-01 1.38558820e-01
-5.93485653e-01 6.23079687e-02 1.38343722e-01 -1.82649344e-01
1.93858370e-01 2.00441092e-01 -1.44216168e+00 6.71644032e-01
9.09497321e-01 -1.71225980e-01 -5.09830832e-01 4.50580418e-02
1.94767982e-01 -6.52828038e-01 -1.10500252e+00 8.98069978e-01
5.21444619e-01 -8.20054948e-01 1.38990891e+00 -6.04844503e-02
-5.75631261e-02 -5.64719617e-01 7.18905479e-02 -8.80532920e-01
-4.39045489e-01 -5.62815368e-01 9.75202955e-03 7.63970435e-01
-9.55595262e-03 -4.29957926e-01 9.60723519e-01 3.88614982e-01
-1.44658968e-01 -1.05731308e+00 -1.28586578e+00 -4.19348806e-01
-1.66240364e-01 -7.20686167e-02 4.36436951e-01 9.39337313e-01
-9.24380198e-02 2.25279197e-01 9.15313289e-02 4.40528184e-01
1.03314352e+00 4.34910983e-01 6.91664219e-01 -1.84439170e+00
6.11338615e-01 -5.08360326e-01 -7.71379232e-01 -1.21688497e+00
-2.73782760e-01 -6.77760601e-01 -2.87563056e-01 -1.28202164e+00
-1.35355130e-01 -9.77631450e-01 -9.56627801e-02 -8.51723179e-02
9.66784358e-02 1.34563178e-01 -1.60548359e-01 3.25871915e-01
-1.60568044e-01 4.07446951e-01 1.50696754e+00 3.27514708e-01
-6.19356155e-01 1.71757102e-01 -8.83391723e-02 1.04151380e+00
5.27403414e-01 -2.82209486e-01 1.64408743e-01 -1.17548309e-01
-1.72607884e-01 2.78944969e-01 2.35287488e-01 -8.51603806e-01
2.09746733e-01 -3.29070300e-01 4.65454996e-01 -1.07629478e+00
4.37505245e-01 -9.55996215e-01 1.82985708e-01 3.45539987e-01
1.55890971e-01 -1.25315841e-02 -4.86394903e-03 2.08124772e-01
1.80144131e-01 -3.92735720e-01 1.19774902e+00 -3.16072434e-01
-6.50556386e-01 3.57600927e-01 -1.17104754e-01 7.87001923e-02
1.30228817e+00 -6.56866431e-01 2.09777385e-01 9.16961879e-02
-3.92628491e-01 1.36994809e-01 8.57051730e-01 1.69507295e-01
1.36463869e+00 -1.49970102e+00 -8.56440604e-01 7.31114030e-01
-3.61952931e-02 3.22094798e-01 -5.03407232e-02 6.54538810e-01
-7.27000117e-01 8.55482966e-02 -5.23278341e-02 -1.00843716e+00
-1.61784458e+00 1.94567785e-01 3.77711296e-01 1.99296385e-01
-7.67033100e-01 8.42246652e-01 4.11088504e-02 -1.99689627e-01
5.14419796e-03 -3.92476141e-01 -2.46704891e-01 -1.71407968e-01
1.64145052e-01 8.78710210e-01 9.66188312e-02 -1.07220876e+00
-5.70980966e-01 1.52094579e+00 9.99230444e-02 2.54737526e-01
1.17732644e+00 2.98441891e-02 -1.14163697e-01 4.04965222e-01
1.35927022e+00 1.52579889e-01 -9.25675750e-01 -1.77615374e-01
-1.10426053e-01 -8.22643876e-01 1.48969859e-01 -1.07957292e-02
-8.67860317e-01 6.19885623e-01 6.27125204e-01 2.66746610e-01
7.18044043e-01 3.62914324e-01 6.70717120e-01 -2.04316020e-01
7.01620758e-01 -9.52934623e-01 -3.25269043e-01 5.17484069e-01
9.85599995e-01 -9.70705628e-01 3.22044671e-01 -9.11559165e-01
1.15129568e-01 1.26030636e+00 4.80745941e-01 -6.40439689e-01
9.91026700e-01 6.76485151e-02 -8.86860266e-02 -6.74881995e-01
4.09424081e-02 -1.64766088e-02 6.85877502e-01 2.92493731e-01
2.51029909e-01 -1.41016290e-01 -4.69281971e-01 -5.65431602e-02
1.71140924e-01 -2.67665595e-01 1.15870774e-01 1.07761467e+00
-8.26146722e-01 -8.23394358e-01 -8.66921008e-01 4.02442008e-01
-2.34016374e-01 3.33246559e-01 -1.19348951e-01 8.82526577e-01
-6.71215728e-02 5.46027899e-01 2.43721992e-01 -1.93206042e-01
6.85290694e-01 -3.81699085e-01 3.98232222e-01 -6.42048120e-01
-1.83294997e-01 4.05972511e-01 -3.72848988e-01 -4.06580240e-01
-2.65456557e-01 -6.99523211e-01 -1.63574886e+00 -1.72950014e-01
-5.92606962e-01 2.71263510e-01 6.83089912e-01 5.15096664e-01
3.70940208e-01 -2.83593237e-01 1.18767500e+00 -1.33611429e+00
-2.10265040e-01 -4.19073761e-01 -7.63081074e-01 3.08784097e-01
9.76917371e-02 -1.05617821e+00 -5.11586785e-01 -2.35558689e-01] | [7.868456840515137, -2.9994735717773438] |
b5481244-0741-48be-b251-1645c63b631c | empirical-evaluation-of-leveraging-named | 1904.10195 | null | http://arxiv.org/abs/1904.10195v1 | http://arxiv.org/pdf/1904.10195v1.pdf | Empirical Evaluation of Leveraging Named Entities for Arabic Sentiment Analysis | Social media reflects the public attitudes towards specific events. Events
are often related to persons, locations or organizations, the so-called Named
Entities. This can define Named Entities as sentiment-bearing components. In
this paper, we dive beyond Named Entities recognition to the exploitation of
sentiment-annotated Named Entities in Arabic sentiment analysis. Therefore, we
develop an algorithm to detect the sentiment of Named Entities based on the
majority of attitudes towards them. This enabled tagging Named Entities with
proper tags and, thus, including them in a sentiment analysis framework of two
models: supervised and lexicon-based. Both models were applied on datasets of
multi-dialectal content. The results revealed that Named Entities have no
considerable impact on the supervised model, while employing them in the
lexicon-based model improved the classification performance and outperformed
most of the baseline systems. | ['Ismail Babaoglu', 'Mourad Gridach', 'Hatem Haddad', 'Hala Mulki'] | 2019-04-23 | null | null | null | null | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-4.82842118e-01 1.31894946e-01 1.81733407e-02 -5.27316868e-01
-3.04828674e-01 -8.88690710e-01 6.67533219e-01 5.61747015e-01
-8.25069010e-01 7.03212678e-01 6.18764043e-01 -9.34460089e-02
3.50763708e-01 -1.07451010e+00 -2.72856086e-01 -5.63676655e-01
1.80120394e-01 1.88588873e-01 2.62009025e-01 -5.94396234e-01
5.26570976e-01 4.37200487e-01 -1.32761323e+00 4.50762302e-01
8.63166928e-01 8.82001698e-01 -1.25204459e-01 2.02916652e-01
-7.05046237e-01 1.09878242e+00 -7.59445906e-01 -8.69470119e-01
-2.70657510e-01 -2.19592571e-01 -6.68298542e-01 1.67503953e-01
-2.68948406e-01 2.99623817e-01 5.33993542e-01 9.32289898e-01
3.71265531e-01 -6.41234368e-02 8.36188793e-01 -6.87861860e-01
-4.16810274e-01 6.90168142e-01 -2.92900294e-01 -1.47069842e-01
4.75291729e-01 -7.47449875e-01 1.04247880e+00 -1.15737307e+00
8.24658990e-01 9.02583599e-01 8.09972644e-01 8.02441686e-02
-5.54788828e-01 -3.51303726e-01 2.58133262e-01 -1.49551511e-01
-1.28738678e+00 -3.58422279e-01 5.51025271e-01 -7.38786697e-01
7.26829112e-01 2.96080194e-04 5.62282622e-01 5.94275594e-01
8.60259831e-02 5.28040290e-01 1.25686014e+00 -7.31452227e-01
3.77569407e-01 8.17854106e-01 4.26984102e-01 5.23460746e-01
3.38551134e-01 -6.79043353e-01 -5.81290960e-01 -2.21595839e-01
-1.44014275e-02 -1.79538250e-01 1.81922406e-01 -1.76768512e-01
-9.97568846e-01 1.03143191e+00 2.63039052e-01 5.78170657e-01
-6.81149721e-01 -5.44409037e-01 5.88261247e-01 5.00008203e-02
9.18628335e-01 5.83370566e-01 -8.45418870e-01 1.21993609e-01
-6.76477730e-01 -1.71115369e-01 1.00909972e+00 5.86030960e-01
9.34916437e-01 -2.92650551e-01 1.04821511e-01 9.53535080e-01
5.82202733e-01 5.99392951e-01 7.89588332e-01 1.80734452e-02
3.17767680e-01 1.33740306e+00 3.02306652e-01 -1.39719450e+00
-7.36907125e-01 -2.79075325e-01 -2.42040128e-01 -3.10958415e-01
3.98036152e-01 -7.68057346e-01 -7.74157584e-01 1.47240055e+00
6.59341633e-01 -3.81297231e-01 4.38797683e-01 6.49435163e-01
9.70367968e-01 6.11961067e-01 3.42155516e-01 -1.99863166e-01
1.76699519e+00 -6.49301887e-01 -8.83169949e-01 -1.41106457e-01
9.64749634e-01 -1.14589870e+00 6.14288449e-01 1.93896830e-01
-7.04078794e-01 -2.70561486e-01 -6.91876531e-01 4.22606140e-01
-1.17365456e+00 4.30089056e-01 7.67054021e-01 1.19675481e+00
-7.81127810e-01 1.71523243e-01 -6.30169690e-01 -5.97124100e-01
5.96093833e-02 1.68557197e-01 -4.66884404e-01 5.20799398e-01
-1.38257873e+00 1.08514023e+00 4.32502896e-01 1.17381297e-01
-1.05393760e-01 1.37563040e-02 -9.71882701e-01 -2.97657460e-01
1.13095976e-01 5.65462634e-02 8.89559448e-01 -1.40152085e+00
-1.29420662e+00 1.22469902e+00 -2.54607111e-01 -2.65850395e-01
8.96979123e-02 -2.96506345e-01 -9.83447790e-01 1.10063449e-01
4.36730295e-01 9.98425633e-02 4.89148021e-01 -1.16286039e+00
-9.96226251e-01 -2.46831924e-01 1.40225947e-01 2.07537621e-01
-8.51514637e-01 6.65580511e-01 -3.06616992e-01 -5.57845175e-01
1.80639848e-01 -1.01125717e+00 -1.38661250e-01 -9.26216424e-01
-3.70845258e-01 -3.64547670e-01 3.69028151e-01 -5.00452340e-01
1.29125059e+00 -2.21977115e+00 -2.29605347e-01 3.43562305e-01
-1.41920760e-01 2.22372547e-01 3.98645878e-01 7.65350461e-01
-2.37913989e-02 1.27769381e-01 1.48159668e-01 -9.58471000e-02
1.28401548e-01 -8.33482482e-03 -4.48523164e-01 2.94866234e-01
3.62694114e-01 6.06662571e-01 -7.47807086e-01 -5.74711621e-01
-1.03816517e-01 3.31345737e-01 -3.34222525e-01 1.97493136e-02
-3.20324712e-02 2.34006509e-01 -8.05184841e-01 5.38869143e-01
6.06414497e-01 -8.95662010e-02 4.66441035e-01 -2.83558309e-01
-4.56884146e-01 4.43726957e-01 -1.22868872e+00 1.00715280e+00
-5.56885660e-01 3.79746318e-01 -2.09184885e-01 -6.94257021e-01
1.06356502e+00 4.79893118e-01 4.09976631e-01 -4.92292583e-01
4.84271556e-01 5.63525736e-01 -2.99726963e-01 -5.17778933e-01
9.68950093e-01 -1.99525207e-01 -5.79830587e-01 4.27170575e-01
6.38871640e-02 1.47170380e-01 4.50998783e-01 1.48828372e-01
4.71823603e-01 2.38246601e-02 6.62620306e-01 -2.68305868e-01
9.22839165e-01 3.03585202e-01 3.97409439e-01 3.55244845e-01
-2.71806680e-02 3.29628617e-01 5.98519027e-01 -2.77326614e-01
-5.25446057e-01 -5.68445742e-01 -2.67948359e-01 1.45553017e+00
-4.31075469e-02 -6.32583022e-01 -8.22567165e-01 -1.01231670e+00
-3.42948169e-01 4.56772596e-01 -5.38942337e-01 2.49322385e-01
-3.04860860e-01 -1.20973027e+00 4.22971338e-01 1.96564913e-01
1.80313215e-01 -1.25737011e+00 -3.71549129e-01 3.55226845e-01
-1.61395028e-01 -1.13164401e+00 -4.90577966e-02 4.60070938e-01
-3.87564838e-01 -9.96104896e-01 -7.10374415e-01 -8.20547760e-01
7.98754930e-01 -1.68557465e-01 1.06087816e+00 -1.94568977e-01
5.84437132e-01 2.75764644e-01 -1.06156588e+00 -9.82179761e-01
-3.33066374e-01 3.13518137e-01 3.26918215e-02 5.04377306e-01
8.15613151e-01 -1.14370927e-01 -2.13575989e-01 4.53801125e-01
-8.23456526e-01 -4.40400898e-01 3.81170452e-01 3.67569655e-01
1.52132064e-01 -2.89777145e-02 8.20920646e-01 -1.44541836e+00
6.59453273e-01 -5.47720909e-01 -1.64119065e-01 7.11261183e-02
-4.73670632e-01 -9.23558250e-02 6.15221381e-01 -1.78559989e-01
-1.24830604e+00 2.64550030e-01 -3.24311852e-01 7.43325233e-01
-2.59419113e-01 1.06547761e+00 -2.88132489e-01 4.20550667e-02
5.78446567e-01 4.81374515e-03 -5.73705912e-01 -2.25485563e-01
2.34663501e-01 9.49369609e-01 -1.18098564e-01 -4.11027431e-01
4.23225909e-01 3.45713824e-01 -3.96762460e-01 -8.34353864e-01
-1.26710236e+00 -7.61736870e-01 -8.65084887e-01 -2.57236719e-01
1.03116608e+00 -1.08644843e+00 -2.30761454e-01 8.15729082e-01
-8.78555417e-01 2.40813434e-01 -2.10555613e-01 7.25921333e-01
-1.68927424e-02 2.34448776e-01 -5.60164511e-01 -8.44551146e-01
-1.48330525e-01 -7.80473530e-01 6.30658150e-01 5.09964943e-01
-4.01402920e-01 -1.32590580e+00 3.17522675e-01 2.70332724e-01
2.91052341e-01 1.33293614e-01 6.30735755e-01 -1.46697748e+00
3.39707345e-01 -4.07309204e-01 3.06497421e-02 3.10507596e-01
1.79148629e-01 1.49751306e-02 -9.53594923e-01 1.92944944e-01
-8.42489451e-02 -2.11536661e-01 5.92671812e-01 3.53518641e-03
1.95267484e-01 -2.28289217e-01 -7.28554651e-02 5.01060709e-02
1.25545585e+00 2.44506329e-01 5.59362352e-01 8.50017786e-01
5.72164297e-01 8.24483931e-01 8.89488101e-01 6.26675487e-01
7.42215693e-01 3.43409270e-01 9.19037312e-02 -3.37740004e-01
4.97780532e-01 7.87913892e-03 6.80689991e-01 1.07376885e+00
-3.97271775e-02 -4.92708027e-01 -1.04604900e+00 7.90467978e-01
-1.64711285e+00 -6.99987173e-01 -6.03117824e-01 1.81275237e+00
9.26797628e-01 2.82016963e-01 1.04178756e-01 1.08958669e-01
9.43513453e-01 2.18193740e-01 4.31634523e-02 -5.07284105e-01
-4.13138032e-01 2.78398424e-01 4.95361775e-01 7.73741752e-02
-1.46960843e+00 1.19958365e+00 5.61681414e+00 6.40230179e-01
-1.10730410e+00 2.15340853e-01 3.90244156e-01 5.45455277e-01
-1.39277518e-01 -9.10145342e-02 -1.16883457e+00 3.49599838e-01
9.51952279e-01 4.80862334e-02 -2.93606818e-01 7.77913451e-01
2.20217839e-01 -2.34977812e-01 -2.80964971e-01 2.69712389e-01
2.94305384e-01 -8.38217795e-01 -1.06902599e-01 -6.18513823e-02
9.01218235e-01 -5.89861944e-02 -2.88154304e-01 7.14445263e-02
3.72292191e-01 -4.36889201e-01 9.91142750e-01 4.94393349e-01
2.34174356e-01 -8.82273614e-01 1.18247271e+00 5.02471700e-02
-1.08235323e+00 1.86907813e-01 -1.69055432e-01 -8.48378018e-02
2.31644139e-01 7.31277645e-01 -8.12539816e-01 5.75273633e-01
5.34004331e-01 6.77918971e-01 -7.79431343e-01 7.61109293e-01
-5.74983537e-01 9.17243302e-01 -1.55097082e-01 -4.13882881e-01
4.22515601e-01 -2.84130782e-01 9.69289690e-02 1.52276087e+00
2.74279505e-01 -1.66654348e-01 1.33361578e-01 -5.28716482e-02
7.37550259e-02 1.01092219e+00 -2.48080894e-01 -2.87124902e-01
2.00410560e-01 1.62389076e+00 -1.39125657e+00 -3.40931863e-01
-7.22201526e-01 7.93317080e-01 9.63075608e-02 -7.04736542e-03
-5.35377860e-01 -6.09281600e-01 1.22172661e-01 1.76961087e-02
4.65455651e-01 9.16958973e-03 -1.80364847e-01 -1.25303090e+00
-1.09796949e-01 -7.11933315e-01 4.59533244e-01 -5.58260500e-01
-1.31801772e+00 8.96761954e-01 -4.22209054e-01 -1.15620148e+00
-1.18442640e-01 -8.73870850e-01 -3.89584303e-01 7.40290999e-01
-1.51471651e+00 -1.31381655e+00 -3.61400247e-02 5.33282280e-01
-1.58351645e-01 -3.12006682e-01 1.07050633e+00 4.54723746e-01
-5.67315638e-01 3.04092407e-01 2.26746023e-01 5.75153947e-01
1.02648759e+00 -1.26542234e+00 3.93228084e-02 8.93112183e-01
3.24442863e-01 7.86137223e-01 5.91943681e-01 -6.83303177e-01
-6.76549017e-01 -8.17953348e-01 1.64024413e+00 -4.67971265e-01
1.01070797e+00 -2.25016654e-01 -5.06380260e-01 4.91434842e-01
2.25988150e-01 -4.43208337e-01 1.30280125e+00 3.55250657e-01
-2.52528608e-01 2.27343112e-01 -1.02507198e+00 3.75291020e-01
3.91285390e-01 -5.46548426e-01 -7.77672112e-01 3.35621089e-01
3.45627457e-01 -1.92129537e-01 -9.51349854e-01 4.67308871e-02
5.01665175e-01 -8.51104021e-01 6.70696616e-01 -6.58526123e-01
4.34715867e-01 -3.83973300e-01 -2.22715616e-01 -1.37579346e+00
5.03108725e-02 3.59741896e-02 4.02666599e-01 1.77509165e+00
9.01893139e-01 -8.99245739e-01 5.93042314e-01 5.36316156e-01
-5.51702641e-02 -2.14726254e-01 -3.74303371e-01 1.23714888e-02
-5.74873574e-02 -3.00195992e-01 5.20512998e-01 1.38487840e+00
3.07091296e-01 4.13988948e-01 -1.41543567e-01 1.50130272e-01
-2.56010503e-01 1.88108370e-01 5.23597300e-01 -1.24770570e+00
2.94476151e-01 -1.50060147e-01 -5.69137812e-01 -4.49418843e-01
2.61945307e-01 -6.57894015e-01 -8.90804306e-02 -1.49858499e+00
-7.95475841e-02 -5.98482251e-01 -2.98894107e-01 5.86961210e-01
-3.72471213e-01 5.18802106e-01 -4.73932596e-03 1.15393475e-01
-8.15270066e-01 1.12423651e-01 7.27813423e-01 1.68743804e-01
-3.28036100e-01 2.03373566e-01 -9.84966815e-01 1.11299586e+00
9.12805617e-01 -7.04639971e-01 9.90522206e-02 3.71670648e-02
1.09235513e+00 -4.26556855e-01 -2.56597549e-01 -7.47834325e-01
2.13149339e-01 -1.36757970e-01 2.77587563e-01 -5.43260694e-01
-2.21079569e-02 -8.26185942e-01 -7.99549296e-02 2.52445847e-01
-1.06216297e-01 5.05459905e-02 9.28297266e-03 9.84630957e-02
-5.25751472e-01 -6.31215155e-01 5.55087984e-01 -9.99395847e-02
-9.33929324e-01 -3.52948815e-01 -8.87286425e-01 4.02773730e-02
1.07422614e+00 -5.20325564e-02 -5.83767220e-02 -2.26072282e-01
-9.43112195e-01 -1.63055375e-01 4.91320580e-01 2.67957896e-01
1.98209304e-02 -9.88897085e-01 -5.86553872e-01 -6.56698570e-02
4.15152192e-01 -4.89807308e-01 8.14574957e-03 7.31777251e-01
-5.68005860e-01 3.80853951e-01 -1.09930441e-01 -7.32693449e-02
-1.17344820e+00 3.02023530e-01 2.22447589e-01 -4.30377424e-01
3.20352256e-01 7.65384436e-01 -3.18649672e-02 -1.06252563e+00
-2.32409611e-01 2.20260009e-01 -1.44999957e+00 1.11541569e+00
3.58029217e-01 6.64538071e-02 3.37732166e-01 -1.46043801e+00
-5.42205095e-01 6.01993084e-01 -7.20440820e-02 -1.37173310e-01
1.49023962e+00 -3.17623347e-01 -4.98719633e-01 7.17217982e-01
8.89096916e-01 1.04114723e+00 -3.16726148e-01 -1.20919392e-01
4.75480229e-01 3.87676135e-02 -8.82415324e-02 -8.56127501e-01
-9.98816550e-01 4.21287358e-01 3.03415686e-01 6.53805375e-01
1.08511353e+00 -1.12883188e-01 2.86375612e-01 2.95242220e-01
4.17468965e-01 -1.23684514e+00 -3.72347772e-01 9.83280361e-01
9.61746499e-02 -1.17601860e+00 -6.13987073e-02 -6.14508212e-01
-1.12739694e+00 1.23420501e+00 3.41697752e-01 1.02103911e-01
9.44907844e-01 2.99507137e-02 6.06723309e-01 -4.53528941e-01
-1.81422845e-01 -6.98231161e-01 4.94946867e-01 3.26932430e-01
8.84538233e-01 1.83667228e-01 -9.41702843e-01 1.01224005e+00
-3.24802130e-01 -1.84203655e-01 6.76752806e-01 1.07238090e+00
-5.10356069e-01 -1.26750910e+00 -4.22830284e-01 3.36749345e-01
-1.13573909e+00 -2.63762802e-01 -6.41550303e-01 5.05160332e-01
4.10976857e-01 1.35848081e+00 4.43056412e-02 -4.00274575e-01
5.30740857e-01 2.28248477e-01 -1.94367737e-01 -7.66997397e-01
-1.32682848e+00 5.33533283e-02 5.07908285e-01 -2.70900335e-02
-1.14265978e+00 -8.21094096e-01 -1.25245488e+00 -4.23693359e-02
-8.55847001e-01 7.46860921e-01 9.46443856e-01 1.13789570e+00
1.41364366e-01 1.75945655e-01 9.96833146e-01 -4.00559008e-01
1.00977845e-01 -1.01926732e+00 -6.86941445e-01 4.71454382e-01
-1.09091684e-01 -3.68911147e-01 -4.57120806e-01 2.80759424e-01] | [11.05174732208252, 6.9406867027282715] |
0ac4ac1f-b11d-43ea-8694-7ac5fc37906b | smac-symbiotic-multi-agent-construction | 2010.08473 | null | https://arxiv.org/abs/2010.08473v1 | https://arxiv.org/pdf/2010.08473v1.pdf | SMAC: Symbiotic Multi-Agent Construction | We present a novel concept of a heterogeneous, distributed platform for autonomous 3D construction. The platform is composed of two types of robots acting in a coordinated and complementary fashion: (i) A collection of communicating smart construction blocks behaving as a form of growable smart matter, and capable of planning and monitoring their own state and the construction progress; and (ii) A team of inchworm-shaped builder robots designed to navigate and modify the 3D structure, following the guidance of the smart blocks. We describe the design of the hardware and introduce algorithms for navigation and construction that support a wide class of 3D structures. We demonstrate the capabilities of our concept and characterize its performance through simulations and real-robot experiments. | ['Carlo Pinciroli', 'Gregory Lewin', 'Hannan Liang', 'Josue Contreras', 'Trevor Rizzo', 'Neel Dhanaraj', 'Caleb Wagner'] | 2020-10-16 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-2.80717641e-01 6.26170516e-01 4.23132598e-01 -4.50779125e-03
1.57057658e-01 -8.18447948e-01 4.98746336e-01 -1.72944427e-01
4.77714807e-01 3.22463930e-01 8.28290880e-02 -3.37725058e-02
-3.14432919e-01 -1.28869104e+00 -6.96875036e-01 -7.92325437e-01
-6.44329965e-01 1.23791838e+00 6.73728347e-01 -8.63581240e-01
1.15223207e-01 6.89605117e-01 -1.81016326e+00 -2.44936004e-01
2.60117173e-01 1.01521277e+00 7.63609588e-01 6.42679930e-01
3.90101492e-01 8.53204846e-01 -1.46673471e-01 5.79037189e-01
3.76794875e-01 9.41109881e-02 -9.85891342e-01 4.05552208e-01
-5.52460730e-01 -2.24655405e-01 -2.63377298e-02 3.10019553e-01
5.83927929e-01 -2.49659717e-01 3.73760879e-01 -1.13947463e+00
-3.10452580e-01 7.15669155e-01 2.18899623e-01 -9.09739375e-01
1.03783560e+00 1.23800538e-01 5.35913944e-01 -7.18598247e-01
1.06906354e+00 1.17091417e+00 6.39394939e-01 4.79030371e-01
-1.02292681e+00 -1.07265092e-01 -2.44260520e-01 -4.24770206e-01
-1.45781767e+00 -6.12838268e-01 6.71515226e-01 -3.53279263e-01
9.75217044e-01 -1.61284178e-01 8.22901785e-01 6.62892044e-01
8.09429705e-01 3.36397171e-01 5.76714873e-01 -5.40613890e-01
8.11858952e-01 -4.54554498e-01 -4.81025964e-01 8.86399746e-01
5.28810620e-01 -1.19675770e-01 -4.04545486e-01 -2.58225828e-01
1.13826334e+00 -2.76608795e-01 5.01598045e-02 -1.23054934e+00
-1.36285555e+00 1.93049476e-01 3.32882851e-01 4.04395521e-01
-6.93228781e-01 8.73329639e-01 8.15922543e-02 3.26028973e-01
-2.36010551e-01 5.10302186e-01 -7.35723734e-01 -2.97540396e-01
2.18086243e-01 1.97419345e-01 1.30214810e+00 1.45086503e+00
9.06248152e-01 6.08363710e-02 6.84297144e-01 3.11552852e-01
6.37639999e-01 5.31340301e-01 8.20861608e-02 -1.42804825e+00
3.79338153e-02 9.39418018e-01 5.14110327e-01 -7.54208803e-01
-9.28502619e-01 1.61280140e-01 -4.31648821e-01 7.34933376e-01
-3.22982699e-01 -1.77647784e-01 -4.50209260e-01 1.25744748e+00
7.82354593e-01 -6.42267525e-01 9.08532143e-02 4.41245407e-01
2.19955653e-01 4.51192021e-01 -4.48667824e-01 2.20033765e-01
8.69358063e-01 -7.03353763e-01 -2.54882574e-01 1.73037916e-01
9.36505973e-01 -2.15229765e-01 6.20711029e-01 4.30034816e-01
-1.28719735e+00 -3.29593629e-01 -1.10919309e+00 3.76341969e-01
-1.26505345e-01 -2.90918827e-01 8.24984133e-01 4.97988433e-01
-1.68036759e+00 4.49157506e-01 -1.11395109e+00 -6.35064662e-01
-1.44012526e-01 5.39559186e-01 -5.05545795e-01 5.21542169e-02
-3.77340138e-01 9.31133568e-01 3.35565895e-01 -1.49728015e-01
-1.36670339e+00 3.01839504e-02 -5.10587990e-01 -8.66272226e-02
2.32038543e-01 -1.05640876e+00 1.53202689e+00 -7.10971430e-02
-2.12136364e+00 8.92428756e-01 7.70979643e-01 3.21713053e-02
-3.22466269e-02 -2.35544950e-01 5.67902997e-02 -9.95302200e-03
2.88889498e-01 3.31248492e-01 1.43188998e-01 -1.60297394e+00
-7.67363071e-01 -4.22937870e-01 6.41555060e-03 2.16145217e-01
9.19482559e-02 -4.10943747e-01 -1.86632693e-01 3.35507877e-02
7.52618015e-01 -1.31660140e+00 -8.64451587e-01 8.18269476e-02
-2.27680907e-01 -1.80067956e-01 1.00505257e+00 2.71054119e-01
6.66509688e-01 -1.91586912e+00 3.69471133e-01 4.29599017e-01
7.93573633e-02 -6.11230671e-01 1.22325845e-01 1.47191465e+00
6.55880392e-01 -2.84752935e-01 3.80038954e-02 -1.02141842e-01
2.18747199e-01 6.67707145e-01 2.15142399e-01 4.26098287e-01
-7.23951459e-01 5.10716677e-01 -1.05082417e+00 -7.59679265e-03
1.86408609e-01 -4.81776074e-02 -5.89392245e-01 1.45149857e-01
-3.90529543e-01 7.39030302e-01 -1.14623821e+00 1.14183199e+00
1.87995732e-01 -8.44099075e-02 6.95454061e-01 3.67453873e-01
-6.13833785e-01 1.35631308e-01 -1.35717499e+00 2.11734104e+00
-7.57460177e-01 -1.50985301e-01 1.14017224e+00 -7.19514191e-01
1.37575758e+00 6.28193796e-01 9.67080891e-01 -1.85008734e-01
3.08274239e-01 4.12722826e-01 -4.17764276e-01 -7.61390865e-01
4.72803861e-01 1.66032761e-01 -8.13990831e-01 8.87488127e-01
-2.36803800e-01 -1.15180361e+00 -2.20993668e-01 5.44780074e-03
2.12333512e+00 4.41082329e-01 2.68310905e-01 -4.66554701e-01
3.69801313e-01 5.27181886e-02 3.08229923e-01 6.64392829e-01
7.25105107e-02 -1.11838125e-01 -1.74389139e-01 -7.25785971e-01
-1.01156223e+00 -1.38396883e+00 3.04634660e-01 1.04953587e+00
8.45454335e-01 -5.59137404e-01 -2.58577257e-01 -7.11370334e-02
3.16701144e-01 4.41460758e-01 -3.39578032e-01 9.44958180e-02
-7.10736871e-01 1.35688826e-01 2.50019848e-01 3.92622858e-01
4.15356576e-01 -1.03971744e+00 -1.42730021e+00 6.17957890e-01
2.57874489e-01 -7.57946193e-01 3.99880767e-01 5.35286725e-01
-7.75477409e-01 -1.01675940e+00 3.13845485e-01 -1.30579519e+00
8.93688798e-01 4.84808743e-01 9.07396913e-01 4.47538942e-01
3.26911509e-02 1.37143612e+00 -7.80673742e-01 -1.44497097e-01
-9.54750299e-01 -2.83731576e-02 4.31449115e-01 -5.98723471e-01
-7.85767317e-01 -1.36327064e+00 -6.12566590e-01 8.17078590e-01
-4.81396019e-01 2.99405664e-01 7.16252327e-01 1.45702707e-02
4.00608242e-01 2.63414502e-01 1.17022492e-01 -1.71889126e-01
3.24845523e-01 -5.70188463e-01 -5.33973634e-01 1.43160388e-01
-3.84668499e-01 -1.17150232e-01 4.38225716e-01 7.40746707e-02
-9.39796388e-01 9.31179345e-01 -8.14284682e-02 4.43011284e-01
-3.69878352e-01 8.22348595e-02 -3.62976015e-01 -5.06635308e-01
4.43913102e-01 1.71255469e-01 7.97607154e-02 -6.46055281e-01
7.02855229e-01 7.93882608e-01 6.20742321e-01 -1.03641284e+00
9.50447142e-01 9.08562481e-01 3.73057127e-01 -6.79138362e-01
2.94655412e-01 -6.39753789e-02 -7.46649444e-01 -7.06316411e-01
6.09121203e-01 -5.84168315e-01 -9.40706432e-01 5.73972106e-01
-1.02236319e+00 -8.58903289e-01 -4.18154299e-01 3.95678841e-02
-1.35812271e+00 -8.66528153e-02 -6.90944254e-01 -6.63500547e-01
-5.33062696e-01 -1.03178024e+00 1.29104650e+00 -1.99237272e-01
-3.14905196e-01 -4.45841461e-01 8.02585185e-01 -7.94372559e-02
7.38814950e-01 7.84337163e-01 6.69270992e-01 9.79258418e-02
-1.05370533e+00 -6.37521565e-01 7.04438567e-01 -2.30376646e-01
5.10824084e-01 9.64350700e-02 -1.96083635e-01 -1.19843215e-01
-7.94028956e-03 -2.83863306e-01 -5.60019851e-01 3.89027111e-02
3.42616439e-01 -1.00467533e-01 -8.92154396e-01 1.36450389e-02
1.36970162e+00 5.44718623e-01 5.86124659e-01 6.85139298e-01
1.97162762e-01 3.76085252e-01 6.66734576e-01 7.65608490e-01
6.99284315e-01 8.74531329e-01 1.02645433e+00 5.42775989e-01
1.68577909e-01 -3.72917086e-01 3.47799480e-01 1.10802078e+00
-4.83212531e-01 -5.33953644e-02 -1.32024503e+00 2.66118437e-01
-1.94132447e+00 -6.56962693e-01 -2.11726189e-01 1.65444827e+00
3.85000318e-01 -5.39377965e-02 -2.03266606e-01 1.19261734e-01
5.08521438e-01 -4.68235105e-01 -6.65576085e-02 -3.10609847e-01
2.94882149e-01 5.31596057e-02 4.56596017e-01 4.11537476e-02
-6.78621173e-01 9.17050183e-01 7.40018988e+00 1.00893237e-01
-5.74355483e-01 7.77409673e-02 -4.30500835e-01 3.74595135e-01
-1.09806180e-01 5.01501441e-01 -2.91061550e-01 2.19987005e-01
6.06712162e-01 -8.98240507e-03 3.60165209e-01 1.42533779e+00
9.04235616e-02 -4.70665544e-01 -8.36243629e-01 5.22098601e-01
-4.15017754e-01 -1.58082843e+00 -3.34236830e-01 7.08122253e-02
8.18692923e-01 3.56367677e-01 -6.32895768e-01 -5.68364002e-02
1.33553541e+00 -1.92708164e-01 1.70191073e+00 4.64938104e-01
2.79288352e-01 -2.18381196e-01 2.64472395e-01 8.11154544e-01
-1.67617249e+00 -5.63630581e-01 1.37667075e-01 -6.62417054e-01
9.56187785e-01 2.90327787e-01 -6.46548212e-01 5.31831086e-01
9.29576874e-01 2.88485587e-01 5.19849919e-02 8.68068635e-01
-4.34362322e-01 -1.21706672e-01 -5.34801781e-01 -5.42364240e-01
4.43628132e-02 -2.97205716e-01 7.36472070e-01 3.82411301e-01
4.87287045e-01 2.66165167e-01 4.47762638e-01 2.70096719e-01
4.04598653e-01 -2.55010813e-01 -9.51670766e-01 5.20407736e-01
7.02789485e-01 1.19921935e+00 -1.03021276e+00 -1.46334067e-01
-7.59203136e-02 5.44409394e-01 -1.13556534e-01 -3.55968475e-01
-3.93804908e-01 -2.40390047e-01 5.01568735e-01 3.30280930e-01
6.01466477e-01 -1.03300178e+00 -4.07088041e-01 -3.85673374e-01
-2.34037027e-01 -2.41803795e-01 -2.51198739e-01 -1.02484381e+00
-6.63122594e-01 3.22620511e-01 -2.42292032e-01 -1.47095668e+00
-1.14991531e-01 -2.31830627e-01 -4.94261026e-01 -5.27761817e-01
-5.83290756e-01 -1.39882374e+00 -6.57610118e-01 3.65803778e-01
3.73565823e-01 -7.17232525e-02 1.12158096e+00 -1.32658303e-01
-3.35143618e-02 -5.21873653e-01 1.35407940e-01 -4.42781836e-01
-2.33075093e-03 -7.99687862e-01 3.75275970e-01 4.35927093e-01
-6.15728974e-01 3.33474725e-01 7.81481087e-01 -7.90874124e-01
-2.33397770e+00 -6.87953413e-01 1.10795714e-01 -5.92186570e-01
6.49917066e-01 -5.78050971e-01 1.01321608e-01 8.07949662e-01
-4.03468385e-02 -5.14660716e-01 2.94978142e-01 -1.03629418e-01
3.59506518e-01 -1.47374034e-01 -1.38467741e+00 5.25108695e-01
1.82997954e+00 1.58241689e-01 -5.62672019e-01 5.17881453e-01
7.30879009e-01 -5.47794878e-01 -1.08628082e+00 3.23598564e-01
8.04255366e-01 -9.41382289e-01 7.07526147e-01 2.75907159e-01
9.58447978e-02 -6.85780108e-01 -5.97957611e-01 -1.12119627e+00
-6.88649178e-01 -1.03104186e+00 1.15351379e-02 9.36305642e-01
6.54089600e-02 -8.20956588e-01 8.66637170e-01 5.19697666e-01
-1.11993122e+00 -5.27155042e-01 -1.14293504e+00 -8.72141957e-01
-5.43739378e-01 -3.40549082e-01 8.79398167e-01 4.48971897e-01
5.86655736e-01 2.41263986e-01 1.57132939e-01 3.99778247e-01
3.37293625e-01 1.67611111e-02 1.55251980e+00 -1.10556817e+00
-1.42591745e-01 -1.97725035e-02 -8.53309989e-01 -1.15202177e+00
-3.37326735e-01 -6.46999657e-01 7.22847402e-01 -2.07141948e+00
-5.70534766e-01 -1.27536428e+00 4.70817149e-01 7.08643675e-01
1.20510101e+00 -5.34671664e-01 -7.28789940e-02 5.45230150e-01
-9.35254753e-01 6.71602130e-01 1.14132774e+00 2.65160114e-01
-4.39248323e-01 1.84440404e-01 -2.64179319e-01 7.10885108e-01
6.48437023e-01 -2.54859209e-01 -3.62930208e-01 -6.45662189e-01
5.69376469e-01 2.81690240e-01 4.96518472e-03 -1.56574786e+00
5.91587245e-01 -2.26938307e-01 -2.62044430e-01 -7.93270826e-01
3.40650648e-01 -1.33806276e+00 8.23311627e-01 1.29938138e+00
4.69862849e-01 2.72363603e-01 -4.47848827e-01 5.91323137e-01
3.13599497e-01 1.27441198e-01 4.39848721e-01 -4.39873010e-01
-8.12477767e-01 -1.41859487e-01 -9.64304805e-01 -8.91664684e-01
1.73150659e+00 -5.18866479e-01 -2.59571522e-01 -1.82361186e-01
-5.04905045e-01 4.33259368e-01 1.29193532e+00 3.07507277e-01
5.24266362e-01 -1.27898705e+00 -2.41911337e-01 1.33637324e-01
-2.04121172e-01 2.67717421e-01 -4.81367745e-02 3.22025001e-01
-1.41688800e+00 2.14789942e-01 -4.54109967e-01 -7.60588825e-01
-6.11321270e-01 3.70658606e-01 2.04838201e-01 1.26756817e-01
-1.09159636e+00 5.12406111e-01 -2.65603513e-01 -8.08827877e-01
-1.04796052e-01 -2.70976841e-01 1.87898830e-01 -6.58198357e-01
1.16369911e-01 6.48905635e-01 8.49769078e-03 -4.07425135e-01
-6.22434258e-01 9.71205711e-01 8.67300749e-01 -1.60169899e-01
1.98286784e+00 -5.45999825e-01 -5.39692640e-01 1.23684853e-01
3.60895663e-01 -8.27450007e-02 -9.99086022e-01 3.14984649e-01
7.07847327e-02 -8.76088440e-02 -5.61172783e-01 -3.08857143e-01
-6.81481183e-01 -3.87120962e-01 3.60395283e-01 6.60785854e-01
9.14150298e-01 7.43409097e-01 7.86228836e-01 7.95625150e-01
1.97854578e+00 -1.36545026e+00 3.76534551e-01 5.61578035e-01
1.10026169e+00 -3.52586538e-01 -7.33608380e-02 -5.14011741e-01
-5.71495108e-02 1.14397562e+00 2.69315332e-01 -5.82117319e-01
9.18321550e-01 9.01164353e-01 -1.24679528e-01 -9.80548859e-01
-9.80811477e-01 1.00468695e-02 -1.01599920e+00 1.14096594e+00
-4.24664199e-01 1.55623630e-01 -9.69365016e-02 3.29068124e-01
-3.79750460e-01 -2.63603684e-02 6.83986902e-01 2.12668729e+00
-1.26181555e+00 -1.06984508e+00 -5.99698961e-01 -1.44601882e-01
6.45132065e-01 9.49826837e-01 -2.99117953e-01 8.05561483e-01
4.23153073e-01 1.01381278e+00 9.19909030e-03 -7.90805042e-01
7.38776386e-01 -5.00689447e-01 4.84452337e-01 -7.68317282e-01
-3.58760923e-01 -4.01958346e-01 4.95248407e-01 -9.42280889e-01
-3.26174021e-01 -6.17474139e-01 -1.69928765e+00 -1.91889212e-01
-1.56617612e-01 -5.14421538e-02 1.09064960e+00 6.04027152e-01
5.16859531e-01 4.03075777e-02 1.29217887e+00 -1.57217777e+00
-2.36835405e-01 -5.22494256e-01 -8.21680188e-01 -1.10833876e-01
-3.71652782e-01 -9.41988587e-01 -8.33924338e-02 1.00882445e-03] | [4.8242292404174805, 0.87785804271698] |
6faf8e1b-93aa-4436-b340-dfd66631209e | one-shot-learning-based-drivers-head-movement | 2306.05291 | null | https://arxiv.org/abs/2306.05291v1 | https://arxiv.org/pdf/2306.05291v1.pdf | One shot learning based drivers head movement identification using a millimetre wave radar sensor | Concentration of drivers on traffic is a vital safety issue; thus, monitoring a driver being on road becomes an essential requirement. The key purpose of supervision is to detect abnormal behaviours of the driver and promptly send warnings to him her for avoiding incidents related to traffic accidents. In this paper, to meet the requirement, based on radar sensors applications, the authors first use a small sized millimetre wave radar installed at the steering wheel of the vehicle to collect signals from different head movements of the driver. The received signals consist of the reflection patterns that change in response to the head movements of the driver. Then, in order to distinguish these different movements, a classifier based on the measured signal of the radar sensor is designed. However, since the collected data set is not large, in this paper, the authors propose One shot learning to classify four cases of driver's head movements. The experimental results indicate that the proposed method can classify the four types of cases according to the various head movements of the driver with a high accuracy reaching up to 100. In addition, the classification performance of the proposed method is significantly better than that of the convolutional neural network model. | ['Yong Hwa Kim', 'Tien Tung Nguyen', 'Seongwook Lee', 'Hong Nhung Nguyen'] | 2023-05-31 | null | null | null | null | ['one-shot-learning'] | ['methodology'] | [ 2.20722705e-01 -2.64977902e-01 1.20497495e-01 -4.89531219e-01
-1.25788167e-01 1.51970237e-01 2.86042690e-01 -3.95941168e-01
-6.08716488e-01 4.16534215e-01 -1.21634662e-01 -3.62458915e-01
-3.47452819e-01 -7.62266040e-01 -1.05569750e-01 -1.03726006e+00
5.04075170e-01 -1.07472427e-01 4.81900901e-01 -3.24667990e-01
4.86819267e-01 7.84151137e-01 -2.06979942e+00 7.03154951e-02
6.56158030e-01 1.18748462e+00 1.85904250e-01 4.10462290e-01
4.46380582e-03 6.39219105e-01 -5.80218613e-01 2.38448039e-01
1.42601356e-01 -3.78540576e-01 1.85035244e-01 4.74204607e-02
1.18909568e-01 -2.44011879e-01 -4.27593857e-01 1.12168276e+00
5.39580464e-01 2.31836766e-01 5.56973279e-01 -1.14131784e+00
8.74882340e-02 6.62796106e-03 -6.55907512e-01 6.09482408e-01
1.31688798e-02 2.03796625e-01 3.63292396e-01 -5.86036205e-01
-7.68380240e-02 7.82753468e-01 2.67424852e-01 5.89080036e-01
-4.62337703e-01 -1.18729186e+00 -6.57772347e-02 1.03410697e+00
-1.20117009e+00 -7.53798485e-01 1.04484284e+00 -4.75070685e-01
2.84835786e-01 1.70726761e-01 4.50367451e-01 8.07431996e-01
6.25902057e-01 3.51980984e-01 1.08610857e+00 -2.45249376e-01
2.14148521e-01 3.50663453e-01 7.83125699e-01 3.93361628e-01
3.70998889e-01 4.91423577e-01 -2.96132177e-01 3.37937593e-01
-7.69326314e-02 3.16985756e-01 -3.15846056e-01 1.77034348e-01
-6.48102701e-01 6.62876129e-01 2.39245862e-01 4.77159053e-01
-6.75743878e-01 -2.89914995e-01 2.39696607e-01 2.73822218e-01
1.23532414e-01 -7.02618584e-02 -3.35292742e-02 -1.63003370e-01
-5.80701888e-01 -8.21579546e-02 6.16272748e-01 4.98652041e-01
6.06529415e-01 1.20068237e-01 -9.04858932e-02 4.33679581e-01
2.32416376e-01 7.42486656e-01 3.37177843e-01 -3.28278124e-01
3.27120781e-01 5.00825107e-01 -9.10146907e-02 -1.21015370e+00
-6.97875917e-01 -5.47414422e-01 -9.52881634e-01 4.07832026e-01
2.16632068e-01 -4.87175137e-01 -8.92798007e-01 1.12465346e+00
2.39356235e-01 1.80308804e-01 2.25683272e-01 8.95742238e-01
1.07113373e+00 7.57553279e-01 -7.30126053e-02 -4.79679704e-01
1.42818761e+00 -3.68253976e-01 -1.15301156e+00 -5.28203428e-01
3.65025282e-01 -4.95917648e-01 3.55925590e-01 4.50280190e-01
-4.19758230e-01 -8.18575561e-01 -1.15576780e+00 7.53097832e-01
-2.90450096e-01 3.41651142e-01 1.92449018e-01 7.68079042e-01
-2.65864074e-01 -8.43156949e-02 -3.83833766e-01 -2.70648718e-01
2.19748363e-01 3.48484889e-02 3.69390182e-04 -2.20684573e-01
-1.33764458e+00 1.06208444e+00 9.73630100e-02 6.30125642e-01
-2.98674881e-01 -5.22718132e-01 -7.13408768e-01 5.60826296e-03
2.74927229e-01 5.88097784e-04 9.85974491e-01 -5.26646018e-01
-1.17241991e+00 5.02459526e-01 -3.22529674e-01 -3.65886450e-01
3.22904468e-01 2.02998310e-01 -1.07807505e+00 -1.92989465e-02
2.58464850e-02 -1.12208072e-03 9.45254087e-01 -8.25106025e-01
-1.18039429e+00 -5.93142807e-01 -4.44996625e-01 -1.45663187e-01
-1.10094294e-01 1.25898749e-01 3.89353223e-02 3.41783613e-01
-3.43815163e-02 -8.34879458e-01 2.57910825e-02 -6.04601741e-01
-1.92932233e-01 -4.48334545e-01 1.29736114e+00 -4.10748392e-01
1.11535406e+00 -2.58617067e+00 -6.19858503e-01 3.18346083e-01
1.80119157e-01 5.73796988e-01 1.00952953e-01 3.30598243e-02
-1.98163521e-02 -7.64469504e-01 -1.01256762e-02 4.12938416e-01
-3.62642586e-01 1.78859066e-02 -1.82199687e-01 6.35070145e-01
1.58106819e-01 4.40592498e-01 -3.16273034e-01 -1.99827045e-01
3.97756815e-01 3.40203553e-01 1.23602465e-01 3.54363084e-01
4.10866737e-01 4.50856745e-01 -6.77843809e-01 2.37270206e-01
8.63280833e-01 4.14749593e-01 -3.95767659e-01 -2.20332906e-01
-5.29641032e-01 8.69492535e-03 -1.02333295e+00 4.67200518e-01
-3.57682049e-01 1.02009904e+00 1.21724539e-01 -1.12819779e+00
1.49997509e+00 3.71425122e-01 3.62633497e-01 -1.29133463e+00
4.24517274e-01 -1.23535711e-02 5.74145675e-01 -9.98772442e-01
2.63226274e-02 -2.82645166e-01 -1.14070252e-01 1.38916984e-01
-6.33061647e-01 4.04990524e-01 1.66116968e-01 -3.03556979e-01
1.19101584e+00 -5.75714350e-01 7.31471851e-02 1.12941511e-01
8.59671533e-01 -1.89517081e-01 7.47754276e-01 4.95162368e-01
-3.71893227e-01 -2.03829065e-01 2.65351921e-01 -5.42234302e-01
-3.05721968e-01 -6.38284266e-01 -2.44985133e-01 7.95703948e-01
3.47201705e-01 3.73685986e-01 -5.66128969e-01 -2.01944321e-01
-4.51463759e-02 1.00760365e+00 -4.10724610e-01 -6.02372348e-01
-5.93708873e-01 -6.90950572e-01 2.61262357e-01 3.15732926e-01
8.63549054e-01 -1.35376263e+00 -1.43733335e+00 7.14380518e-02
-2.29148455e-02 -1.21081924e+00 4.95611168e-02 -1.61975101e-02
-2.73163140e-01 -1.25805950e+00 -2.90705860e-01 -8.06763232e-01
4.30824578e-01 7.24363387e-01 2.28853479e-01 -6.74607828e-02
-6.53004944e-01 9.00435895e-02 -2.66834676e-01 -9.79910791e-01
-3.28907251e-01 -9.37255546e-02 1.06727667e-02 7.16813743e-01
9.91130590e-01 -2.84532398e-01 -4.72284913e-01 3.69976223e-01
-4.13544148e-01 -1.91109493e-01 1.03464770e+00 1.58132404e-01
-8.19310024e-02 6.66577995e-01 9.49162722e-01 -6.49221778e-01
7.14048684e-01 -3.77533942e-01 -7.34318674e-01 -1.68764040e-01
-4.56885159e-01 -2.31587961e-01 6.00561976e-01 -1.93649516e-01
-1.19753206e+00 -2.04288159e-02 -1.53109074e-01 -1.42340675e-01
-8.08961511e-01 2.38683328e-01 -3.50020170e-01 7.07132071e-02
4.28879797e-01 3.92901123e-01 3.50741923e-01 -2.65026122e-01
-8.69993940e-02 1.13984156e+00 5.29826939e-01 5.38383245e-01
7.08454013e-01 1.91559240e-01 3.08636099e-01 -1.45485044e+00
-7.94813514e-01 -8.36805105e-01 -3.79171818e-01 -8.02816093e-01
9.78853822e-01 -5.49468875e-01 -1.15043044e+00 6.34826660e-01
-1.04089725e+00 4.34486479e-01 2.18772247e-01 8.66353035e-01
-2.14373320e-01 1.05828710e-01 -8.29021335e-02 -1.20720136e+00
-3.62621188e-01 -9.02323127e-01 4.92461890e-01 6.25721335e-01
5.34945391e-02 -5.14663041e-01 -2.29393318e-01 4.07695353e-01
5.67051947e-01 4.33659516e-02 7.48995721e-01 -5.52262962e-01
-3.97252083e-01 -7.38211513e-01 -1.65378258e-01 1.41680583e-01
4.06735986e-01 -1.50625378e-01 -1.01130688e+00 -3.05404589e-02
4.45563465e-01 2.55912304e-01 8.41388226e-01 5.28488576e-01
8.52645695e-01 5.55207878e-02 -4.99343157e-01 3.61718208e-01
1.05976450e+00 9.48969901e-01 1.01236749e+00 1.58084631e-01
3.48483741e-01 1.04254055e+00 9.41443563e-01 2.76487947e-01
-3.19034420e-02 3.82769018e-01 4.97148126e-01 -2.21409649e-02
1.45686924e-01 2.98061311e-01 4.59089339e-01 6.02607131e-01
3.59558538e-02 -5.04040755e-02 -5.26658237e-01 3.88165534e-01
-1.61718309e+00 -1.44286335e+00 -7.39099860e-01 2.08996892e+00
1.78386554e-01 3.51653486e-01 -1.10978016e-03 6.00269496e-01
9.29672778e-01 1.40765682e-01 -5.27306676e-01 -4.43242222e-01
5.06599665e-01 1.51293650e-01 5.30682385e-01 3.40566874e-01
-8.24001610e-01 3.18790525e-01 5.16184616e+00 4.23797816e-01
-1.48076999e+00 -2.60723501e-01 7.53218587e-03 -3.08745354e-02
2.14230299e-01 -3.49615067e-01 -1.21006835e+00 7.97020793e-01
1.13191772e+00 -6.61087409e-02 -7.37747550e-02 4.72747147e-01
6.67192936e-01 -4.22008127e-01 -5.88875532e-01 1.11438143e+00
2.19689339e-01 -6.60821736e-01 -5.33462942e-01 1.56074584e-01
-5.59392460e-02 -3.09297681e-01 1.88160557e-02 3.28472883e-01
-3.78878981e-01 -7.85994112e-01 1.53319329e-01 9.21652138e-01
3.72017592e-01 -1.16771293e+00 9.98582721e-01 9.50179875e-01
-1.02721739e+00 -4.35679108e-01 -3.73557687e-01 -2.70063847e-01
3.58094335e-01 8.20869088e-01 -8.44341874e-01 1.05368882e-01
5.97444236e-01 5.54063976e-01 -3.06553781e-01 1.06007826e+00
-3.28262597e-01 8.09022844e-01 2.16138344e-02 -4.01522726e-01
8.76024142e-02 -2.60276377e-01 6.51027024e-01 1.08815038e+00
4.34866518e-01 3.64868701e-01 -1.30430549e-01 7.39366472e-01
4.64005768e-01 -1.65986091e-01 -1.04517782e+00 2.41620198e-01
2.65304893e-01 1.52074027e+00 -2.67061114e-01 -4.73204302e-03
-3.21329743e-01 2.96354175e-01 -3.50340605e-01 2.51760364e-01
-8.25220346e-01 -1.09540975e+00 5.77418387e-01 4.56186116e-01
3.33841443e-01 1.85572598e-02 -7.54415616e-02 -2.25937977e-01
2.02467173e-01 -2.49020711e-01 3.88669707e-02 -6.59443557e-01
-8.77993107e-01 4.34497684e-01 -8.07098597e-02 -1.34322107e+00
-2.25780755e-01 -5.59937418e-01 -1.10206759e+00 1.02221525e+00
-1.67243636e+00 -5.39740801e-01 -8.29761863e-01 5.11888087e-01
4.82639104e-01 -3.84693980e-01 3.44205111e-01 3.27820092e-01
-7.80806541e-01 4.05767471e-01 -3.46764892e-01 1.44918188e-01
5.64083517e-01 -3.94514948e-01 -8.21205899e-02 8.62497866e-01
-5.58322966e-01 1.97087064e-01 7.60473311e-01 -5.65925181e-01
-1.23100102e+00 -1.07500148e+00 1.03390706e+00 1.73086286e-01
2.75055557e-01 -1.04934499e-01 -9.32584107e-01 2.44700953e-01
2.06900373e-01 -8.88280645e-02 5.59439480e-01 -2.89735794e-01
3.11234713e-01 -6.88282192e-01 -1.07287824e+00 2.73751050e-01
5.05931914e-01 -1.18637487e-01 -8.80843282e-01 -8.54296312e-02
9.55806747e-02 4.41628471e-02 -2.62332708e-01 4.91086632e-01
5.08648634e-01 -1.09524846e+00 3.73684376e-01 -3.45247388e-01
-3.92884947e-02 -3.97322416e-01 2.37096980e-01 -1.30562043e+00
-4.64703470e-01 -1.38267219e-01 3.03067982e-01 8.75905693e-01
1.82787061e-01 -1.14967704e+00 6.64456606e-01 3.19203138e-01
-1.18054867e-01 -4.36159641e-01 -9.75474358e-01 -4.30896550e-01
-5.88910341e-01 -4.29435700e-01 3.46396446e-01 4.22143489e-01
-1.96770318e-02 6.74228370e-01 -2.03164622e-01 4.14453268e-01
7.71369100e-01 -1.25493677e-02 8.31944406e-01 -1.59325647e+00
3.14752698e-01 -2.11724058e-01 -7.63609469e-01 -8.76368046e-01
1.75306588e-01 -5.91769218e-01 5.36840856e-01 -1.51795173e+00
-1.50832444e-01 -2.25327030e-01 -3.60479981e-01 6.22386001e-02
3.77609916e-02 -1.14896797e-01 -1.57957941e-01 -1.52453437e-01
-6.54749349e-02 4.00010765e-01 1.01518655e+00 -1.25411943e-01
-1.82431698e-01 8.59884739e-01 -6.37846708e-01 7.03472614e-01
9.65762079e-01 -2.50247121e-01 -4.13394243e-01 4.60455008e-02
-9.69394892e-02 2.77993172e-01 4.64073658e-01 -1.40628636e+00
6.14106297e-01 -2.28729129e-01 2.66972393e-01 -1.09713590e+00
3.78876984e-01 -1.05680621e+00 -9.85826701e-02 5.84142447e-01
-7.67871961e-02 -2.81104326e-01 1.50248423e-01 6.13739252e-01
-1.45732597e-01 -2.84590900e-01 9.55774665e-01 2.49780044e-01
-9.27744448e-01 4.58728708e-02 -9.37377930e-01 -3.60558897e-01
1.43567157e+00 -4.33853686e-01 -4.17077184e-01 -3.81606758e-01
-2.95087159e-01 2.52435148e-01 -4.22190130e-01 5.56222856e-01
9.53049839e-01 -1.10789669e+00 -6.45747244e-01 7.66781986e-01
2.71566361e-01 -2.63226181e-01 4.69008625e-01 1.10821819e+00
5.08131459e-02 7.54090130e-01 -4.14609462e-01 -5.84798396e-01
-1.55526328e+00 3.68769735e-01 4.31690246e-01 4.70095366e-01
-8.12605441e-01 2.41130397e-01 7.59418681e-02 1.31228611e-01
1.47031590e-01 -9.60155204e-03 -9.67817485e-01 2.79864728e-01
8.96118939e-01 5.81229329e-01 2.54807651e-01 -6.92034602e-01
-4.24078137e-01 7.50909269e-01 -1.48738623e-01 1.76244155e-01
1.05788279e+00 4.08555605e-02 8.70191380e-02 4.18299884e-01
1.12623954e+00 6.38751909e-02 -9.56805885e-01 -2.29169950e-01
-2.19673470e-01 -4.43598866e-01 2.30075479e-01 -4.51119453e-01
-1.20203578e+00 1.19267869e+00 9.16865945e-01 3.48495692e-01
1.20852649e+00 -1.34024113e-01 9.28244293e-01 4.10943687e-01
3.20423841e-01 -1.07238638e+00 -3.97144824e-01 3.25609922e-01
4.75927234e-01 -1.09826016e+00 -2.90478259e-01 -2.83746600e-01
-6.41274393e-01 1.10922563e+00 6.33584857e-01 6.28133789e-02
7.87093282e-01 1.84897244e-01 4.00226504e-01 -3.32823992e-01
-6.98853016e-01 -5.34650445e-01 3.18154782e-01 7.41176605e-01
9.45972279e-04 1.04522608e-01 -3.50827217e-01 3.42464328e-01
-1.84714898e-01 3.08146253e-02 5.40509820e-01 7.04811871e-01
-1.19822836e+00 -4.38925445e-01 -5.10152519e-01 6.53700888e-01
-1.81982443e-01 4.87690806e-01 -1.78939134e-01 5.55010498e-01
2.52110779e-01 1.54095995e+00 3.44032377e-01 -5.54653049e-01
7.73684800e-01 1.11312307e-01 4.57285754e-02 -2.87984878e-01
1.25199541e-01 -3.61105829e-01 -4.46350947e-02 -2.11423814e-01
-3.12043577e-01 -4.50621337e-01 -1.40325606e+00 -9.96026918e-02
-1.75232857e-01 2.38057390e-01 7.33653724e-01 1.20661843e+00
8.28059316e-02 7.68343687e-01 1.14802349e+00 -4.20396596e-01
-6.09430552e-01 -1.05781317e+00 -7.77594030e-01 2.69513041e-01
5.13609588e-01 -8.34177852e-01 -7.09041417e-01 -4.17266965e-01] | [7.999454975128174, -0.7330774068832397] |
06d2c7e5-97a5-4efb-b9eb-f1ca6e382295 | invalidator-automated-patch-correctness | 2301.01113 | null | https://arxiv.org/abs/2301.01113v2 | https://arxiv.org/pdf/2301.01113v2.pdf | Invalidator: Automated Patch Correctness Assessment via Semantic and Syntactic Reasoning | Automated program repair (APR) faces the challenge of test overfitting, where generated patches pass validation tests but fail to generalize. Existing methods for patch assessment involve generating new tests or manual inspection, which can be time-consuming or biased. In this paper, we propose a novel technique, INVALIDATOR, to automatically assess the correctness of APR-generated patches via semantic and syntactic reasoning. INVALIDATOR leverages program invariants to reason about program semantics while also capturing program syntax through language semantics learned from a large code corpus using a pre-trained language model. Given a buggy program and the developer-patched program, INVALIDATOR infers likely invariants on both programs. Then, INVALIDATOR determines that an APR-generated patch overfits if: (1) it violates correct specifications or (2) maintains erroneous behaviors from the original buggy program. In case our approach fails to determine an overfitting patch based on invariants, INVALIDATOR utilizes a trained model from labeled patches to assess patch correctness based on program syntax. The benefit of INVALIDATOR is threefold. First, INVALIDATOR leverages both semantic and syntactic reasoning to enhance its discriminative capability. Second, INVALIDATOR does not require new test cases to be generated, but instead only relies on the current test suite and uses invariant inference to generalize program behaviors. Third, INVALIDATOR is fully automated. Experimental results demonstrate that INVALIDATOR outperforms existing methods in terms of Accuracy and F-measure, correctly identifying 79% of overfitting patches and detecting 23% more overfitting patches than the best baseline. | ['Quyet-Thang Huynh', 'Bui Quang-Huy', 'Nhat-Hoa Tran', 'David Lo', 'Xuan Bach D. Le', 'Duc-Minh Luong', 'Thanh Le-Cong'] | 2023-01-03 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [ 2.42302611e-01 1.29720539e-01 -5.90147257e-01 -2.14162394e-01
-1.19907749e+00 -8.95584464e-01 -1.01527400e-01 3.98429841e-01
6.08556390e-01 4.21873927e-01 -2.04730377e-01 -7.81538785e-01
2.85615921e-01 -9.27081943e-01 -1.15522194e+00 1.02350004e-01
-7.13139102e-02 4.53098044e-02 5.93976319e-01 7.77349994e-02
4.46343035e-01 -1.43242076e-01 -1.80442381e+00 6.42740965e-01
1.45673907e+00 6.06248200e-01 -5.34089953e-02 8.10869634e-01
9.85120982e-02 7.50345051e-01 -7.76056767e-01 -3.19400549e-01
8.09243545e-02 -7.14785755e-01 -9.37789202e-01 1.72815286e-02
6.91903949e-01 -2.58916557e-01 4.04678673e-01 1.34094417e+00
-2.04228789e-01 -4.73698825e-01 6.47518337e-02 -1.40430939e+00
-4.60160971e-01 6.23537660e-01 -4.10130471e-01 -1.99476127e-02
7.25085139e-01 4.44160521e-01 1.24559498e+00 -5.93640745e-01
4.83005375e-01 8.63937736e-01 9.78949368e-01 4.31960166e-01
-1.58884585e+00 -3.76730502e-01 4.29876670e-02 -3.38914841e-01
-1.24567020e+00 -8.14006180e-02 6.12402380e-01 -6.54504538e-01
1.36731851e+00 5.04427314e-01 4.16613132e-01 8.03125679e-01
4.36138153e-01 3.52725834e-01 1.05130374e+00 -4.39493716e-01
5.01012623e-01 1.31975576e-01 5.63304007e-01 1.13068485e+00
6.27779663e-01 1.74603701e-01 -1.13374190e-02 -1.01268077e+00
2.76495237e-02 -1.32022977e-01 -2.47021899e-01 -1.55670136e-01
-8.94343913e-01 6.08429074e-01 1.58047929e-01 2.83235580e-01
7.68756866e-03 2.60563314e-01 5.19761443e-01 4.63683158e-01
1.04862101e-01 1.07542074e+00 -7.31347263e-01 -3.89680296e-01
-1.02852488e+00 3.69282126e-01 1.00509131e+00 1.05948937e+00
1.32155490e+00 4.54687215e-02 -1.54983466e-02 4.61171180e-01
2.86396444e-01 6.38120711e-01 5.41846633e-01 -5.78616977e-01
4.97283429e-01 1.52467287e+00 -1.67083248e-01 -1.04498720e+00
-6.81806281e-02 -3.98438722e-01 1.89906254e-01 2.62484908e-01
-7.17456043e-02 2.32965261e-01 -7.28258848e-01 1.54239058e+00
1.32573023e-01 1.18230321e-02 -1.05479442e-01 4.84672666e-01
3.76140356e-01 2.75150299e-01 -3.25287938e-01 1.30695626e-01
9.68916535e-01 -7.55016565e-01 1.29891351e-01 -6.19302094e-01
1.14874816e+00 -5.98351836e-01 1.49720979e+00 3.06168765e-01
-7.36383557e-01 -3.67228478e-01 -1.33189142e+00 6.30578578e-01
-4.47393879e-02 2.00318471e-01 3.88648927e-01 6.55057013e-01
-1.05877221e+00 5.60720265e-01 -9.19977307e-01 -1.24385454e-01
2.43016362e-01 2.24031046e-01 -3.41819227e-01 -1.48182020e-01
-4.31079566e-01 2.50666559e-01 3.34347188e-01 -4.50502157e-01
-1.10716069e+00 -8.32485080e-01 -1.20369446e+00 2.63769597e-01
5.80493748e-01 -3.83408844e-01 1.40126300e+00 -1.38322592e+00
-7.82835782e-01 4.88846809e-01 -3.18852991e-01 -1.28664926e-01
4.12088595e-02 -7.83568621e-02 -4.70715731e-01 -1.07506827e-01
4.99750912e-01 -1.28996044e-01 7.33093143e-01 -1.29168844e+00
-5.80349803e-01 -2.86869377e-01 3.42585742e-01 -8.38859260e-01
-1.52899176e-01 -2.09205851e-01 -2.87230045e-01 -4.34929103e-01
5.41524291e-02 -1.07899702e+00 6.84548169e-02 -5.06260455e-01
-6.75437212e-01 1.44960999e-01 8.90398443e-01 -8.08425069e-01
1.69067860e+00 -2.28212738e+00 -2.16428027e-01 5.81220984e-01
3.61674041e-01 1.24807708e-01 -3.83475602e-01 3.58925641e-01
-1.25082165e-01 5.75936496e-01 -6.84760213e-01 3.49386245e-01
1.52879998e-01 1.52760565e-01 -5.99300802e-01 2.29980126e-01
5.78308582e-01 8.54601443e-01 -1.02757740e+00 -2.95956075e-01
-3.63788188e-01 -1.64633274e-01 -1.19011152e+00 1.64732531e-01
-9.32389855e-01 -6.51154965e-02 -4.27016377e-01 1.14406347e+00
5.57246208e-01 -1.82488874e-01 1.38467327e-01 1.88054711e-01
6.19542077e-02 5.57868123e-01 -8.66547048e-01 1.26121402e+00
-5.53007960e-01 3.64882767e-01 -4.05184209e-01 -7.24774420e-01
1.01791561e+00 7.58759817e-03 -6.17345907e-02 -4.45470214e-01
-4.14297849e-01 7.10845828e-01 -1.72412544e-01 -9.68196452e-01
2.12960765e-01 1.73757315e-01 -5.45681834e-01 8.68766487e-01
5.02174050e-02 -1.29885525e-01 2.08837077e-01 -1.36415549e-02
2.03128433e+00 4.49195921e-01 3.39824289e-01 -1.27605006e-01
4.59391475e-01 4.76930588e-01 1.02526104e+00 8.84321868e-01
1.12683728e-01 3.52657497e-01 1.07972670e+00 -3.01316231e-01
-9.07032669e-01 -9.86670434e-01 1.72301203e-01 7.94761658e-01
-6.02409393e-02 -9.11108911e-01 -8.58487427e-01 -1.41221011e+00
1.86261863e-01 9.02173400e-01 -6.25915229e-01 -6.58032119e-01
-5.74943662e-01 -2.94333339e-01 7.56516755e-01 6.68581784e-01
2.57779211e-01 -7.42370188e-01 -6.79188967e-01 5.54556623e-02
-1.64944574e-01 -4.64321882e-01 -6.63899958e-01 -6.59515336e-02
-8.54659677e-01 -1.77398825e+00 4.61703211e-01 -6.77213907e-01
1.13991189e+00 9.69505087e-02 1.24752176e+00 8.73127580e-01
-4.75005358e-01 4.70825940e-01 -5.03022552e-01 1.97885588e-01
-1.03163671e+00 8.24647695e-02 -4.42713678e-01 -3.40763181e-01
4.40854818e-01 -4.53068107e-01 2.76141893e-02 5.22753060e-01
-1.06949842e+00 -5.03735065e-01 5.78526139e-01 9.82204080e-01
5.45928240e-01 2.85310328e-01 4.13927406e-01 -1.17593610e+00
4.39148158e-01 -5.32345653e-01 -9.80962753e-01 4.33623880e-01
-7.95947731e-01 3.57510149e-01 8.23027670e-01 -4.73910928e-01
-8.77338469e-01 8.65828842e-02 1.44125959e-02 -2.52540559e-01
-1.11138776e-01 9.52534676e-01 -2.26656377e-01 -1.05960734e-01
1.10250139e+00 2.07076341e-01 -1.74666986e-01 -2.60629296e-01
-3.20970863e-01 4.33241993e-01 5.58639884e-01 -1.10757375e+00
9.67021883e-01 -8.37931633e-02 -3.87352794e-01 -1.40248463e-01
-5.13452172e-01 -1.39402121e-01 -2.39927545e-01 -4.57556956e-02
3.33829552e-01 -4.88636434e-01 -3.97138566e-01 8.69220942e-02
-1.00937080e+00 -4.68429506e-01 -1.49193332e-01 1.02063239e-01
-3.39054555e-01 3.64633560e-01 -3.18583995e-01 -6.45436883e-01
-1.20815858e-01 -1.46963072e+00 1.30622518e+00 -8.65643695e-02
-6.64141774e-01 -7.28394389e-01 3.97780269e-01 2.00653225e-01
4.05707657e-01 5.01460791e-01 1.48194146e+00 -6.61127269e-01
-6.85062647e-01 -7.52042949e-01 2.45216489e-01 4.11484689e-01
2.95423239e-01 4.25235540e-01 -8.22883964e-01 -4.37245578e-01
-3.89372185e-02 -2.77511150e-01 2.15408862e-01 -3.51658434e-01
8.73874605e-01 -7.13630378e-01 -4.99662429e-01 3.58162314e-01
1.65241706e+00 4.01148514e-04 4.34170336e-01 3.15288603e-01
4.50491488e-01 4.20269221e-01 6.27477348e-01 2.78923959e-01
2.38784835e-01 4.40329850e-01 6.46448314e-01 3.34289044e-01
-2.31742468e-02 -5.76347172e-01 9.36752260e-01 4.89651620e-01
4.66326773e-01 3.76198679e-01 -1.38402057e+00 6.75358713e-01
-1.70059395e+00 -9.10442352e-01 -2.18948409e-01 2.57347178e+00
1.01739168e+00 2.16792330e-01 1.52177483e-01 1.29872397e-01
5.30627728e-01 -4.69716638e-01 -6.74039900e-01 -5.51356673e-01
2.41976887e-01 2.65877187e-01 4.21708785e-02 6.13823950e-01
-7.12275624e-01 6.96424901e-01 5.61721706e+00 3.08644950e-01
-1.23216057e+00 9.97338519e-02 1.58677071e-01 2.65953839e-01
-7.91398704e-01 7.68855929e-01 -7.08294868e-01 3.66490126e-01
7.72993386e-01 -3.05942535e-01 4.51549143e-01 1.37583113e+00
-2.66218215e-01 -3.64875913e-01 -1.51961339e+00 5.52820675e-02
4.68581021e-01 -1.08697569e+00 -3.24599564e-01 5.48197813e-02
8.33405554e-01 -5.75117096e-02 -1.36149645e-01 7.71290064e-01
2.35307783e-01 -7.68205583e-01 7.63612807e-01 4.45757627e-01
4.82087255e-01 -5.43765068e-01 8.53141129e-01 3.71571183e-01
-1.05371523e+00 -1.54684588e-01 -9.71184596e-02 -5.20620774e-03
-5.95513821e-01 6.77541494e-01 -1.18071485e+00 3.23945612e-01
7.08986819e-01 5.23910522e-01 -1.33708489e+00 1.11330235e+00
-4.87075061e-01 9.91989076e-01 -3.76371294e-02 6.87822476e-02
-1.24011897e-01 2.56895214e-01 6.00214601e-01 1.10924995e+00
4.30398405e-01 -6.55622840e-01 6.42619193e-01 1.54668176e+00
2.26738304e-01 -1.42848387e-01 -8.69270384e-01 -2.33446687e-01
4.91998225e-01 9.40775931e-01 -4.18509632e-01 -2.83727407e-01
-5.19036174e-01 5.18589437e-01 2.76845664e-01 3.27720821e-01
-9.63366747e-01 -5.74021816e-01 5.45965075e-01 1.99651450e-01
3.81927550e-01 1.21144265e-01 -3.88565570e-01 -1.21298993e+00
7.55874991e-01 -1.31506467e+00 2.86571413e-01 -6.34790719e-01
-8.72865915e-01 7.28419244e-01 -1.21076494e-01 -1.23406196e+00
-3.21985155e-01 -3.35568428e-01 -8.73462975e-01 5.96923947e-01
-1.06089389e+00 -9.60651100e-01 -3.77269715e-01 1.94993764e-02
3.44651610e-01 4.20569815e-02 8.65463197e-01 1.24058230e-02
-5.99830508e-01 9.24200535e-01 -6.35803401e-01 -6.02088980e-02
5.76117814e-01 -1.18796659e+00 4.41923767e-01 1.33544683e+00
-2.05207825e-01 9.66315329e-01 7.00090408e-01 -1.08524048e+00
-1.73719215e+00 -1.56443381e+00 8.08388591e-01 -6.09779060e-01
8.85866404e-01 -3.28691602e-01 -1.26785564e+00 9.24952209e-01
-3.80695492e-01 1.40719935e-01 7.17465580e-01 7.97482207e-02
-1.10839200e+00 1.14008784e-02 -1.18324983e+00 4.30661410e-01
6.52451813e-01 -8.04697573e-01 -6.38780296e-01 1.52627990e-01
7.28774011e-01 -2.30150804e-01 -9.32468951e-01 5.04920065e-01
2.69691437e-01 -1.03880787e+00 2.88414299e-01 -5.36494255e-01
5.50193489e-01 -8.31847966e-01 -3.52271080e-01 -1.01770866e+00
-4.45934944e-02 -4.85353053e-01 -1.10540137e-01 1.27911615e+00
8.94069731e-01 -8.89359534e-01 5.10254323e-01 6.88232183e-01
-4.43513602e-01 -7.80265749e-01 -5.42021692e-01 -1.05470955e+00
-2.36012354e-01 -6.33437395e-01 9.69903529e-01 9.09869075e-01
6.05843067e-01 5.80049418e-02 2.31044188e-01 5.64042270e-01
2.79133499e-01 5.94837070e-01 1.04137874e+00 -9.36281860e-01
-9.56105828e-01 -4.40033466e-01 -5.56667745e-01 -3.68629724e-01
5.03481925e-01 -1.10317099e+00 4.08953488e-01 -9.18913007e-01
4.24035430e-01 -3.83022755e-01 1.51546881e-01 1.16428506e+00
-4.24986273e-01 9.30725858e-02 -3.88746947e-01 9.21401083e-02
-5.79450488e-01 2.41136178e-02 4.80359137e-01 -4.40190494e-01
-3.08737487e-01 6.75311014e-02 -7.98896492e-01 5.97245574e-01
7.57830083e-01 -6.77326262e-01 -2.89412528e-01 -2.15701446e-01
7.59204388e-01 1.88895673e-01 7.44164348e-01 -1.08911300e+00
-1.30618647e-01 -1.18912376e-01 -1.69366196e-01 -5.26643284e-02
-5.77607989e-01 -5.81978142e-01 3.62521827e-01 7.97034383e-01
-1.62567839e-01 4.27440166e-01 4.87411261e-01 5.54557145e-01
-3.70701134e-01 -7.29561508e-01 3.66099238e-01 4.99695428e-02
-5.12811303e-01 -1.07307702e-01 -3.91439110e-01 1.80796996e-01
1.12626672e+00 -2.12927118e-01 -6.62140667e-01 3.37346822e-01
-2.90194809e-01 -4.77002040e-02 1.24425495e+00 5.03424644e-01
5.95496178e-01 -1.11505485e+00 -3.44776541e-01 6.47216678e-01
7.28387952e-01 -3.94170374e-01 -2.09871799e-01 8.34773898e-01
-4.58477378e-01 2.60187723e-02 3.32552016e-01 -8.72250557e-01
-1.22549093e+00 6.27782702e-01 1.96775585e-01 -1.36219606e-01
-5.21797776e-01 3.89301598e-01 -1.04681393e-02 -7.97981441e-01
-2.25090295e-01 -6.80683374e-01 5.97905517e-01 -8.50777268e-01
4.54964548e-01 8.53385925e-02 4.27990317e-01 -2.11924031e-01
-5.01408935e-01 4.50353444e-01 -2.21993569e-02 2.10346401e-01
1.20235312e+00 8.06037903e-01 -6.52224541e-01 3.14840913e-01
1.19481015e+00 5.56649923e-01 -9.58102643e-01 -1.68680567e-02
3.85442883e-01 -6.96044147e-01 -2.43212104e-01 -1.01677895e+00
-8.03818703e-01 4.57889557e-01 4.49331664e-02 2.23623276e-01
1.10723639e+00 1.95697829e-01 5.04606009e-01 3.78438324e-01
7.49791980e-01 -5.66121340e-01 2.54535735e-01 3.83094996e-01
7.11615801e-01 -1.00109112e+00 -2.55379498e-01 -4.46407229e-01
-1.28056049e-01 1.13889146e+00 1.23509181e+00 -2.37748742e-01
1.32304668e-01 5.57139575e-01 -4.29100960e-01 -2.30877966e-01
-8.40171814e-01 3.20352554e-01 2.58899748e-01 5.39856553e-01
4.14600551e-01 -5.73507696e-03 -1.24081522e-01 8.29244971e-01
-1.94058433e-01 -3.71525101e-02 6.56282365e-01 1.31672955e+00
-5.86148083e-01 -1.22253323e+00 -6.75100088e-01 7.31809855e-01
-6.62839711e-02 -8.91959667e-02 -6.72454119e-01 6.51537776e-01
1.45642594e-01 9.08027828e-01 -3.05106342e-01 -7.32976198e-01
3.47230405e-01 9.98320207e-02 4.95625854e-01 -1.17982209e+00
-7.98937500e-01 -2.19476745e-01 8.33475813e-02 -8.12416077e-01
2.81732023e-01 -6.65269136e-01 -1.31452978e+00 1.09969601e-01
-5.76080978e-01 2.63577580e-01 2.62655407e-01 8.75469148e-01
9.08704162e-01 5.61796367e-01 6.15643620e-01 -1.92894891e-01
-7.03558981e-01 -6.20307386e-01 4.22466313e-03 3.70516926e-01
5.10050118e-01 -5.95091403e-01 -4.27579910e-01 2.41033465e-01] | [7.590681552886963, 7.712461948394775] |
d63c59b2-ce11-4d39-9cd9-a747f345dff0 | fpcc-net-fast-point-cloud-clustering-for | 2012.14618 | null | https://arxiv.org/abs/2012.14618v5 | https://arxiv.org/pdf/2012.14618v5.pdf | FPCC: Fast Point Cloud Clustering based Instance Segmentation for Industrial Bin-picking | Instance segmentation is an important pre-processing task in numerous real-world applications, such as robotics, autonomous vehicles, and human-computer interaction. Compared with the rapid development of deep learning for two-dimensional (2D) image tasks, deep learning-based instance segmentation of 3D point cloud still has a lot of room for development. In particular, distinguishing a large number of occluded objects of the same class is a highly challenging problem, which is seen in a robotic bin-picking. In a usual bin-picking scene, many identical objects are stacked together and the model of the objects is known. Thus, the semantic information can be ignored; instead, the focus in the bin-picking is put on the segmentation of instances. Based on this task requirement, we propose a Fast Point Cloud Clustering (FPCC) for instance segmentation of bin-picking scene. FPCC includes a network named FPCC-Net and a fast clustering algorithm. FPCC-net has two subnets, one for inferring the geometric centers for clustering and the other for describing features of each point. FPCC-Net extracts features of each point and infers geometric center points of each instance simultaneously. After that, the proposed clustering algorithm clusters the remaining points to the closest geometric center in feature embedding space. Experiments show that FPCC also surpasses the existing works in bin-picking scenes and is more computationally efficient. Our code and data are available at https://github.com/xyjbaal/FPCC. | ['Kazuhiro Kosuge', 'Fangzhou Lin', 'Diyi Liu', 'Shogo Arai', 'Yajun Xu'] | 2020-12-29 | null | null | null | null | ['3d-instance-segmentation-1'] | ['computer-vision'] | [-2.17398852e-01 -1.43781304e-01 -2.96946727e-02 -4.08331394e-01
-4.07986313e-01 -4.34149861e-01 4.37948406e-01 3.23509395e-01
-3.06310356e-01 4.81531136e-02 -4.13112760e-01 -1.91735089e-01
-2.90362328e-01 -9.06557679e-01 -7.84727871e-01 -8.67833734e-01
-7.41330385e-02 9.88699555e-01 4.88542974e-01 5.00938632e-02
3.28849226e-01 8.57506990e-01 -1.50952387e+00 -8.50507021e-02
6.97639227e-01 1.22752976e+00 6.34298027e-01 2.19047174e-01
-5.74051261e-01 -8.87130946e-02 -3.99598151e-01 -1.21641964e-01
4.79697049e-01 5.83797060e-02 -5.69232166e-01 4.05957550e-01
1.47659078e-01 -2.29642794e-01 -1.86606497e-01 1.18039083e+00
1.31109610e-01 2.57204860e-01 6.23785436e-01 -1.75949669e+00
-2.56541044e-01 3.95964026e-01 -8.45452726e-01 -1.54504329e-01
-1.78106889e-01 1.51752606e-02 7.55167961e-01 -8.52422535e-01
2.45024338e-01 1.33914661e+00 3.86295110e-01 3.61436188e-01
-7.95678854e-01 -8.31469774e-01 2.12495729e-01 3.92980903e-01
-1.42444777e+00 5.62844276e-02 1.01025605e+00 -5.74366450e-01
3.37190419e-01 2.16764398e-02 8.42691004e-01 4.08483356e-01
-1.21616997e-01 1.00192356e+00 4.72898751e-01 3.76864523e-02
4.12633181e-01 2.75672786e-02 3.98347139e-01 4.88094568e-01
3.98684591e-01 -2.65395284e-01 8.12129751e-02 8.91861841e-02
8.97105157e-01 6.89289570e-01 -1.62146650e-02 -7.83388495e-01
-1.25239289e+00 7.78605223e-01 8.80568206e-01 1.63593426e-01
-3.16575319e-01 2.05046311e-01 2.59102404e-01 -2.37945110e-01
1.95083722e-01 1.18154630e-01 -3.74770820e-01 1.53096721e-01
-8.45220804e-01 3.64623070e-01 6.54615104e-01 1.35219705e+00
1.17492473e+00 -4.73915011e-01 1.59988329e-01 6.67595804e-01
3.53056490e-01 2.21630514e-01 2.56829798e-01 -7.89266288e-01
4.20682520e-01 1.05655563e+00 1.24591395e-01 -1.38036370e+00
-4.40430731e-01 -2.92327970e-01 -9.08029735e-01 1.96918249e-01
2.13453069e-01 9.16580558e-02 -1.01575053e+00 1.15753043e+00
7.91137516e-01 3.28952849e-01 -2.51283735e-01 1.16793406e+00
7.77916372e-01 7.24290907e-01 -1.77894935e-01 1.77020460e-01
1.33144999e+00 -1.03722692e+00 -3.54347557e-01 -2.78797507e-01
2.61405170e-01 -5.82684159e-01 6.77253544e-01 2.65405208e-01
-6.40448689e-01 -6.05590045e-01 -9.41531599e-01 -2.14131013e-01
-5.49765229e-01 2.13399142e-01 7.88840830e-01 1.94432020e-01
-6.96044385e-01 5.36930621e-01 -1.03911483e+00 -3.43133330e-01
8.16677690e-01 5.53144097e-01 -3.56982648e-01 -3.92163217e-01
-5.51543534e-01 3.75329435e-01 6.88801050e-01 4.15422827e-01
-6.58412993e-01 -3.63707632e-01 -9.44283843e-01 6.92723393e-02
6.07129455e-01 -4.17732269e-01 9.39287126e-01 -5.71864843e-01
-1.12074745e+00 6.05948746e-01 -2.31068164e-01 -1.33101717e-01
5.03077924e-01 -2.77213275e-01 -2.69905925e-02 2.29176530e-03
3.35054517e-01 7.85222232e-01 9.03139412e-01 -1.56801915e+00
-9.28368092e-01 -6.92877173e-01 -2.50410885e-02 2.00431257e-01
-1.32896379e-02 -3.52692902e-01 -9.83580768e-01 -7.62740709e-03
7.66118526e-01 -9.16925907e-01 -3.92846972e-01 9.27898735e-02
-7.75601923e-01 -5.66809714e-01 1.24619734e+00 -2.50661552e-01
6.53042197e-01 -2.39232373e+00 8.03407580e-02 2.35036820e-01
4.53993350e-01 5.11932671e-02 4.79149818e-02 3.06938410e-01
-6.24348260e-02 8.85627046e-02 -4.38200504e-01 -4.90956843e-01
2.17056051e-01 1.93676412e-01 -9.18098316e-02 6.60509229e-01
3.40492338e-01 6.48142576e-01 -9.62852061e-01 -6.31841481e-01
6.67860091e-01 2.80444354e-01 -4.89254385e-01 1.51861042e-01
-3.02999526e-01 3.36434543e-01 -6.18096948e-01 6.22440219e-01
1.20250893e+00 -1.34525746e-01 -2.76658237e-01 -2.57927507e-01
-1.15384430e-01 -2.06986845e-01 -1.40673578e+00 1.62192094e+00
-2.35187523e-02 3.47620279e-01 6.83067366e-02 -1.24263072e+00
9.93936062e-01 -1.27297297e-01 7.83545792e-01 -1.52200803e-01
2.51570374e-01 2.14035898e-01 -5.00513092e-02 -4.81078118e-01
4.43198323e-01 1.16986200e-01 -8.31339806e-02 5.72732687e-02
-7.38326833e-02 -4.33523864e-01 1.86806321e-01 7.26468861e-02
7.51091123e-01 -3.77347991e-02 -3.38158980e-02 -1.05308332e-01
3.70837867e-01 3.76392186e-01 7.37444162e-01 3.10260266e-01
-2.38804817e-01 5.95145464e-01 2.59036481e-01 -3.80139112e-01
-9.03495073e-01 -9.51964617e-01 -2.38742262e-01 5.34661770e-01
9.09016550e-01 -1.59619659e-01 -7.36707330e-01 -5.53904533e-01
3.42738211e-01 4.80941415e-01 -3.73829782e-01 -1.76244490e-02
-3.73451620e-01 -4.48207200e-01 -8.53452384e-02 3.93540919e-01
5.30665994e-01 -9.91872370e-01 -5.56212366e-01 3.25639248e-02
4.40823846e-03 -1.09290683e+00 -2.74274558e-01 2.09187195e-01
-8.59719157e-01 -1.35983419e+00 -3.65748316e-01 -9.37712729e-01
8.25098634e-01 8.01243901e-01 7.43628204e-01 2.20491096e-01
-3.94412935e-01 -3.03939916e-02 -3.89279485e-01 -6.78310156e-01
2.69238770e-01 1.50586188e-01 3.57605293e-02 5.22549413e-02
8.56243491e-01 -4.03441668e-01 -7.83603013e-01 4.62679386e-01
-9.20324445e-01 1.08918577e-01 6.91463709e-01 5.22976279e-01
8.95522118e-01 5.24783373e-01 1.74783483e-01 -6.25736654e-01
1.08435839e-01 -6.62210941e-01 -8.05179417e-01 -2.65172392e-01
3.29238400e-02 -3.08116436e-01 5.56099772e-01 -2.62548268e-01
-3.46477360e-01 3.91251445e-01 -4.82619293e-02 -8.64911318e-01
-5.96579909e-01 3.08907330e-01 -5.46866357e-01 2.01280445e-01
7.47562721e-02 5.91075644e-02 -4.92836945e-02 -5.90868652e-01
2.71113902e-01 7.24099100e-01 4.00300264e-01 -3.45055699e-01
1.03421354e+00 6.29815996e-01 3.63162085e-02 -8.86695862e-01
-5.81829846e-01 -1.00082242e+00 -9.81483698e-01 -1.45661682e-01
9.71661091e-01 -7.26707280e-01 -8.63256693e-01 3.84896696e-01
-1.19824719e+00 -1.83061376e-01 -2.05507413e-01 4.32697207e-01
-4.89763826e-01 2.52259135e-01 -1.90770686e-01 -6.83736920e-01
1.25966677e-02 -1.41815960e+00 1.44617617e+00 4.51092243e-01
3.21712494e-01 -6.13853216e-01 -1.20526239e-01 4.22344506e-01
-1.75250873e-01 3.37891459e-01 9.64642465e-01 -7.73077369e-01
-9.07127261e-01 -4.30306643e-01 -4.59931105e-01 1.09336033e-01
2.32557148e-01 1.21227078e-01 -7.08075881e-01 -2.07359985e-01
2.17797793e-02 1.11778185e-01 6.93705499e-01 5.31351268e-01
1.52155828e+00 5.33344634e-02 -7.61229277e-01 6.46938741e-01
1.55048919e+00 3.27773631e-01 3.88967454e-01 2.46948466e-01
1.01925409e+00 6.30080104e-01 8.00599515e-01 2.46592224e-01
3.78586501e-01 5.19914091e-01 9.76262510e-01 -2.30377674e-01
2.99020022e-01 -1.27369359e-01 -1.66047215e-01 5.65481186e-01
1.26185939e-01 -1.05464019e-01 -9.98499870e-01 7.25183129e-01
-2.05348587e+00 -8.07198167e-01 -2.14797527e-01 2.09072495e+00
2.56954521e-01 2.58840006e-02 1.18896283e-01 2.77322143e-01
1.12203038e+00 -7.95943514e-02 -8.77782762e-01 1.26306862e-02
4.03071672e-01 -1.24875776e-01 4.79282647e-01 2.65890568e-01
-1.32345521e+00 1.03279424e+00 4.04831648e+00 8.68227065e-01
-1.04450655e+00 -1.16909795e-01 5.08358657e-01 1.81755483e-01
1.24797054e-01 7.90398754e-03 -8.86472642e-01 6.29562199e-01
1.50125414e-01 2.90267952e-02 2.93486416e-01 1.23680043e+00
1.63665354e-01 -1.67407826e-01 -1.25125575e+00 1.12677681e+00
-1.09484717e-01 -1.06816947e+00 7.14326650e-02 2.48153284e-01
3.74443173e-01 1.31820053e-01 -1.46617129e-01 2.23820090e-01
1.85042456e-01 -9.13042188e-01 6.48009360e-01 2.62366116e-01
4.14801270e-01 -9.23502088e-01 8.96312714e-01 6.92650557e-01
-1.33996630e+00 -2.20938891e-01 -8.92283976e-01 9.37724411e-02
7.65007408e-03 9.54890251e-01 -9.09781516e-01 6.93473637e-01
8.86507630e-01 7.24871874e-01 -3.62783611e-01 1.59950376e+00
-2.22640373e-02 1.71114400e-01 -5.22357881e-01 5.39728589e-02
4.35492158e-01 -6.51185870e-01 4.88526285e-01 9.66979980e-01
3.38215023e-01 1.55456826e-01 6.39749229e-01 9.27459002e-01
-9.11767855e-02 -5.82774654e-02 -5.56623578e-01 6.06775954e-02
6.98715687e-01 1.58925986e+00 -1.17881966e+00 -2.82187641e-01
-1.26781851e-01 9.36386585e-01 2.39710808e-01 1.45877019e-01
-7.53912807e-01 -7.04004943e-01 8.42295706e-01 7.01588467e-02
5.17545819e-01 -5.54860771e-01 -2.86403090e-01 -7.99434125e-01
2.34303195e-02 -3.17218512e-01 6.63239807e-02 -5.96149385e-01
-1.18746197e+00 3.75694156e-01 3.24829221e-02 -1.40368390e+00
2.44298533e-01 -7.40201473e-01 -7.04329491e-01 6.98187530e-01
-1.47954726e+00 -9.80782449e-01 -8.20535839e-01 6.54574156e-01
7.31870770e-01 2.36311406e-01 3.43303114e-01 2.05020443e-01
-7.61026740e-01 7.72894174e-02 2.11794719e-01 4.36892837e-01
2.73783535e-01 -1.26133192e+00 2.75786102e-01 6.39755189e-01
7.36074820e-02 5.49244165e-01 3.84451389e-01 -6.37751758e-01
-1.45876706e+00 -1.46217346e+00 4.43825394e-01 -2.74376392e-01
3.93345445e-01 -6.85979962e-01 -8.10849845e-01 5.09743214e-01
-2.05700547e-01 1.55483633e-01 3.49049270e-01 -2.03793660e-01
2.92003900e-02 -2.33217910e-01 -1.19198573e+00 4.83002603e-01
9.48923290e-01 -6.80914372e-02 -4.00690734e-01 6.53962076e-01
9.23746407e-01 -5.55085957e-01 -6.40392482e-01 2.77717233e-01
1.77051634e-01 -8.89162242e-01 1.02190578e+00 -2.94976711e-01
4.00464386e-01 -6.80237293e-01 -8.73963535e-02 -1.15268850e+00
-4.38666821e-01 -1.62630081e-01 1.81177810e-01 1.03818583e+00
-2.87959687e-02 -4.30995494e-01 9.42329645e-01 4.89840478e-01
-4.95458484e-01 -8.28247547e-01 -8.42425704e-01 -6.82989061e-01
-5.19174859e-02 -4.22262967e-01 9.49185252e-01 8.46030772e-01
-3.02958399e-01 1.50415868e-01 3.61890882e-01 6.82054996e-01
8.30639780e-01 5.20801485e-01 1.10185122e+00 -1.62833512e+00
1.32043570e-01 -5.07124245e-01 -8.31563771e-01 -1.15475655e+00
5.47791757e-02 -7.94290543e-01 3.26791883e-01 -1.87724841e+00
8.64844918e-02 -9.31101859e-01 -8.33997950e-02 4.54178900e-01
-9.33123007e-02 6.38326406e-02 3.09325486e-01 4.04370695e-01
-5.68782449e-01 5.68682253e-01 1.15727150e+00 -3.23833793e-01
-3.22849214e-01 2.82494426e-01 -4.13723439e-01 7.67864645e-01
7.87695169e-01 -4.01555002e-01 -2.29852527e-01 -5.59367537e-01
-3.16809028e-01 -1.59585685e-01 4.75614160e-01 -1.14847791e+00
5.17207861e-01 -2.64976859e-01 4.64369476e-01 -1.17862368e+00
5.15561461e-01 -1.28306913e+00 2.63869390e-02 3.19442511e-01
3.03166628e-01 -1.62417889e-01 1.95497960e-01 6.60230517e-01
-1.80092409e-01 -3.85943025e-01 6.25242472e-01 -2.89367318e-01
-8.37811887e-01 8.03162932e-01 9.97646600e-02 -4.24033552e-01
1.38702559e+00 -5.68278730e-01 -1.92286208e-01 8.51892978e-02
-4.79441047e-01 6.58356190e-01 5.69000065e-01 4.08923656e-01
7.87119806e-01 -1.14509523e+00 -4.14570987e-01 2.42447153e-01
1.10716157e-01 1.14504337e+00 2.28701070e-01 6.97179914e-01
-6.12796903e-01 3.29356343e-01 -1.63549855e-02 -1.07864606e+00
-9.52423513e-01 7.37131238e-01 1.14046752e-01 3.17885250e-01
-7.21505404e-01 8.53884041e-01 5.08955359e-01 -5.78161180e-01
3.31312865e-01 -4.80712891e-01 -2.21228585e-01 -9.06130951e-03
2.26097375e-01 2.49140143e-01 8.69854018e-02 -6.76199853e-01
-3.76132429e-01 7.30961084e-01 -1.39288798e-01 3.36700112e-01
1.49393761e+00 7.02453628e-02 -4.31574732e-01 5.39254844e-01
1.17432630e+00 -3.42981607e-01 -1.24803269e+00 -1.19815715e-01
6.97797984e-02 -5.72234333e-01 -1.84816178e-02 -1.80710241e-01
-1.31582367e+00 1.06625438e+00 4.84670430e-01 2.40041435e-01
9.06022370e-01 2.05719948e-01 8.24520528e-01 4.48415339e-01
6.04839683e-01 -9.39721048e-01 -4.97237705e-02 4.95707750e-01
6.58550799e-01 -1.20542073e+00 5.06983977e-03 -6.79188192e-01
-3.48224849e-01 1.06746018e+00 6.70686424e-01 -4.33084995e-01
8.50592494e-01 -1.24487787e-01 -7.25980029e-02 -4.59704965e-01
-2.73794588e-02 -2.33441725e-01 6.72866404e-02 6.05611086e-01
-1.66488349e-01 1.76725388e-01 1.22781985e-01 6.16668582e-01
-4.13501590e-01 -2.53287882e-01 1.60443529e-01 7.74886310e-01
-6.20087504e-01 -7.38584936e-01 -6.07240438e-01 4.99604255e-01
1.31479740e-01 2.97769725e-01 -2.69714564e-01 8.33097994e-01
5.73512197e-01 8.61205459e-01 5.26281178e-01 -4.87510294e-01
2.58460909e-01 -8.73877704e-02 7.20678344e-02 -8.07139814e-01
-2.33125344e-01 9.43486243e-02 -5.99072337e-01 -6.11094594e-01
-2.74989784e-01 -6.25477612e-01 -1.60955489e+00 -1.61733001e-01
-5.18929899e-01 2.61723161e-01 1.14431000e+00 9.27078485e-01
3.69895488e-01 3.98209453e-01 7.84696698e-01 -1.42748034e+00
-9.06280801e-02 -7.24368930e-01 -6.72908843e-01 4.18585598e-01
2.65158653e-01 -9.72468495e-01 -3.42822134e-01 -1.40213162e-01] | [7.983469009399414, -3.1718127727508545] |
c9ec1740-b1de-414b-9560-3b12847e9844 | codekgc-code-language-model-for-generative | 2304.09048 | null | https://arxiv.org/abs/2304.09048v1 | https://arxiv.org/pdf/2304.09048v1.pdf | CodeKGC: Code Language Model for Generative Knowledge Graph Construction | Current generative knowledge graph construction approaches usually fail to capture structural knowledge by simply flattening natural language into serialized texts or a specification language. However, large generative language model trained on structured data such as code has demonstrated impressive capability in understanding natural language for structural prediction and reasoning tasks. Intuitively, we address the task of generative knowledge graph construction with code language model: given a code-format natural language input, the target is to generate triples which can be represented as code completion tasks. Specifically, we develop schema-aware prompts that effectively utilize the semantic structure within the knowledge graph. As code inherently possesses structure, such as class and function definitions, it serves as a useful model for prior semantic structural knowledge. Furthermore, we employ a rationale-enhanced generation method to boost the performance. Rationales provide intermediate steps, thereby improving knowledge extraction abilities. Experimental results indicate that the proposed approach can obtain better performance on benchmark datasets compared with baselines. Code and datasets are available in https://github.com/zjunlp/DeepKE/tree/main/example/llm. | ['Ningyu Zhang', 'Huajun Chen', 'Wei Guo', 'Feiyu Xiong', 'Yinuo Jiang', 'Jing Chen', 'Zhen Bi'] | 2023-04-18 | null | null | null | null | ['graph-construction'] | ['graphs'] | [ 9.64119807e-02 5.80611587e-01 -4.69838083e-01 -4.42798346e-01
-7.06175804e-01 -7.48394191e-01 5.65307915e-01 8.73880833e-03
4.15869832e-01 5.99384725e-01 4.76558417e-01 -6.66935861e-01
1.21188767e-01 -1.08178186e+00 -9.99815643e-01 -1.05908372e-01
1.86771989e-01 2.79118389e-01 8.05815607e-02 -1.54883996e-01
2.29687661e-01 -3.74923646e-01 -1.28594112e+00 6.38168633e-01
1.22393775e+00 5.47170162e-01 4.12876427e-01 2.70031273e-01
-6.45063221e-01 1.43138146e+00 -1.79631650e-01 -8.39856803e-01
-3.21528725e-02 -5.79145014e-01 -1.27094746e+00 1.68034213e-03
-4.36981320e-02 -2.09275395e-01 -2.28597075e-01 1.16989863e+00
-9.45288464e-02 -1.57776430e-01 4.34107929e-01 -1.37205982e+00
-1.10833716e+00 1.51267862e+00 -4.34556782e-01 -1.17393218e-01
6.13243818e-01 1.26402348e-01 1.29007900e+00 -1.04041946e+00
6.65249646e-01 1.14892673e+00 3.62525791e-01 6.71167672e-01
-1.06346142e+00 -5.63641965e-01 2.08598912e-01 1.90246686e-01
-1.41196954e+00 -3.62217814e-01 9.06721056e-01 -6.46493673e-01
1.19486117e+00 1.11894205e-01 5.97783327e-01 1.03377497e+00
-2.37198807e-02 1.09269345e+00 5.61085522e-01 -3.70564550e-01
-3.89573351e-02 2.26159766e-01 4.02736813e-01 1.19731641e+00
4.64726388e-01 -1.62722707e-01 -3.81534606e-01 -3.30471605e-01
6.62997782e-01 1.26360878e-01 -4.31085110e-01 -5.81835806e-01
-1.00563931e+00 9.48659539e-01 5.39763153e-01 8.14887602e-03
-1.39069572e-01 3.23914886e-01 3.84422749e-01 1.81686267e-01
3.13596390e-02 4.56576586e-01 -4.56337154e-01 -1.72677130e-01
-6.44074082e-01 1.45495743e-01 8.96602273e-01 1.75783682e+00
1.07662547e+00 1.87234148e-01 -1.16477810e-01 7.05570817e-01
5.33370614e-01 4.35726523e-01 4.75722283e-01 -5.77600837e-01
8.35565865e-01 1.23935521e+00 -3.18010956e-01 -7.88863599e-01
5.80485947e-02 -5.00133991e-01 -5.39438188e-01 -4.31253463e-01
-4.69579250e-02 -1.66760534e-01 -7.51496255e-01 1.69980884e+00
1.74166068e-01 1.36477590e-01 4.02991682e-01 6.32595956e-01
1.02197075e+00 6.20843589e-01 7.00308383e-02 4.35841829e-02
1.47754979e+00 -1.08822215e+00 -3.46659750e-01 -5.05613983e-01
8.15379322e-01 -4.78721619e-01 1.28316724e+00 1.02655128e-01
-8.69228482e-01 -3.45256656e-01 -8.65051985e-01 -1.47142872e-01
-1.45002261e-01 2.51848370e-01 1.08646619e+00 5.32061398e-01
-8.44227731e-01 -5.06097414e-02 -8.53074253e-01 -1.47487357e-01
5.83250165e-01 -2.44958885e-02 -4.46297899e-02 -4.23999190e-01
-1.02989924e+00 3.02132070e-01 9.33522880e-01 -1.70777455e-01
-1.20244527e+00 -8.41608346e-01 -1.26263058e+00 3.70239109e-01
8.09854388e-01 -1.09160769e+00 1.37121224e+00 -8.03310871e-01
-1.19346011e+00 5.81432343e-01 -1.30018100e-01 -4.26438898e-01
1.72374360e-02 -2.64383078e-01 -1.92827195e-01 -2.46569946e-01
1.54315397e-01 4.16162401e-01 5.46546578e-01 -1.50866175e+00
-4.54748124e-01 -5.22478111e-02 6.37648642e-01 5.77021502e-02
-2.39238665e-01 -8.31708312e-02 -6.41362667e-01 -4.77018863e-01
-8.68498832e-02 -8.61661971e-01 -1.38898626e-01 -6.67766094e-01
-8.72736633e-01 -3.44981104e-01 4.61528927e-01 -6.53803051e-01
1.61864436e+00 -2.01917744e+00 1.11109041e-01 2.47787505e-01
3.93395662e-01 -6.77347630e-02 4.42521870e-02 7.34879255e-01
-1.51736170e-01 3.83411020e-01 -3.81891310e-01 4.42200564e-02
3.29641581e-01 2.30285764e-01 -7.02532828e-01 -2.08054498e-01
3.82566869e-01 1.53273702e+00 -1.04208267e+00 -5.34413636e-01
-1.62507460e-01 1.61199525e-01 -9.08285916e-01 3.26351523e-01
-8.14214826e-01 -1.68405194e-02 -8.29306126e-01 6.15822375e-01
3.10934097e-01 -7.52208650e-01 5.16244233e-01 5.06472960e-02
3.01160425e-01 6.55683041e-01 -8.60988736e-01 2.10645366e+00
-7.27738261e-01 2.04202861e-01 -5.00441611e-01 -7.92620838e-01
9.53689277e-01 2.55526930e-01 -1.26644939e-01 -3.21321547e-01
-1.85614169e-01 -1.99515279e-02 -7.58539960e-02 -6.49250925e-01
3.26861769e-01 1.35531202e-01 -3.81398141e-01 5.41512072e-01
1.22967627e-04 -1.62280172e-01 3.07612956e-01 8.60044003e-01
1.30150664e+00 4.52983290e-01 5.55210233e-01 -1.58774674e-01
5.48809171e-01 3.17340642e-01 6.34401083e-01 4.11896795e-01
5.91748059e-01 9.30974260e-02 8.81255984e-01 -1.93552926e-01
-8.13581705e-01 -1.13972020e+00 4.18518305e-01 9.91578519e-01
6.54766560e-02 -1.13859355e+00 -7.06282735e-01 -9.39540446e-01
-2.33576283e-01 1.08597255e+00 -4.89145309e-01 -3.97786707e-01
-5.36381006e-01 -5.67864776e-01 5.25727689e-01 8.20748210e-01
4.74213839e-01 -1.04689527e+00 -3.20327967e-01 1.13037169e-01
-3.97323340e-01 -9.53327596e-01 -4.27694082e-01 -7.75959641e-02
-7.41000414e-01 -1.39179993e+00 8.20521414e-02 -8.53778303e-01
1.08851945e+00 1.22168370e-01 1.58464015e+00 5.05733013e-01
-1.25093564e-01 3.60086828e-01 -5.38423836e-01 -1.51196882e-01
-6.50601268e-01 1.64190635e-01 -6.31458879e-01 -3.74423712e-01
4.32927102e-01 -7.29122937e-01 -4.86907750e-01 -3.01500726e-02
-1.02707958e+00 7.12998033e-01 6.92706585e-01 8.11990023e-01
4.11048830e-01 1.39129786e-02 5.41828692e-01 -1.46709895e+00
7.26161778e-01 -7.63643980e-01 -6.28402352e-01 5.36850214e-01
-7.11930990e-01 5.85101604e-01 9.16801870e-01 2.30032415e-03
-1.41720605e+00 2.06264526e-01 5.70328794e-02 -1.14814170e-01
-5.57504483e-02 1.05712986e+00 -3.16004783e-01 4.58338290e-01
7.06171155e-01 7.55931854e-01 -4.49567825e-01 -2.69730628e-01
6.06833041e-01 4.06163663e-01 4.74746972e-01 -1.25485563e+00
1.11120856e+00 7.73876458e-02 -4.34168100e-01 -2.26112053e-01
-7.98066199e-01 -2.79553682e-01 -3.39593649e-01 2.90978044e-01
5.96861005e-01 -1.09397638e+00 -3.96656394e-01 -1.11318603e-01
-1.16596019e+00 -3.23473245e-01 -1.19363636e-01 9.84964743e-02
-5.81443727e-01 3.17680359e-01 -4.90616024e-01 -5.68235338e-01
-5.77210844e-01 -1.13377666e+00 1.08151925e+00 1.77335888e-01
-2.03788221e-01 -1.18591368e+00 1.01523377e-01 4.59031850e-01
2.77232409e-01 1.68193206e-01 1.54945099e+00 -6.88022196e-01
-1.17453945e+00 1.64788648e-01 -1.79808840e-01 1.07046701e-01
2.75808901e-01 -2.13187132e-02 -5.79315722e-01 -8.59883055e-02
-3.40023637e-01 -4.57490116e-01 5.97630501e-01 -2.64333457e-01
1.27585292e+00 -6.84568524e-01 -4.04469103e-01 6.03441417e-01
1.69105256e+00 1.84097275e-01 6.38432324e-01 -1.33322570e-02
9.66151237e-01 2.70343035e-01 2.21495479e-01 4.84604299e-01
9.35111165e-01 4.38340485e-01 3.25545162e-01 2.17744812e-01
-3.31050277e-01 -9.63711560e-01 4.65898484e-01 1.07977486e+00
3.03374380e-02 -6.15243837e-02 -1.35228384e+00 7.20397651e-01
-2.04196310e+00 -9.64334548e-01 -2.08230659e-01 1.75806665e+00
1.23854375e+00 -4.29252796e-02 -1.34094641e-01 -3.47772419e-01
4.38134670e-01 -1.55731902e-01 -4.35125440e-01 7.70523623e-02
2.99365520e-01 7.30084702e-02 3.01589016e-02 5.39262295e-01
-5.84552169e-01 1.22018111e+00 4.87925625e+00 8.95789146e-01
-8.30096364e-01 1.12845838e-01 8.18013623e-02 3.02809894e-01
-1.00907147e+00 5.67658782e-01 -7.76167572e-01 4.50487912e-01
7.15266228e-01 -9.10809755e-01 5.88151991e-01 1.20264447e+00
-6.90684468e-02 1.88724175e-01 -1.34474385e+00 7.54927218e-01
-8.34489241e-02 -1.48339307e+00 5.15331805e-01 -2.11166263e-01
8.98593843e-01 -1.56383380e-01 -2.16142252e-01 7.49660790e-01
9.35157776e-01 -9.93467152e-01 6.00368083e-01 4.28899527e-01
6.90514028e-01 -6.18532240e-01 5.05850196e-01 4.55044180e-01
-1.54059565e+00 -5.71483783e-02 -3.25664788e-01 -7.68281566e-03
-1.31173134e-01 4.39411312e-01 -1.33200192e+00 9.77075040e-01
1.98880970e-01 8.20278168e-01 -8.27076912e-01 6.99524462e-01
-9.18020904e-01 8.12234938e-01 1.14975967e-01 -1.05409406e-01
1.26343757e-01 5.85668832e-02 1.40332490e-01 1.30978286e+00
2.60481894e-01 1.28438905e-01 3.35901082e-01 1.41619182e+00
-4.20031190e-01 1.51923180e-01 -7.85003662e-01 -3.91706645e-01
5.48153937e-01 1.25036454e+00 -5.79768598e-01 -4.77665424e-01
-7.82337487e-01 6.22254968e-01 5.95063090e-01 6.01594031e-01
-1.12084985e+00 -2.99051493e-01 4.29222614e-01 2.77579203e-02
1.93964660e-01 -1.10566862e-01 -1.39289305e-01 -1.54769015e+00
1.86128020e-01 -8.80303264e-01 4.29292083e-01 -1.01765370e+00
-9.39452350e-01 5.52816033e-01 2.24556908e-01 -9.01023805e-01
-4.37730014e-01 -2.88016021e-01 -6.88537657e-01 6.41027987e-01
-1.27991223e+00 -1.36026335e+00 -4.33668584e-01 7.00958669e-01
8.18883419e-01 -2.10574061e-01 7.68762350e-01 1.10727139e-01
-4.16217387e-01 5.91076851e-01 -4.06404406e-01 4.81062293e-01
1.69495046e-01 -1.46061432e+00 6.27632260e-01 1.25239921e+00
4.55194294e-01 1.27808690e+00 3.67113322e-01 -9.16636884e-01
-1.78515112e+00 -1.43489385e+00 5.80672026e-01 -6.46235645e-01
8.46137881e-01 -5.02143979e-01 -1.00197351e+00 1.17307293e+00
3.24787557e-01 -2.80128837e-01 8.32030535e-01 1.67804420e-01
-7.68831134e-01 2.49029368e-01 -5.06737232e-01 4.99355525e-01
1.48265874e+00 -7.22685099e-01 -8.84114504e-01 4.05583769e-01
9.72132027e-01 -5.87575614e-01 -7.92560399e-01 2.07347691e-01
1.54892534e-01 -5.93571901e-01 6.89551055e-01 -8.66402268e-01
1.06820536e+00 -6.04589820e-01 -1.37579277e-01 -1.31937814e+00
-2.55593270e-01 -7.29719877e-01 -4.45513487e-01 1.32600760e+00
8.67093921e-01 -4.32498693e-01 6.84838176e-01 7.76904285e-01
-4.55281705e-01 -8.49550545e-01 -1.66431308e-01 -6.61389291e-01
-1.96574733e-01 -4.36642885e-01 8.48193765e-01 9.95605290e-01
4.79390591e-01 8.99603546e-01 -2.05838978e-01 2.35732824e-01
4.82829839e-01 7.76621580e-01 1.03875721e+00 -9.40095305e-01
-5.60927451e-01 -1.99668959e-01 -1.41019776e-01 -1.21022809e+00
4.61461842e-01 -1.52994227e+00 -1.73874617e-01 -1.88136446e+00
7.44748175e-01 -4.59431112e-01 7.86607936e-02 1.04826379e+00
-4.24908251e-01 -4.39874262e-01 -5.10384142e-02 8.79902691e-02
-7.96061873e-01 5.84579408e-01 1.19042516e+00 -2.73311973e-01
7.98693299e-02 -2.29067162e-01 -1.19020569e+00 5.26490629e-01
6.97594821e-01 -4.92703468e-01 -1.22445238e+00 -6.46432936e-01
7.28030860e-01 3.45303029e-01 3.71549994e-01 -6.43718839e-01
4.41091239e-01 -3.54025126e-01 -2.06080586e-01 -7.47129768e-02
-1.94153279e-01 -7.42624104e-01 5.96448421e-01 3.71228576e-01
-4.05676454e-01 -2.19872594e-03 1.49456054e-01 5.05821109e-01
-3.15575033e-01 -3.80637914e-01 2.71862537e-01 -3.21553379e-01
-1.01435816e+00 3.84365022e-01 8.97794813e-02 4.36848313e-01
9.31979001e-01 1.23439983e-01 -7.97384083e-01 -1.41336456e-01
-4.32291925e-01 4.01073605e-01 5.99080324e-01 7.22200274e-01
7.56593943e-01 -1.35879016e+00 -5.49515843e-01 1.73304662e-01
6.01772726e-01 2.44958773e-01 6.34900108e-02 5.17145216e-01
-4.52117652e-01 3.80169809e-01 2.64716055e-02 -2.86601454e-01
-1.04034412e+00 7.89998412e-01 1.45012528e-01 -3.38824660e-01
-5.20305216e-01 8.32316160e-01 6.80655599e-01 -4.03655201e-01
-1.49214327e-01 -6.42903924e-01 1.86796449e-02 -6.20855927e-01
3.85228872e-01 -2.54988432e-01 -9.66309831e-02 -1.16238214e-01
-3.62565547e-01 1.83224872e-01 -1.59585983e-01 4.03456807e-01
1.37902999e+00 7.70111196e-03 -2.89544851e-01 5.66724166e-02
8.70698154e-01 2.45928332e-01 -9.66480196e-01 -5.23095548e-01
2.61577845e-01 -3.24609667e-01 -3.11667413e-01 -9.02629793e-01
-9.93318021e-01 7.02501416e-01 -3.27549696e-01 1.60358205e-01
9.99830365e-01 3.27159494e-01 6.85153127e-01 6.38496399e-01
6.50402844e-01 -3.80936563e-01 1.54473171e-01 5.93574941e-01
8.49099219e-01 -1.07785237e+00 -2.53187031e-01 -8.55813026e-01
-6.91951573e-01 1.09235406e+00 9.24891472e-01 1.98751420e-01
2.76386797e-01 4.23124284e-01 -2.77773291e-01 -5.85218787e-01
-1.19427156e+00 -2.49615625e-01 2.62026519e-01 4.69693452e-01
6.88937426e-01 1.68061763e-01 -3.78020704e-02 1.01783228e+00
-5.22818983e-01 1.73793845e-02 5.47050714e-01 9.46067512e-01
-4.54960704e-01 -1.26173079e+00 1.20523281e-01 5.17434359e-01
-1.89307943e-01 -6.74748659e-01 -3.99474949e-01 6.84249341e-01
-9.44105461e-02 6.75232649e-01 -5.06867290e-01 -3.37961555e-01
2.74958372e-01 3.54706854e-01 4.95056361e-01 -1.27847886e+00
-3.65962476e-01 -3.63955379e-01 3.33719432e-01 -4.76081491e-01
-1.59211382e-01 -4.39614028e-01 -1.72465789e+00 -2.26985618e-01
-1.80046111e-01 5.87034047e-01 2.45386198e-01 8.37347150e-01
4.22183186e-01 7.09600627e-01 3.55987698e-01 9.84523594e-02
-3.21904957e-01 -6.96837246e-01 -1.86724767e-01 4.26539749e-01
-1.28832310e-01 -5.53287327e-01 2.76972763e-02 7.90830374e-01] | [7.886507987976074, 7.874252796173096] |
f76a75ab-defd-44d7-b261-6dd6ae8c64bf | region-adaptive-texture-enhancement-for | 2005.12486 | null | https://arxiv.org/abs/2005.12486v1 | https://arxiv.org/pdf/2005.12486v1.pdf | Region-adaptive Texture Enhancement for Detailed Person Image Synthesis | The ability to produce convincing textural details is essential for the fidelity of synthesized person images. However, existing methods typically follow a ``warping-based'' strategy that propagates appearance features through the same pathway used for pose transfer. However, most fine-grained features would be lost due to down-sampling, leading to over-smoothed clothes and missing details in the output images. In this paper we presents RATE-Net, a novel framework for synthesizing person images with sharp texture details. The proposed framework leverages an additional texture enhancing module to extract appearance information from the source image and estimate a fine-grained residual texture map, which helps to refine the coarse estimation from the pose transfer module. In addition, we design an effective alternate updating strategy to promote mutual guidance between two modules for better shape and appearance consistency. Experiments conducted on DeepFashion benchmark dataset have demonstrated the superiority of our framework compared with existing networks. | ['Zhanning Gao', 'Xuansong Xie', 'Xinfeng Zhang', 'Wen Gao', 'Shanshe Wang', 'Lingbo Yang', 'Siwei Ma', 'Peiran Ren', 'Pan Wang'] | 2020-05-26 | null | null | null | null | ['pose-transfer'] | ['computer-vision'] | [ 4.15982127e-01 1.62964761e-01 2.57263094e-01 -4.48176265e-01
-3.37814420e-01 -3.71872514e-01 6.05591834e-01 -4.52698886e-01
4.74397019e-02 7.42079556e-01 3.31357628e-01 4.09402072e-01
4.50727418e-02 -8.99242043e-01 -8.31756890e-01 -8.01193416e-01
4.17196721e-01 9.78169031e-03 1.64613664e-01 -3.30773979e-01
-4.82762083e-02 2.60010600e-01 -1.39549851e+00 3.49618524e-01
9.76554334e-01 9.69017506e-01 1.58638824e-02 3.25969547e-01
1.17288105e-01 6.20233834e-01 -4.33690190e-01 -7.76023388e-01
4.50642437e-01 -3.99039090e-01 -5.35330534e-01 4.65553939e-01
9.17688310e-01 -6.37526572e-01 -5.56747973e-01 1.08780062e+00
3.93474013e-01 2.17555821e-01 6.11561775e-01 -9.69864428e-01
-8.78037930e-01 2.53920555e-01 -8.99537325e-01 -3.50822002e-01
4.85907435e-01 4.01188433e-01 6.31632745e-01 -8.05286348e-01
7.82220542e-01 1.50946259e+00 7.31099367e-01 7.51355290e-01
-1.35705471e+00 -6.39112234e-01 4.02497530e-01 -5.94069734e-02
-1.15863562e+00 -4.95494246e-01 1.06849635e+00 -2.01013342e-01
2.51157075e-01 2.16335684e-01 8.62238109e-01 1.40880549e+00
3.06165248e-01 6.77757502e-01 1.44164622e+00 -9.38506275e-02
-1.21663854e-01 8.52761120e-02 -4.10940766e-01 1.08472300e+00
3.31201494e-01 5.02142966e-01 -7.85559416e-01 8.66909251e-02
1.35147548e+00 1.39169265e-02 -3.62599015e-01 -4.71822888e-01
-1.17066967e+00 3.81814420e-01 7.44603217e-01 -8.33161697e-02
-4.82078791e-01 2.59293139e-01 -8.19709711e-03 2.70393103e-01
5.62492311e-01 2.35909268e-01 -7.28908507e-03 1.60310879e-01
-8.97483110e-01 4.06335860e-01 3.76194537e-01 7.98139215e-01
7.05724299e-01 2.08220705e-01 -5.76869071e-01 8.78095865e-01
3.51841331e-01 5.41693330e-01 -5.94456941e-02 -1.03515434e+00
4.33485359e-01 5.65111041e-01 2.90902376e-01 -1.31470668e+00
-7.30731785e-02 -5.86724520e-01 -1.19476414e+00 4.91271883e-01
6.44246876e-01 -9.30857211e-02 -1.12006879e+00 1.78727579e+00
5.34534633e-01 8.05071220e-02 -3.52305204e-01 1.24090934e+00
7.86477268e-01 3.85273904e-01 -1.80730168e-02 3.69739562e-01
1.32399297e+00 -1.19877112e+00 -7.10799038e-01 -1.27230421e-01
-2.93913335e-01 -8.41329336e-01 1.04042506e+00 4.01761800e-01
-1.38716805e+00 -8.72924268e-01 -9.64543879e-01 -1.54099047e-01
9.48678795e-03 2.77409971e-01 5.88814735e-01 5.66498339e-01
-1.02930117e+00 8.45754027e-01 -6.49811745e-01 -1.82720065e-01
6.20766163e-01 1.93934768e-01 -4.95574474e-01 -1.26390412e-01
-9.53772783e-01 6.00877404e-01 -7.48426616e-02 5.08421957e-01
-8.36029708e-01 -7.51824796e-01 -8.71114612e-01 -1.83875576e-01
1.44009203e-01 -1.22383296e+00 7.38216996e-01 -1.22480488e+00
-1.81255221e+00 5.84416270e-01 -6.32449016e-02 5.42344563e-02
9.59835172e-01 -1.67226583e-01 -7.72758275e-02 2.48106882e-01
-8.45512748e-03 9.38539267e-01 1.43711627e+00 -1.53153098e+00
-4.68259543e-01 -2.67023951e-01 1.33726805e-01 4.24225539e-01
-1.64234057e-01 -4.20271069e-01 -6.17608666e-01 -1.25054145e+00
1.04803458e-01 -8.73141229e-01 -2.51105994e-01 6.46954358e-01
-5.25031269e-01 2.91407764e-01 6.40121996e-01 -1.12399471e+00
8.04039061e-01 -2.07387996e+00 5.34873784e-01 3.98822010e-01
4.82455641e-01 -2.21543089e-01 -2.71548361e-01 1.05922528e-01
1.39953047e-01 -3.59092802e-01 -2.05180138e-01 -5.95880866e-01
1.00942671e-01 -9.05692726e-02 -2.55066991e-01 4.62113470e-01
4.86308694e-01 9.74375010e-01 -8.27707767e-01 -3.28218192e-01
3.97417575e-01 9.32499588e-01 -6.12897575e-01 2.27335319e-01
-1.95671432e-02 9.40399706e-01 -4.44453180e-01 7.45764494e-01
8.35264564e-01 -2.21514255e-01 -3.80053138e-03 -6.57008111e-01
8.52478594e-02 -1.91318557e-01 -1.11732268e+00 1.91499090e+00
-4.75278586e-01 2.48746827e-01 3.11810941e-01 -5.08704424e-01
9.70598340e-01 1.35716423e-01 4.09265131e-01 -7.10875034e-01
2.55223483e-01 -6.69633746e-02 -2.59813935e-01 -1.51307121e-01
5.01877189e-01 -5.52585796e-02 2.31157560e-02 2.68276423e-01
-1.42702833e-01 -1.13642260e-01 -2.38144904e-01 2.13371348e-02
6.99614286e-01 6.98867261e-01 -2.39250630e-01 -2.05117837e-01
5.24967790e-01 -3.70266795e-01 4.24645841e-01 5.84994912e-01
-2.09056571e-01 9.55667973e-01 1.48497177e-02 -6.19890213e-01
-1.32886934e+00 -1.31226587e+00 7.02337474e-02 7.78664827e-01
4.94517714e-01 -1.89745128e-01 -1.02165973e+00 -6.17916048e-01
-5.54384850e-02 1.53512999e-01 -9.38098252e-01 -1.83493644e-01
-5.11931300e-01 -5.50822020e-01 4.44074124e-01 5.13999879e-01
9.64940012e-01 -9.49443877e-01 -2.86451429e-01 3.27403307e-01
-5.13587415e-01 -8.87130082e-01 -8.54386985e-01 -5.79816997e-01
-6.66495860e-01 -7.71325827e-01 -1.17924094e+00 -6.88840449e-01
1.18881595e+00 1.14679039e-01 8.43080521e-01 2.72744358e-01
-3.89263391e-01 7.64160976e-02 -8.86379778e-02 3.66093554e-02
-1.00165665e-01 -1.74803048e-01 -3.83464135e-02 4.64431286e-01
-3.23982388e-01 -5.86695313e-01 -1.05913913e+00 2.73521274e-01
-7.70483017e-01 5.37369847e-01 7.58921087e-01 1.02027154e+00
4.22898293e-01 1.88728333e-01 4.16858643e-01 -5.43397844e-01
5.11714756e-01 9.66904461e-02 -3.35468948e-01 2.88747847e-01
-3.04785430e-01 7.38369375e-02 6.37821496e-01 -5.00584245e-01
-1.66219962e+00 6.71112835e-02 -7.62518793e-02 -3.67818296e-01
8.97891074e-03 -2.39054367e-01 -1.97951302e-01 -3.57994884e-01
3.56672347e-01 4.14134204e-01 1.87205866e-01 -4.68035966e-01
4.83017623e-01 1.63272709e-01 6.76789284e-01 -9.49296653e-01
1.26102293e+00 8.55367899e-01 -8.78301710e-02 -5.35250902e-01
-8.20571840e-01 1.55928463e-01 -5.24576366e-01 -4.99315858e-01
7.43771851e-01 -1.00491047e+00 -8.91375482e-01 8.66004050e-01
-9.42259789e-01 -3.55447233e-01 -2.25626409e-01 8.53385925e-02
-4.42083895e-01 4.10531938e-01 -8.60573709e-01 -5.64045787e-01
-5.64566255e-01 -1.03363967e+00 1.28406835e+00 3.70459944e-01
-9.83372331e-02 -8.73431742e-01 -1.25651672e-01 6.20460272e-01
6.50290549e-01 3.93975496e-01 4.78270292e-01 5.02907515e-01
-7.15632558e-01 1.49704069e-01 -5.57120562e-01 2.28123516e-01
3.40133339e-01 -9.21888575e-02 -1.00299370e+00 -5.70253074e-01
-2.94449389e-01 -2.88370728e-01 8.60207081e-01 2.96127051e-01
1.18609405e+00 -3.92441452e-01 -1.90903068e-01 7.66750276e-01
1.15378630e+00 -3.87858659e-01 6.65563703e-01 1.65275976e-01
1.07548463e+00 8.61874282e-01 5.01346469e-01 3.38755101e-01
4.78792280e-01 7.75405884e-01 -8.90357569e-02 -6.34910285e-01
-6.78558350e-01 -5.07973909e-01 4.35142130e-01 5.78363776e-01
-5.18084884e-01 1.01304062e-01 -2.31442809e-01 2.37158999e-01
-1.75254869e+00 -9.14354086e-01 2.83539355e-01 2.07047272e+00
1.07417703e+00 4.56637070e-02 1.32444695e-01 -1.37434274e-01
6.54234469e-01 1.49769634e-01 -4.82917935e-01 1.87260676e-02
-1.86389819e-01 2.45177299e-01 2.70565510e-01 6.00675464e-01
-9.68374729e-01 9.52531517e-01 5.91360760e+00 8.39587390e-01
-9.91011441e-01 -1.63224325e-01 7.50418961e-01 -3.73787135e-02
-5.16486168e-01 -2.69367605e-01 -4.66695964e-01 5.16919672e-01
1.72324404e-01 1.90458134e-01 5.65298915e-01 4.63818550e-01
2.96351105e-01 1.00443810e-01 -8.23643029e-01 9.24969375e-01
5.95347248e-02 -1.24391353e+00 3.31372499e-01 -4.13966691e-03
1.02072966e+00 -6.66231632e-01 3.71122211e-01 -7.68793970e-02
3.76459390e-01 -1.01347184e+00 1.06571758e+00 8.50265324e-01
1.02714682e+00 -9.23941374e-01 3.90446275e-01 -2.14333907e-01
-1.47900164e+00 1.77683607e-01 -3.18066180e-01 1.10077016e-01
2.42422640e-01 5.92445016e-01 -3.13150376e-01 7.60487795e-01
8.37809980e-01 7.40535855e-01 -6.10612273e-01 7.41355956e-01
-3.04684162e-01 1.44881532e-01 -5.56479692e-02 4.69820112e-01
1.31265037e-02 -3.90741050e-01 5.04560351e-01 8.96432161e-01
2.17386052e-01 -1.42598078e-02 8.34931508e-02 1.18115270e+00
-2.90878303e-02 -9.34235826e-02 -4.32020038e-01 3.56711656e-01
3.24570686e-01 1.34529495e+00 -7.08454728e-01 -3.98968279e-01
-3.61329496e-01 1.51422668e+00 3.53942722e-01 5.20848989e-01
-9.74205554e-01 -2.64755160e-01 6.08045220e-01 3.15184385e-01
2.03086749e-01 -2.56441720e-02 -3.47629547e-01 -1.34229136e+00
9.36603546e-02 -1.07383120e+00 -4.67782505e-02 -6.93969548e-01
-1.56995976e+00 6.27800941e-01 -3.03998530e-01 -1.20759118e+00
1.32615909e-01 -3.24692667e-01 -5.02739847e-01 1.02797091e+00
-1.42791903e+00 -1.66483057e+00 -6.45314574e-01 7.03978539e-01
4.80500340e-01 9.85924378e-02 4.33641016e-01 3.65488082e-01
-6.01750016e-01 8.22994173e-01 -2.30539590e-01 1.14424668e-01
9.32590365e-01 -1.19074261e+00 5.80134451e-01 8.44978750e-01
-3.02325934e-01 7.98451126e-01 6.54445708e-01 -9.25132692e-01
-1.32859254e+00 -1.34085238e+00 3.23681355e-01 -4.33218777e-01
2.29074612e-01 -3.35978985e-01 -7.25739002e-01 4.71422583e-01
2.08972484e-01 -7.29599372e-02 2.08992928e-01 -1.82148620e-01
-5.02093911e-01 -2.58937985e-01 -1.20933974e+00 9.01173472e-01
1.17724872e+00 -3.75904888e-01 -3.35903138e-01 -1.56383440e-01
5.01718223e-01 -5.22024632e-01 -9.70182776e-01 4.13340390e-01
1.08823824e+00 -9.65877116e-01 1.24859858e+00 -1.41050071e-01
6.68860734e-01 -5.28584480e-01 1.80169359e-01 -1.30819106e+00
-6.61338627e-01 -7.36353815e-01 -9.40039158e-02 1.23283720e+00
5.27268499e-02 -4.97105241e-01 9.01727796e-01 5.97693622e-01
1.33412927e-01 -6.73859775e-01 -5.50428629e-01 -6.31208181e-01
-1.78238332e-01 9.12813321e-02 7.57702529e-01 7.78951466e-01
-4.68725741e-01 1.39544636e-01 -9.57958043e-01 3.89283970e-02
1.25501919e+00 1.06746972e-01 8.49316835e-01 -9.16076243e-01
-4.57854271e-01 -3.97689819e-01 -1.29617140e-01 -1.13747895e+00
-1.43047705e-01 -5.01956046e-01 1.35620102e-01 -1.32374299e+00
3.63148183e-01 -4.81692731e-01 -1.56563640e-01 3.77635419e-01
-3.86541516e-01 8.42560768e-01 2.04306155e-01 -1.90656690e-03
-3.46637577e-01 8.34196627e-01 2.01891851e+00 -1.03539146e-01
-1.43461404e-02 -1.73734769e-01 -7.87682056e-01 6.85498476e-01
5.41307092e-01 -2.47570146e-02 -5.39781749e-01 -5.25000513e-01
-4.46982682e-02 -1.23727307e-01 8.67603123e-01 -9.57452536e-01
-1.34665919e-02 -1.71400696e-01 1.06207418e+00 -4.03240889e-01
6.09289885e-01 -7.24365234e-01 4.84607726e-01 4.15711105e-01
-1.86459631e-01 -1.77789465e-01 1.82119757e-02 6.81191623e-01
8.96248315e-03 5.90547442e-01 9.65350449e-01 -2.47520544e-02
-3.82010043e-01 5.81017673e-01 1.05303243e-01 -3.46814364e-01
7.54494190e-01 -4.02334869e-01 -2.51182646e-01 -4.09899145e-01
-6.62460208e-01 -7.45358840e-02 9.61599708e-01 6.49538219e-01
8.61515164e-01 -1.63121617e+00 -8.17847133e-01 4.34706360e-01
-3.77732143e-02 -2.16575842e-02 6.41732216e-01 7.68961370e-01
-4.95582014e-01 3.90784480e-02 -5.45407176e-01 -4.05425191e-01
-1.01431596e+00 2.78105080e-01 4.20014262e-01 -1.35703683e-01
-1.07322872e+00 9.29906905e-01 7.63376236e-01 -3.66016179e-01
1.91109300e-01 -1.31214067e-01 5.90304323e-02 -3.13685089e-01
6.17329597e-01 2.80271977e-01 -2.09659532e-01 -7.79974461e-01
-2.69298375e-01 8.21333528e-01 -1.93850219e-01 -2.29631320e-01
1.25618315e+00 -4.91803288e-01 4.40368289e-03 -1.02579318e-01
7.05739856e-01 -2.38452610e-02 -2.01207900e+00 -2.85263985e-01
-6.85725093e-01 -8.24152410e-01 1.04077347e-02 -9.54158723e-01
-1.40706575e+00 5.62056065e-01 5.25558770e-01 -4.14913446e-01
1.24217546e+00 -2.63481468e-01 9.16712403e-01 -1.42770439e-01
4.77589697e-01 -1.00715816e+00 4.45235759e-01 -3.53225991e-02
1.12722838e+00 -9.95668292e-01 6.73793256e-02 -7.03030527e-01
-5.29833019e-01 9.44081843e-01 8.44228029e-01 -2.61389285e-01
2.60088921e-01 1.96551681e-01 4.45903605e-03 -1.73740104e-01
-5.60669899e-01 1.20878354e-01 7.25058615e-01 7.44750202e-01
2.59067804e-01 1.68458428e-02 -2.95909261e-03 6.11176968e-01
-2.17570171e-01 -1.42832339e-01 2.45870769e-01 6.30778432e-01
-2.06699818e-01 -1.11880314e+00 -5.50733626e-01 2.89048821e-01
-3.14038754e-01 -3.08380369e-02 -3.94903004e-01 4.33126628e-01
2.63843149e-01 7.63153255e-01 -1.03664257e-01 -3.80799830e-01
3.56723249e-01 -5.26434302e-01 9.82754052e-01 -2.94278264e-01
-5.01466155e-01 2.95011550e-01 2.87807006e-02 -8.69157970e-01
-3.66292298e-01 -4.41672623e-01 -7.92897046e-01 -6.41903222e-01
5.71203232e-02 -3.21317524e-01 2.73474842e-01 7.17775583e-01
2.46179998e-01 7.58461058e-01 6.20830119e-01 -1.12275517e+00
-2.33937040e-01 -7.22485840e-01 -5.02687156e-01 7.63750434e-01
3.72558564e-01 -8.66099954e-01 -1.00106508e-01 3.44244093e-01] | [11.956600189208984, -0.8515157103538513] |
b70530fb-b260-42c9-94d6-e7d014d82940 | tab2kg-semantic-table-interpretation-with | 2302.01150 | null | https://arxiv.org/abs/2302.01150v1 | https://arxiv.org/pdf/2302.01150v1.pdf | Tab2KG: Semantic Table Interpretation with Lightweight Semantic Profiles | Tabular data plays an essential role in many data analytics and machine learning tasks. Typically, tabular data does not possess any machine-readable semantics. In this context, semantic table interpretation is crucial for making data analytics workflows more robust and explainable. This article proposes Tab2KG - a novel method that targets at the interpretation of tables with previously unseen data and automatically infers their semantics to transform them into semantic data graphs. We introduce original lightweight semantic profiles that enrich a domain ontology's concepts and relations and represent domain and table characteristics. We propose a one-shot learning approach that relies on these profiles to map a tabular dataset containing previously unseen instances to a domain ontology. In contrast to the existing semantic table interpretation approaches, Tab2KG relies on the semantic profiles only and does not require any instance lookup. This property makes Tab2KG particularly suitable in the data analytics context, in which data tables typically contain new instances. Our experimental evaluation on several real-world datasets from different application domains demonstrates that Tab2KG outperforms state-of-the-art semantic table interpretation baselines. | ['Elena Demidova', 'Simon Gottschalk'] | 2023-02-02 | null | null | null | null | ['one-shot-learning'] | ['methodology'] | [ 3.00638109e-01 7.17331529e-01 -4.61593568e-01 -7.18235195e-01
-5.82971156e-01 -7.34868169e-01 4.56932068e-01 1.11548007e+00
1.83089226e-01 8.45411420e-01 8.50626752e-02 -3.04248661e-01
-4.52216297e-01 -1.38674629e+00 -1.05956829e+00 -1.07324801e-01
7.28405192e-02 1.17885959e+00 5.86550593e-01 -3.66784632e-01
3.68653610e-02 -8.21250230e-02 -2.12999511e+00 9.33681905e-01
9.80773628e-01 1.29328430e+00 -9.64299813e-02 -6.51094988e-02
-1.17789626e+00 9.61916029e-01 -4.87693131e-01 -8.33520412e-01
2.12376073e-01 -1.19059747e-02 -1.05432796e+00 -4.38020565e-02
3.39164466e-01 2.03555584e-01 6.19490221e-02 1.14778352e+00
-1.78426191e-01 1.12131834e-01 1.37020871e-01 -1.65384042e+00
-7.73313403e-01 9.54896390e-01 1.06478587e-01 -1.37078911e-01
4.73185122e-01 -4.14592803e-01 1.19915771e+00 -5.85320115e-01
9.70096052e-01 1.47853339e+00 6.38561130e-01 5.51502168e-01
-1.16300094e+00 -3.08696300e-01 3.59776437e-01 4.65007395e-01
-1.07266605e+00 -6.66372553e-02 5.08656919e-01 -2.74250716e-01
7.70055830e-01 4.19931322e-01 3.88891906e-01 9.04407084e-01
-2.52069443e-01 6.24907792e-01 7.15835094e-01 -4.07305300e-01
7.25201190e-01 5.26556492e-01 4.64243799e-01 3.58349025e-01
1.01723540e+00 -6.75094724e-01 -7.83247948e-01 1.08622611e-02
2.48360321e-01 1.24228761e-01 1.12214230e-01 -1.00628901e+00
-1.21698296e+00 4.79405195e-01 2.88050175e-01 6.84564263e-02
-1.73928872e-01 -4.86047678e-02 7.89191365e-01 1.62306815e-01
3.42048496e-01 6.97919548e-01 -7.89437830e-01 -1.83674365e-01
-2.88327128e-01 2.48453721e-01 1.02187192e+00 1.56621301e+00
9.60587502e-01 -4.98819619e-01 6.76487610e-02 6.99422657e-01
-1.42515302e-02 4.44923192e-01 2.47169599e-01 -6.58419609e-01
8.26588213e-01 1.54805374e+00 2.08600596e-01 -9.18057144e-01
-2.47234777e-01 1.10746779e-01 -2.91561872e-01 -2.27013797e-01
4.69545633e-01 5.49270034e-01 -9.60420728e-01 1.27663481e+00
5.55873871e-01 3.47826295e-02 5.12235582e-01 7.47122228e-01
1.12118220e+00 3.28693241e-01 3.41570199e-01 1.61104783e-01
1.68865812e+00 -6.61967158e-01 -1.12406552e+00 -3.76387328e-01
7.99994767e-01 -1.71277881e-01 1.51662791e+00 2.44176701e-01
-5.08790016e-01 -3.24718326e-01 -1.18843520e+00 -4.01499808e-01
-1.46093917e+00 -5.57769001e-01 8.55737031e-01 5.56348741e-01
-3.12562317e-01 6.19834304e-01 -6.65902019e-01 -8.19350898e-01
5.25411963e-01 1.73624367e-01 -5.44880271e-01 -2.05798417e-01
-1.43898368e+00 7.06250370e-01 1.23187554e+00 -1.55831560e-01
-3.30483019e-01 -1.03722966e+00 -1.22305596e+00 3.20314676e-01
1.23095679e+00 -6.31087363e-01 1.12974513e+00 -4.96891975e-01
-5.76140285e-01 8.89939666e-01 -2.44281083e-01 -8.09948027e-01
1.77551240e-01 -1.73491672e-01 -9.02533531e-01 -6.23202510e-02
3.46729428e-01 3.44226360e-01 3.24865490e-01 -1.48312974e+00
-7.75461257e-01 -6.54518902e-01 3.40995729e-01 8.78815278e-02
-4.80329335e-01 -5.74876010e-01 -4.94118959e-01 -3.91362429e-01
3.77137005e-01 -3.59310955e-01 1.06344350e-01 -1.58418324e-02
-5.38330615e-01 -2.42074922e-01 9.42401767e-01 -3.63936096e-01
1.38619483e+00 -1.78162169e+00 -2.36577600e-01 3.05976778e-01
2.65266806e-01 7.12494925e-02 4.49709833e-01 5.56467295e-01
-9.13193170e-03 3.48366827e-01 -3.36902350e-01 1.56915545e-01
2.99203128e-01 7.36461341e-01 -5.61297536e-01 -5.49560606e-01
1.14147954e-01 1.07306027e+00 -1.13038790e+00 -4.73039567e-01
3.27290326e-01 -2.46886313e-02 -3.01506102e-01 -2.66127121e-02
-9.83253658e-01 3.96517813e-02 -5.64640284e-01 8.72554660e-01
6.04122579e-01 -5.77403069e-01 5.82109749e-01 -4.12350655e-01
3.83693218e-01 3.76846731e-01 -1.32499516e+00 1.84518874e+00
-2.78849036e-01 6.83952048e-02 -7.00291634e-01 -8.69711578e-01
1.30642676e+00 3.36632803e-02 3.77228200e-01 -8.08874488e-01
-1.86271980e-01 4.85277742e-01 -7.01765001e-01 -5.69648266e-01
5.52250922e-01 -1.39368623e-01 -4.60575074e-01 1.74016371e-01
-4.29174006e-02 6.83419630e-02 5.94840705e-01 3.38275731e-01
9.54410732e-01 2.39644021e-01 5.91827929e-01 -3.66506577e-01
7.50460088e-01 7.45571613e-01 7.16941714e-01 6.75559938e-01
3.17245454e-01 3.61703873e-01 7.98231781e-01 -1.09650195e+00
-1.03147864e+00 -1.23523009e+00 1.66823305e-02 8.16160738e-01
6.74722791e-01 -9.41694558e-01 -8.13400149e-01 -1.06013358e+00
5.45815527e-01 1.06871569e+00 -8.43598783e-01 -1.91356152e-01
-2.67877072e-01 -3.33979666e-01 2.69554317e-01 6.92548335e-01
4.11543906e-01 -1.13356555e+00 -4.15040225e-01 2.72409856e-01
-3.67568165e-01 -1.47801852e+00 2.37925872e-01 2.85528392e-01
-8.64210546e-01 -1.56908023e+00 5.31030834e-01 -4.65871423e-01
8.26699734e-01 -2.53829770e-02 1.50179398e+00 7.84373935e-03
-1.55301258e-01 1.16397858e-01 -6.53932691e-01 -8.22682083e-01
-5.15373051e-01 2.35369310e-01 -1.80758372e-01 -3.73971686e-02
1.20648503e+00 -3.30898494e-01 -1.11563459e-01 3.37635577e-01
-1.23761594e+00 2.37840697e-01 1.43821454e-02 4.65516597e-01
8.44405711e-01 4.13576335e-01 6.48384631e-01 -1.69060814e+00
2.75392026e-01 -4.94302124e-01 -7.52163887e-01 8.12806606e-01
-9.70090270e-01 6.07367516e-01 9.78569567e-01 1.44339696e-01
-1.26343799e+00 -8.83614346e-02 4.82432812e-01 -1.76432237e-01
-2.91972607e-01 6.55138552e-01 -7.26253390e-01 6.79424226e-01
6.44013584e-01 -1.64657563e-01 -7.01201856e-02 -7.43976533e-01
3.47833961e-01 5.30044675e-01 7.57623017e-01 -8.04290116e-01
1.12878239e+00 8.40187192e-01 1.93403527e-01 -7.40136430e-02
-1.37882507e+00 -5.06870329e-01 -8.94332886e-01 1.11810528e-01
8.79268765e-01 -6.94234371e-01 -7.66228557e-01 -8.11846331e-02
-6.26567304e-01 -3.30020301e-02 -6.57489121e-01 -6.01321794e-02
-6.80710435e-01 -1.01466402e-01 2.61877235e-02 -5.26233315e-01
-2.04437062e-01 -6.84829891e-01 1.08141637e+00 3.21402140e-02
-2.63782531e-01 -1.12314439e+00 -4.15672570e-01 6.53609574e-01
6.09393455e-02 4.83787060e-01 1.50200307e+00 -1.25428879e+00
-8.26168597e-01 -1.40536577e-01 -2.94454962e-01 -1.10614695e-01
4.83475536e-01 -4.41585362e-01 -9.01848912e-01 1.99647978e-01
-5.32621801e-01 -1.38844699e-01 3.45577925e-01 -4.54190075e-01
1.62629509e+00 -4.90640223e-01 -3.40281337e-01 4.16763246e-01
1.58553541e+00 5.26469350e-01 6.81751251e-01 8.95337403e-01
8.97060037e-01 8.84368896e-01 1.23419511e+00 5.12998700e-01
6.80347383e-01 6.12627685e-01 5.87768018e-01 1.47970870e-01
1.00682929e-01 -9.28901255e-01 -3.77856970e-01 2.32402667e-01
5.45095682e-01 -3.43193620e-01 -1.21311736e+00 6.44102037e-01
-2.36315799e+00 -8.62254262e-01 -2.39551708e-01 2.40226912e+00
7.82928407e-01 3.55467260e-01 -1.09873280e-01 4.45665359e-01
7.32349515e-01 -2.68348485e-01 -7.34758377e-01 -3.45109552e-01
-2.11603746e-01 1.00169316e-01 6.94335997e-01 2.11794987e-01
-1.07253563e+00 1.11627233e+00 4.70298767e+00 5.20336807e-01
-4.46114093e-01 -1.42413393e-01 3.65812927e-01 1.10909961e-01
-6.07114911e-01 2.21266776e-01 -9.45899606e-01 4.06557292e-01
8.05026829e-01 -6.96004689e-01 4.48290318e-01 1.03555548e+00
-5.93980365e-02 5.50919259e-03 -1.56367898e+00 9.45665359e-01
-7.39489645e-02 -1.70534539e+00 7.87378907e-01 -1.51803285e-01
2.35065624e-01 -8.34650517e-01 -2.99086362e-01 3.40837449e-01
2.38483325e-01 -1.10789812e+00 6.73119426e-01 3.14978659e-01
8.14515650e-01 -7.58669734e-01 8.31224024e-01 -7.42575675e-02
-1.38294685e+00 -2.02693388e-01 -4.18396533e-01 1.61464244e-01
-8.10660571e-02 4.82201993e-01 -1.18249130e+00 1.02319658e+00
1.14526737e+00 1.03175938e+00 -8.36790740e-01 5.70002377e-01
-2.80889332e-01 1.24436133e-01 -3.05142682e-02 2.64310181e-01
-2.97007840e-02 -8.44537169e-02 2.04907253e-01 7.36418545e-01
2.38683566e-01 6.74162805e-02 1.53266182e-02 8.88913155e-01
-2.60361165e-01 -1.98042821e-04 -7.19177663e-01 -1.58426926e-01
8.59848619e-01 1.01310265e+00 -8.78622353e-01 -7.66552925e-01
-5.10073304e-01 5.40047050e-01 1.83894411e-01 9.11896676e-02
-6.11020088e-01 -5.35491467e-01 1.04924273e+00 4.53943253e-01
2.50278562e-01 3.11602056e-01 -8.67888570e-01 -1.19001389e+00
5.56550086e-01 -8.42434943e-01 1.09308159e+00 -9.08876002e-01
-1.26173604e+00 4.66283977e-01 3.52515548e-01 -1.28292525e+00
-3.39460790e-01 -6.71075523e-01 7.83913508e-02 3.83235246e-01
-1.44235492e+00 -1.12101376e+00 -9.38966632e-01 5.81765175e-01
5.35572648e-01 1.15740381e-01 9.89499211e-01 2.16847435e-01
-3.08577955e-01 2.93956071e-01 -2.09538460e-01 1.33357912e-01
6.92469299e-01 -1.64139867e+00 8.53196442e-01 7.13288844e-01
2.29522381e-02 7.82370627e-01 8.19707990e-01 -9.46437955e-01
-1.46149337e+00 -1.40072525e+00 1.00707197e+00 -8.25541317e-01
8.48004937e-01 -8.69747579e-01 -1.40662408e+00 9.91389811e-01
-2.86575109e-01 2.79894710e-01 6.99291587e-01 2.08300963e-01
-7.08736718e-01 -6.69104040e-01 -1.34667075e+00 4.73691761e-01
1.53311622e+00 -3.66877466e-01 -9.63923752e-01 2.63776094e-01
8.85092556e-01 -4.35690671e-01 -1.12626052e+00 5.47954857e-01
3.59307438e-01 -6.78949237e-01 9.77881312e-01 -1.02194798e+00
2.88571626e-01 -7.63667226e-01 -2.64795542e-01 -1.07768655e+00
2.51244128e-01 -3.76340330e-01 -5.04180968e-01 1.23273063e+00
6.14664733e-01 -5.58976054e-01 9.51270878e-01 1.07520127e+00
-2.28832569e-02 -3.00905764e-01 -6.18862092e-01 -1.15744889e+00
-4.58826125e-01 -3.06713521e-01 1.43978119e+00 1.14725935e+00
3.05301845e-01 2.33766343e-02 3.24115828e-02 3.18759769e-01
7.91647851e-01 4.01228011e-01 9.01915967e-01 -1.83639693e+00
3.69197845e-01 1.20638236e-01 -7.71514833e-01 -2.35346764e-01
9.43179280e-02 -1.10264814e+00 -2.52514184e-01 -1.96818602e+00
6.88046888e-02 -7.42854595e-01 -4.94807988e-01 8.96153986e-01
-8.36555883e-02 -2.16217384e-01 4.43094298e-02 -1.04156872e-02
-8.05079341e-01 2.15262696e-01 8.18956852e-01 -1.84511065e-01
-1.29188046e-01 -5.27808368e-01 -9.90054190e-01 4.75488812e-01
7.73578644e-01 -5.80851555e-01 -1.04876685e+00 -3.64173621e-01
4.65385109e-01 -3.39465529e-01 1.79223076e-01 -1.04239881e+00
2.15707108e-01 -4.00810510e-01 8.51801634e-02 -4.99218524e-01
8.95579904e-02 -1.24715364e+00 4.43248898e-01 3.43042724e-02
-3.38322163e-01 -2.80003157e-02 3.57728899e-01 5.74087441e-01
-4.39440012e-01 7.56388530e-03 3.69487137e-01 -2.21408859e-01
-1.43499029e+00 1.58368275e-01 2.97830254e-01 4.98067111e-01
1.19937921e+00 -4.38601822e-01 -6.94352448e-01 1.09083615e-01
-7.82642484e-01 4.65872318e-01 7.54044831e-01 1.03137338e+00
4.71005708e-01 -1.32113647e+00 -1.64121792e-01 3.13735425e-01
1.10989833e+00 4.32303756e-01 -6.34162948e-02 1.60276040e-01
-6.16349876e-01 6.01868093e-01 -3.65334243e-01 -4.15840149e-01
-1.10866034e+00 1.09046721e+00 1.42416880e-01 -1.56023785e-01
-8.27159345e-01 2.73622662e-01 1.57434955e-01 -7.07179487e-01
2.31851503e-01 -4.85473424e-01 -1.15851521e-01 1.27893180e-01
5.34920454e-01 1.32245034e-01 6.85556412e-01 -1.97386872e-02
-5.91172516e-01 1.04338326e-01 -1.11726917e-01 3.66347849e-01
1.26467049e+00 -1.56553105e-01 -1.54772684e-01 8.15013707e-01
6.32243097e-01 -2.99803406e-01 -7.67901361e-01 -4.21135575e-01
9.60524619e-01 -7.49357343e-01 -5.37661552e-01 -1.09259593e+00
-6.79101527e-01 5.05170166e-01 -1.08227424e-01 3.25240701e-01
9.81601775e-01 1.68971226e-01 8.33374977e-01 6.60190284e-01
7.51244962e-01 -1.19559455e+00 -1.24410473e-01 2.29018286e-01
5.69039881e-01 -1.35725999e+00 -1.95913568e-01 -1.10334718e+00
-6.20614111e-01 1.22916818e+00 1.03409231e+00 5.40125370e-01
2.67426729e-01 2.08409101e-01 1.46628022e-01 -5.43238699e-01
-9.43383813e-01 -3.22388411e-01 2.27084637e-01 7.97592521e-01
1.33834809e-01 5.42420447e-02 6.31240979e-02 7.72233546e-01
-2.83965260e-01 6.67042062e-02 4.91709858e-01 1.16245234e+00
-4.74961191e-01 -1.41982579e+00 -2.56494701e-01 6.63869798e-01
-1.36838229e-02 -6.38382956e-02 -5.41097522e-01 9.41466987e-01
1.91289753e-01 9.23031747e-01 2.67441332e-01 -2.30485067e-01
6.68335557e-01 5.07574320e-01 -5.50217554e-02 -9.06307518e-01
-3.55311632e-01 -8.15299630e-01 3.81349623e-01 -7.78824806e-01
-8.22168067e-02 -4.51351404e-01 -1.77474785e+00 -3.46014202e-01
2.22203091e-01 4.49846894e-01 6.01884782e-01 8.46827209e-01
5.59722483e-01 6.14437640e-01 2.74502300e-02 3.38580996e-01
-7.94320628e-02 -3.02568436e-01 -4.72083718e-01 1.06094491e+00
1.64985042e-02 -9.58785832e-01 6.67888224e-02 2.86089957e-01] | [9.342069625854492, 8.00100040435791] |
f266d564-6895-4ef7-8052-096632485d82 | cellular-segmentation-and-composition-in | 2203.02510 | null | https://arxiv.org/abs/2203.02510v1 | https://arxiv.org/pdf/2203.02510v1.pdf | Cellular Segmentation and Composition in Routine Histology Images using Deep Learning | Identification and quantification of nuclei in colorectal cancer haematoxylin \& eosin (H\&E) stained histology images is crucial to prognosis and patient management. In computational pathology these tasks are referred to as nuclear segmentation, classification and composition and are used to extract meaningful interpretable cytological and architectural features for downstream analysis. The CoNIC challenge poses the task of automated nuclei segmentation, classification and composition into six different types of nuclei from the largest publicly known nuclei dataset - Lizard. In this regard, we have developed pipelines for the prediction of nuclei segmentation using HoVer-Net and ALBRT for cellular composition. On testing on the preliminary test set, HoVer-Net achieved a PQ of 0.58, a PQ+ of 0.58 and finally a mPQ+ of 0.35. For the prediction of cellular composition with ALBRT on the preliminary test set, we achieved an overall $R^2$ score of 0.53, consisting of 0.84 for lymphocytes, 0.70 for epithelial cells, 0.70 for plasma and .060 for eosinophils. | ['Adam Shephard', 'Manahil Raza', 'Srijay Deshpande', 'Raja Muhammad Saad Bashir', 'Muhammad Dawood'] | 2022-03-04 | null | null | null | null | ['nuclear-segmentation'] | ['medical'] | [-2.14269999e-02 2.78307050e-01 8.95233974e-02 -3.52758467e-02
-8.94590616e-01 -7.22749054e-01 3.46354663e-01 9.61738050e-01
-7.79383957e-01 8.22689414e-01 -7.70203546e-02 -3.40038151e-01
-1.23122232e-02 -8.35382879e-01 -2.35309768e-02 -1.20450008e+00
-3.87871545e-03 9.30263460e-01 1.38199121e-01 1.71997979e-01
2.89154887e-01 7.02621877e-01 -1.20469213e+00 2.93584198e-01
4.24494296e-01 7.89565742e-01 -1.78650379e-01 1.43579793e+00
7.66400099e-02 7.07488716e-01 -2.39504263e-01 -3.24418575e-01
-8.76738727e-02 -2.76759207e-01 -9.38469589e-01 -8.76169875e-02
1.95008013e-02 8.86683259e-03 1.74124241e-01 8.42425883e-01
7.28092551e-01 -2.19071701e-01 1.06018627e+00 -8.49925578e-01
2.22294152e-01 5.80915451e-01 -5.99334300e-01 4.82078731e-01
5.49999103e-02 4.14291650e-01 1.01492572e+00 -6.68959677e-01
9.79531884e-01 5.60840666e-01 9.59404528e-01 3.56323093e-01
-1.38634896e+00 -3.37556392e-01 -8.02557468e-01 -2.27448314e-01
-1.58091915e+00 -2.93841153e-01 -1.49198040e-01 -6.47016406e-01
1.02534127e+00 6.05586708e-01 9.74788725e-01 1.27152547e-01
1.94254383e-01 4.90292758e-01 1.13263130e+00 -2.28294760e-01
4.37726647e-01 1.12788469e-01 1.89318821e-01 6.69553757e-01
3.70606422e-01 -2.09736183e-01 -1.39738411e-01 -5.16944677e-02
4.65244442e-01 -1.37350217e-01 -6.24280721e-02 3.55353713e-01
-1.27757800e+00 6.49062872e-01 2.55580962e-01 3.64969790e-01
-2.51340568e-01 9.74190906e-02 6.31920516e-01 -1.63655639e-01
2.82774359e-01 5.12329876e-01 -3.72438461e-01 9.69061570e-04
-1.07242310e+00 1.32661626e-01 7.55645096e-01 4.36036676e-01
5.59108496e-01 -3.75425845e-01 -1.66098088e-01 6.74385369e-01
2.71838158e-01 4.14456099e-01 5.93307018e-01 -8.63492250e-01
-5.57631373e-01 8.84376645e-01 -1.65640488e-01 -7.57993996e-01
-1.09190524e+00 -3.63338441e-01 -1.12391138e+00 2.34094169e-02
8.40921342e-01 6.08422756e-02 -7.62223065e-01 1.20023072e+00
5.52481055e-01 -1.65259734e-01 5.12518026e-02 7.12506413e-01
1.02541840e+00 3.48112404e-01 3.15199792e-01 -1.09393097e-01
1.65930736e+00 -3.98912758e-01 -3.40777248e-01 1.89406231e-01
1.10877395e+00 -8.59780371e-01 7.32634783e-01 5.08241296e-01
-9.72448349e-01 -8.42286199e-02 -8.07990849e-01 -1.11930914e-01
-4.89664406e-01 2.40471512e-01 6.31612539e-01 5.66648066e-01
-1.38758099e+00 4.68757957e-01 -9.83572006e-01 -5.70621312e-01
6.68964803e-01 7.67998815e-01 -5.99937797e-01 6.70300052e-02
-4.69853640e-01 7.27903187e-01 4.77658778e-01 1.08065106e-01
-4.96347398e-01 -8.59573781e-01 -5.40067494e-01 -1.08708821e-01
-4.45956439e-01 -8.02877843e-01 1.04315495e+00 -3.83989096e-01
-1.14872670e+00 1.57138538e+00 -4.52572899e-03 -4.78766173e-01
3.98342311e-01 6.04648948e-01 1.47349000e-01 4.33632910e-01
1.35837495e-01 9.20199990e-01 -9.90888849e-02 -9.10896540e-01
-1.04968357e+00 -3.32277685e-01 -4.07199740e-01 1.56282917e-01
1.47804990e-01 -1.16690338e-01 -2.54037887e-01 -2.96687633e-01
2.44382750e-02 -8.26050043e-01 -4.55874592e-01 -1.85667549e-03
-3.73463392e-01 -2.11902320e-01 1.89227879e-01 -9.37611043e-01
9.70883489e-01 -1.96751714e+00 -7.22001269e-02 6.35564029e-01
6.96887016e-01 4.09918837e-02 2.69017786e-01 4.39626770e-03
2.33283974e-02 5.28913081e-01 -1.13362700e-01 -1.40351847e-01
-6.47274852e-02 -7.66809881e-02 5.63577712e-01 9.41449702e-01
1.89247519e-01 8.25597703e-01 -1.10497248e+00 -8.65046799e-01
1.44183695e-01 6.01917148e-01 -5.75332880e-01 4.69361581e-02
2.14697123e-02 6.32516816e-02 5.46529256e-02 1.05196965e+00
5.38434684e-01 -3.42884302e-01 3.93390477e-01 -9.05236006e-02
3.48320454e-02 -2.58514673e-01 -8.01394641e-01 9.21718776e-01
1.11585669e-01 5.63342035e-01 4.29837614e-01 -3.71430039e-01
7.90951788e-01 2.91547567e-01 5.72488606e-01 -2.81599462e-01
7.42774367e-01 5.07690966e-01 2.42357299e-01 -1.67879179e-01
3.75591666e-01 -6.70297980e-01 -4.40571047e-02 3.16308111e-01
1.50928825e-01 -3.00236553e-01 7.62389183e-01 1.43274918e-01
1.43376887e+00 -4.14737850e-01 5.90620160e-01 -6.37784362e-01
7.02902734e-01 3.19150686e-01 4.02039170e-01 5.05222194e-02
-5.50451756e-01 6.84353411e-01 9.42722321e-01 -2.47531280e-01
-1.11031425e+00 -7.61972785e-01 -3.31810892e-01 7.39001274e-01
-2.35487521e-01 -3.84896100e-01 -7.85348296e-01 -4.56289291e-01
-1.59981638e-01 2.53852874e-01 -7.62552261e-01 1.73483178e-01
-2.86050767e-01 -1.33029842e+00 8.08452129e-01 4.80944008e-01
1.30352308e-03 -1.09039283e+00 -4.72243071e-01 -1.56853851e-02
-1.34414807e-01 -7.75502086e-01 -8.23060200e-02 5.39896011e-01
-6.52823031e-01 -1.44562542e+00 -6.25963509e-01 -8.26028049e-01
1.05937576e+00 -3.22004110e-01 1.13090658e+00 6.60896063e-01
-7.75403440e-01 1.14980023e-02 -1.77943051e-01 -2.81925797e-01
-4.71987456e-01 1.28373817e-01 -1.48653373e-01 -5.03954947e-01
2.61093736e-01 -2.93293238e-01 -9.28272247e-01 1.38429731e-01
-9.58048344e-01 -1.14107147e-01 6.44501030e-01 1.04797578e+00
1.31421804e+00 2.06535891e-01 1.96989611e-01 -1.14176834e+00
1.72589663e-02 -4.61514682e-01 -5.18369019e-01 -1.91301219e-02
-4.47987616e-01 -4.03584659e-01 8.78010631e-01 1.23379333e-02
-4.43013191e-01 1.15543850e-01 -3.86781394e-01 1.59730256e-01
-2.09975049e-01 4.29314524e-01 2.67186761e-01 -2.49057580e-02
5.76042771e-01 -7.91704953e-02 2.21575275e-01 1.10459819e-01
-9.36782509e-02 6.12348795e-01 4.88909811e-01 3.22810523e-02
5.09755194e-01 7.15030015e-01 5.01509309e-01 -9.08547342e-01
-5.54360986e-01 -9.79659379e-01 -6.72924697e-01 -1.83941219e-02
9.18576062e-01 -7.77391970e-01 -1.12042260e+00 4.78266448e-01
-3.85052741e-01 -6.72201633e-01 -3.17703307e-01 3.50032717e-01
-5.32741606e-01 2.09852174e-01 -9.75865901e-01 -4.55886781e-01
-8.95895541e-01 -9.84236479e-01 9.93310630e-01 4.06336814e-01
-8.15852225e-01 -1.22345281e+00 4.64418203e-01 4.08259749e-01
1.39830977e-01 5.45563102e-01 9.50053155e-01 -8.58943582e-01
-4.27032672e-02 -4.59485471e-01 -2.19353482e-01 -8.69106799e-02
-2.47358277e-01 6.15640283e-01 -9.44026291e-01 -2.12118939e-01
-4.65630949e-01 -1.49047121e-01 9.51250136e-01 5.49012721e-01
8.55816305e-01 -1.34693757e-01 -5.59295177e-01 8.63650084e-01
1.63107431e+00 1.06591471e-01 6.56980872e-01 4.13917363e-01
2.45055646e-01 5.82610726e-01 6.88586295e-01 3.67153227e-01
2.25995243e-01 -1.39656097e-01 3.12300533e-01 -4.12017405e-01
-1.12217724e-01 1.91317454e-01 -1.48289889e-01 3.81630749e-01
5.26640266e-02 -1.60349786e-01 -1.46161211e+00 8.74600887e-01
-1.21699166e+00 -7.83232689e-01 -4.14937556e-01 1.62412381e+00
9.16067660e-01 1.94701731e-01 2.38576248e-01 3.46350253e-01
5.95195591e-01 -5.58094621e-01 -2.52086461e-01 -3.56505722e-01
-1.31717801e-01 5.66068053e-01 5.71635306e-01 4.76312518e-01
-1.10808635e+00 6.93027854e-01 6.63548279e+00 9.35477257e-01
-1.09885478e+00 -4.19130862e-01 1.35176194e+00 -7.31302518e-03
5.84273487e-02 -1.89733118e-01 -1.04313290e+00 4.08611655e-01
1.04216850e+00 2.70637795e-02 1.73246726e-01 3.64785016e-01
1.60473377e-01 -6.31711304e-01 -7.82478511e-01 6.71917200e-01
-2.26330981e-01 -1.32130885e+00 -3.30898315e-01 3.56045306e-01
5.43977380e-01 2.14090839e-01 -1.60443455e-01 1.92556128e-01
3.61543357e-01 -1.51351631e+00 1.69422895e-01 6.23753369e-01
1.14538991e+00 -8.39078248e-01 1.58891070e+00 1.75494954e-01
-9.76440132e-01 2.50062972e-01 -3.93097103e-01 3.03996474e-01
-2.78176516e-01 6.78863525e-01 -1.69566786e+00 1.52881742e-01
4.72353995e-01 4.25233603e-01 -7.14294434e-01 1.16701591e+00
1.77174255e-01 6.34562314e-01 -6.64429188e-01 -2.24316433e-01
2.31634825e-02 3.94981727e-02 1.37940543e-02 1.59492421e+00
1.99950784e-01 3.39046299e-01 -1.86777294e-01 4.19769466e-01
3.32500562e-02 2.98869014e-01 2.67018974e-01 -1.69171691e-01
2.78477818e-01 2.07609248e+00 -1.77406871e+00 -2.80777842e-01
3.12885582e-01 2.59765685e-01 1.58159614e-01 -2.30856732e-01
-5.82960904e-01 -3.90711427e-01 2.07145140e-01 3.41196150e-01
3.86266112e-02 4.25295562e-01 -4.63824511e-01 -6.08768463e-01
-8.01398754e-01 -5.63335061e-01 7.24735260e-01 -4.13267881e-01
-1.38812447e+00 3.49663585e-01 -6.18976593e-01 -8.75260651e-01
5.65634742e-02 -9.00956154e-01 -6.93706751e-01 8.37567627e-01
-1.30477357e+00 -1.31240380e+00 -3.56756955e-01 -8.26201886e-02
-2.68768854e-02 1.97723523e-01 1.06513262e+00 -2.92749889e-02
-6.76145077e-01 4.24425244e-01 1.50615960e-01 3.47920567e-01
6.09388173e-01 -1.92369485e+00 -2.34675303e-01 2.73275226e-01
-5.34406185e-01 4.18189436e-01 6.43384874e-01 -4.69500601e-01
-1.02694619e+00 -1.28130996e+00 1.00048804e+00 -4.32594955e-01
6.17076576e-01 8.48016888e-02 -5.45166492e-01 3.46987933e-01
-3.66599374e-02 2.04181299e-01 1.70039487e+00 -2.45156124e-01
2.07059830e-01 1.48200810e-01 -1.58641505e+00 4.32796299e-01
1.04117595e-01 -1.98124617e-01 3.75812054e-02 3.45218748e-01
-5.88953942e-02 -5.47909856e-01 -1.52381027e+00 1.15483940e-01
6.56519473e-01 -9.15787458e-01 7.96261847e-01 -9.28655565e-02
3.68175149e-01 -5.50573528e-01 6.69567659e-02 -8.41372669e-01
-5.16523540e-01 -1.49227247e-01 3.26081097e-01 9.48353231e-01
4.27418441e-01 -2.32892141e-01 1.18208265e+00 3.33589256e-01
-1.64686337e-01 -1.13154495e+00 -9.27021086e-01 1.09495416e-01
2.18645766e-01 9.19875514e-04 1.57948956e-01 7.41901636e-01
5.38686037e-01 2.43229389e-01 5.78315437e-01 -2.12297171e-01
4.65604097e-01 -2.20063418e-01 6.70138717e-01 -1.15140939e+00
-1.52389631e-01 -8.33025694e-01 -6.04656160e-01 -1.15281664e-01
-5.82849123e-02 -1.24468243e+00 -5.03749661e-02 -1.65115869e+00
7.73164988e-01 -3.98886144e-01 -2.66191065e-01 5.80656648e-01
-2.83642828e-01 9.38985109e-01 -3.03313851e-01 2.72296011e-01
-6.16604865e-01 -2.74872929e-01 9.67137039e-01 -5.87525428e-04
1.73677050e-03 -1.88065067e-01 -9.19796348e-01 9.00842071e-01
1.04931533e+00 -2.79420346e-01 2.19504058e-01 5.38441837e-01
4.72632766e-01 -5.05038071e-04 2.96004206e-01 -8.66372466e-01
1.08063579e-01 -4.73517329e-02 9.31561887e-01 -1.08968997e+00
1.25623107e-01 -3.83874416e-01 3.10265750e-01 9.43900347e-01
-8.99195224e-02 -2.40075424e-01 1.05986625e-01 1.93150118e-01
1.27477513e-03 -1.78833917e-01 1.06403399e+00 -3.90579045e-01
-1.02017328e-01 5.02521135e-02 -7.40301967e-01 -1.95284054e-01
1.08766496e+00 -6.04829848e-01 -5.36372542e-01 1.72584593e-01
-9.66121614e-01 5.88210344e-01 8.02992821e-01 -7.23058224e-01
4.46032256e-01 -8.19267929e-01 -7.17683017e-01 1.75460184e-03
9.08058062e-02 5.18658280e-01 3.59601080e-01 1.35502434e+00
-1.34994161e+00 4.64619458e-01 -1.64307788e-01 -8.28213751e-01
-1.69417524e+00 9.01547913e-03 6.03069961e-01 -9.32716906e-01
-7.43841305e-02 1.21483207e+00 1.03685960e-01 -5.51392674e-01
-3.40844184e-01 -5.09561062e-01 -5.72388768e-01 3.75679433e-01
4.21975821e-01 6.56327903e-01 1.10264361e-01 -8.89979959e-01
-3.03397536e-01 3.21694285e-01 -1.17916171e-03 1.53921217e-01
1.46219993e+00 3.51534374e-02 -7.87895560e-01 2.49796659e-01
1.06162143e+00 3.42371836e-02 -6.95826232e-01 4.03956473e-01
9.91633460e-02 -6.25445321e-02 1.80621773e-01 -8.87809277e-01
-8.27049971e-01 5.29915631e-01 4.81915325e-01 1.43924206e-01
1.08280849e+00 -1.67783350e-01 5.26409388e-01 6.78688586e-02
-2.00167403e-01 -9.37568963e-01 -3.00111860e-01 5.87605655e-01
1.84857160e-01 -9.56572235e-01 2.96525866e-01 -4.86931533e-01
-3.50838065e-01 1.28744853e+00 4.52264458e-01 -1.76496524e-02
5.74971139e-01 5.43157160e-01 6.21084198e-02 -4.53328371e-01
-9.35924470e-01 -8.36703479e-02 4.20905501e-02 5.02891362e-01
8.13545763e-01 1.90313578e-01 -3.73089969e-01 8.22293639e-01
-7.04582989e-01 -1.73121803e-02 5.89096308e-01 7.10648835e-01
-5.28990746e-01 -7.65139699e-01 -4.15713310e-01 8.21829617e-01
-9.99921799e-01 6.78622946e-02 -5.78067005e-01 8.48118961e-01
1.99606702e-01 4.39433545e-01 3.73029888e-01 -1.43499270e-01
-7.28977397e-02 -1.56577364e-01 2.52833873e-01 -5.79611480e-01
-1.07672060e+00 5.86393893e-01 1.35838866e-01 -7.97673613e-02
-4.56828713e-01 -7.37411976e-01 -1.87580884e+00 -6.17244720e-01
-4.48494494e-01 2.62498528e-01 2.09167719e-01 5.09938478e-01
-2.25497186e-01 5.34249663e-01 3.07135969e-01 -5.01949072e-01
-1.10424329e-02 -1.02980137e+00 -1.11942995e+00 1.69185236e-01
3.33307505e-01 -4.25483333e-03 -6.52245283e-01 4.42531228e-01] | [15.057774543762207, -3.1275789737701416] |
69125399-4a06-428a-8ae0-73286eeba4cb | synthesizing-diverse-human-motions-in-3d | 2305.12411 | null | https://arxiv.org/abs/2305.12411v2 | https://arxiv.org/pdf/2305.12411v2.pdf | Synthesizing Diverse Human Motions in 3D Indoor Scenes | We present a novel method for populating 3D indoor scenes with virtual humans that can navigate the environment and interact with objects in a realistic manner. Existing approaches rely on high-quality training sequences that capture a diverse range of human motions in 3D scenes. However, such motion data is costly, difficult to obtain and can never cover the full range of plausible human-scene interactions in complex indoor environments. To address these challenges, we propose a reinforcement learning-based approach to learn policy networks that predict latent variables of a powerful generative motion model that is trained on a large-scale motion capture dataset (AMASS). For navigating in a 3D environment, we propose a scene-aware policy training scheme with a novel collision avoidance reward function. Combined with the powerful generative motion model, we can synthesize highly diverse human motions navigating 3D indoor scenes, meanwhile effectively avoiding obstacles. For detailed human-object interactions, we carefully curate interaction-aware reward functions by leveraging a marker-based body representation and the signed distance field (SDF) representation of the 3D scene. With a number of important training design schemes, our method can synthesize realistic and diverse human-object interactions (e.g.,~sitting on a chair and then getting up) even for out-of-distribution test scenarios with different object shapes, orientations, starting body positions, and poses. Experimental results demonstrate that our approach outperforms state-of-the-art human-scene interaction synthesis frameworks in terms of both motion naturalness and diversity. Video results are available on the project page: https://zkf1997.github.io/DIMOS. | ['Siyu Tang', 'Thabo Beeler', 'Shaofei Wang', 'Yan Zhang', 'Kaifeng Zhao'] | 2023-05-21 | null | null | null | null | ['human-object-interaction-detection'] | ['computer-vision'] | [-4.74822409e-02 -2.25558370e-01 8.22947174e-02 -1.93549737e-01
-4.39173818e-01 -4.26365018e-01 4.72052604e-01 -5.37395358e-01
-2.95720100e-01 6.83723509e-01 2.84808964e-01 -1.91134661e-01
-4.67107669e-02 -7.67559111e-01 -9.34664309e-01 -4.94903058e-01
-1.25358239e-01 6.63960576e-01 3.19831014e-01 -4.08676773e-01
-2.07360134e-01 5.69705784e-01 -1.59700751e+00 -1.20673947e-01
7.45296180e-01 5.03279150e-01 6.28038585e-01 1.12731957e+00
3.08438122e-01 6.88011408e-01 -5.39270937e-01 -6.00453839e-02
3.75029504e-01 -3.83496225e-01 -5.26525617e-01 2.32876763e-01
2.32516035e-01 -6.33882105e-01 -5.29940784e-01 4.27914649e-01
8.04708540e-01 6.46062493e-01 4.56295490e-01 -1.32492399e+00
-1.72109783e-01 1.26477219e-02 -3.26384425e-01 -1.18528448e-01
1.00794458e+00 6.91609859e-01 4.60006982e-01 -6.44856632e-01
9.84718204e-01 1.41490710e+00 4.46263373e-01 9.92182553e-01
-1.03972137e+00 -4.03891206e-01 2.81247020e-01 -1.39323166e-02
-1.16002214e+00 -2.13209480e-01 8.02975059e-01 -4.43127692e-01
9.66389358e-01 4.08220947e-01 1.09847474e+00 1.75646484e+00
3.42179269e-01 9.68616247e-01 5.76153040e-01 -1.41665801e-01
3.43297631e-01 -3.70514035e-01 -5.05897462e-01 7.15125203e-01
1.68952852e-01 2.48758689e-01 -4.29362178e-01 -4.93133813e-02
1.40932333e+00 8.76348540e-02 -3.89177412e-01 -1.15543056e+00
-1.57715142e+00 5.10312676e-01 3.80751669e-01 -1.36159554e-01
-3.82971883e-01 5.09985447e-01 1.08509749e-01 -2.81634301e-01
-5.46228848e-02 2.13398501e-01 -3.41213286e-01 -3.90570819e-01
-4.08673137e-01 1.08536553e+00 7.03825057e-01 1.32722807e+00
2.89332837e-01 1.40889809e-01 -2.95413136e-01 5.54217219e-01
3.76313448e-01 7.97673225e-01 2.78255790e-02 -1.47820807e+00
5.73125482e-01 2.22325444e-01 6.19070292e-01 -8.94411087e-01
-4.77355629e-01 -9.75789428e-02 -8.58249962e-01 3.96495014e-01
4.71327901e-01 -3.29244345e-01 -9.09078240e-01 1.75522614e+00
9.60938752e-01 2.26436362e-01 -1.85431436e-01 1.22554743e+00
7.84005284e-01 5.71436286e-01 6.07133284e-02 2.59092808e-01
1.04111946e+00 -1.15323889e+00 -5.64031243e-01 -3.00508916e-01
4.80578929e-01 -4.94931787e-01 1.26904988e+00 2.00630240e-02
-1.24757588e+00 -7.99744487e-01 -7.78762341e-01 9.50343162e-03
1.51547164e-01 -1.76544398e-01 7.30905950e-01 4.73232597e-01
-8.76394928e-01 5.57672024e-01 -1.19896710e+00 -4.19214666e-01
1.91305444e-01 2.84496158e-01 -3.48760337e-01 -2.71166265e-01
-9.26226020e-01 6.44065022e-01 1.14056624e-01 6.07593311e-03
-1.21820652e+00 -6.09222114e-01 -1.18103027e+00 -2.88352430e-01
4.77289081e-01 -1.55543232e+00 1.30462778e+00 -2.71120459e-01
-1.72898304e+00 5.19484818e-01 -1.80979475e-01 -1.08254282e-02
9.81385767e-01 -6.56495333e-01 1.01964874e-02 3.15163583e-02
2.01128721e-01 8.62840593e-01 4.60369557e-01 -1.64634693e+00
-3.98168743e-01 -1.08826369e-01 1.54288396e-01 6.65016055e-01
3.80173504e-01 -4.06335354e-01 -6.97110772e-01 -6.70357645e-01
-8.94718319e-02 -1.22851014e+00 -8.36540818e-01 2.46769294e-01
-4.70904320e-01 2.97092617e-01 7.15224862e-01 -3.78988534e-01
8.53421152e-01 -1.80746114e+00 5.20757973e-01 7.42312074e-02
-1.21230841e-01 6.12826198e-02 7.72634596e-02 3.61561656e-01
3.85056317e-01 -2.34289579e-02 -1.95179418e-01 -6.50487363e-01
1.70205072e-01 4.71536845e-01 -8.59245509e-02 2.61734843e-01
-1.14488386e-01 1.07128775e+00 -1.22058666e+00 -4.08319861e-01
6.01737618e-01 7.44545102e-01 -8.79497409e-01 5.08738756e-01
-3.45184088e-01 1.31803012e+00 -6.70335233e-01 6.39181197e-01
5.31973660e-01 -2.51655847e-01 1.74154148e-01 3.30543965e-01
1.92612290e-01 -2.22414974e-02 -1.48036551e+00 2.30401826e+00
-4.32898521e-01 1.00538962e-01 2.26526195e-03 -3.64904195e-01
6.99607849e-01 1.65397346e-01 5.45256615e-01 -4.88110900e-01
1.33180514e-01 6.90079555e-02 -3.21280748e-01 -6.50684476e-01
5.87200344e-01 6.06779233e-02 -3.56157780e-01 1.60668716e-01
-1.98904887e-01 -4.83570844e-01 -1.48527492e-02 4.34980467e-02
1.20137691e+00 7.90990233e-01 9.64730456e-02 6.42720833e-02
3.00053507e-01 -8.80103484e-02 5.38766742e-01 8.14372718e-01
-2.68977404e-01 7.49461830e-01 -8.13464671e-02 -5.71478486e-01
-1.31790268e+00 -1.46271801e+00 2.52524018e-01 6.97390318e-01
6.45966291e-01 -3.05088192e-01 -5.94334602e-01 -3.16259593e-01
5.14616594e-02 5.61517477e-01 -4.08170611e-01 -1.97490193e-02
-9.86859739e-01 -2.79007763e-01 2.33217999e-01 6.14829361e-01
5.50116122e-01 -1.24286675e+00 -1.23123038e+00 2.04408228e-01
-4.49066699e-01 -1.34288597e+00 -4.71969604e-01 -2.57244080e-01
-7.12967396e-01 -8.91115189e-01 -1.03875387e+00 -6.81518674e-01
5.50468504e-01 3.76883268e-01 1.19410217e+00 -9.98761207e-02
-4.04692382e-01 7.14354694e-01 -3.23407739e-01 -3.23500037e-02
-3.43679726e-01 -2.82684565e-01 1.74620286e-01 -4.43076581e-01
-2.62853622e-01 -6.48229361e-01 -9.47267115e-01 5.64932704e-01
-5.69571376e-01 4.72039729e-01 1.78854525e-01 7.16002941e-01
6.33387566e-01 -1.94840252e-01 -7.68710375e-02 -2.95641571e-01
2.91080683e-01 -5.04817963e-01 -2.93770671e-01 -1.00474924e-01
2.88515985e-01 -2.49342814e-01 4.57346112e-01 -6.87902510e-01
-1.09920013e+00 1.76785976e-01 -1.84096083e-01 -6.19551122e-01
-4.75303501e-01 -1.41906276e-01 -4.20697421e-01 1.58850163e-01
6.26734853e-01 4.13858406e-02 -3.92671585e-01 -1.52220815e-01
4.35617328e-01 -9.47890338e-03 6.71120167e-01 -1.00950360e+00
9.37903881e-01 6.58318222e-01 1.02321193e-01 -6.28588676e-01
-2.83259332e-01 -2.98261821e-01 -6.45352662e-01 -4.63414848e-01
1.09412134e+00 -1.04311013e+00 -1.07040739e+00 6.41497672e-01
-1.04087079e+00 -9.48300481e-01 -2.21466616e-01 7.72757173e-01
-1.13301849e+00 2.11170763e-01 -5.65945923e-01 -9.39290285e-01
9.37739760e-02 -1.29124951e+00 1.50819170e+00 2.78879732e-01
-6.20824873e-01 -8.18713903e-01 1.91761672e-01 4.16167527e-01
1.59887716e-01 9.66356039e-01 4.76363987e-01 2.98442781e-01
-9.44213688e-01 1.22564100e-01 4.32776541e-01 -3.18247259e-01
1.29543692e-01 -6.41035065e-02 -5.18949449e-01 -3.06022406e-01
-3.90041590e-01 -2.99466789e-01 3.37307006e-01 6.88716769e-01
1.08426595e+00 -9.55205411e-02 -5.47097087e-01 8.04799616e-01
1.04836011e+00 2.81993836e-01 6.44889235e-01 4.23645645e-01
1.00007296e+00 5.36028028e-01 8.45071018e-01 7.53145099e-01
6.17038667e-01 1.00597465e+00 4.25370544e-01 1.04982732e-02
-4.21437919e-02 -7.21418381e-01 1.91551894e-01 3.71037960e-01
-3.62603277e-01 -5.39393425e-01 -1.01433945e+00 5.15458763e-01
-2.02296925e+00 -1.06880283e+00 -5.99246286e-02 2.06470227e+00
4.62130606e-01 2.40239948e-01 2.25038260e-01 -3.99159975e-02
3.90563428e-01 1.60057724e-01 -6.51972711e-01 -1.38042152e-01
1.15304150e-01 -1.92222334e-02 2.12333426e-01 5.93516529e-01
-9.86909628e-01 9.88671303e-01 5.81902838e+00 4.28993136e-01
-7.27800190e-01 -1.61939293e-01 2.62377232e-01 -3.37666959e-01
-4.65717614e-01 -2.06589207e-01 -7.71145046e-01 3.70867848e-01
5.56926429e-01 1.17878005e-01 2.74036735e-01 9.93155003e-01
5.39265037e-01 -2.28849307e-01 -9.43828940e-01 1.14296722e+00
-2.48947486e-01 -1.20035982e+00 -7.60498270e-02 5.87897487e-02
7.93174505e-01 -2.68092722e-01 5.78622296e-02 3.27983379e-01
6.62434280e-01 -1.01758087e+00 1.03997898e+00 5.95851660e-01
6.05509281e-01 -7.13318110e-01 2.51114130e-01 7.77427018e-01
-1.35910487e+00 4.63847667e-02 -7.02229813e-02 -2.39142731e-01
7.75885463e-01 1.49264291e-01 -5.44429660e-01 5.82740843e-01
9.83334303e-01 5.48542321e-01 -1.15621602e-02 1.05382681e+00
-3.27849448e-01 1.27266096e-02 -3.03033143e-01 -9.74169299e-02
1.19791411e-01 -1.12169139e-01 8.08308959e-01 8.63426089e-01
4.90246356e-01 3.73440117e-01 4.61129546e-01 8.58242035e-01
3.11217070e-01 -2.27119833e-01 -9.07781363e-01 4.44583774e-01
2.77969986e-01 9.03572679e-01 -6.87008679e-01 -2.37835586e-01
1.16475888e-01 1.23091984e+00 9.31001827e-02 5.98182797e-01
-1.27470124e+00 9.94185954e-02 9.04341459e-01 2.72154987e-01
3.35671484e-01 -8.21849525e-01 -1.18332945e-01 -1.30417216e+00
2.36931950e-01 -6.86932921e-01 -1.06576085e-01 -1.02418578e+00
-8.44701350e-01 5.02399027e-01 4.48568463e-01 -1.44606185e+00
-4.88480031e-01 -2.98687011e-01 -4.83052790e-01 6.62136972e-01
-9.84184921e-01 -1.25111270e+00 -7.48022676e-01 8.33649755e-01
7.20147073e-01 2.21425444e-01 7.32600391e-01 1.65550008e-01
-3.04502606e-01 2.52982944e-01 -2.69178182e-01 -9.03187394e-02
4.39675838e-01 -1.03295195e+00 8.29881310e-01 6.14132881e-01
-1.28369197e-01 5.44050634e-01 8.15114439e-01 -8.58920276e-01
-1.52906072e+00 -1.00424743e+00 3.26290190e-01 -8.46312225e-01
2.95881853e-02 -6.86182141e-01 -6.22789741e-01 6.80969298e-01
-2.92649776e-01 2.00150698e-01 4.84210253e-01 -2.61632562e-01
1.86182126e-01 4.14027542e-01 -1.05227661e+00 1.14798963e+00
1.86766028e+00 -2.99291573e-02 -2.26817921e-01 1.46794200e-01
7.72782862e-01 -1.17343700e+00 -6.18429005e-01 5.99606633e-01
7.87488639e-01 -9.31569636e-01 1.43176305e+00 -5.41140735e-01
5.39351285e-01 -4.92544979e-01 -3.14745516e-01 -1.15267301e+00
-3.79029244e-01 -8.19972813e-01 -4.99751478e-01 5.50438225e-01
-7.62052909e-02 -1.29466504e-01 9.95499551e-01 7.36995578e-01
-1.23669915e-01 -9.32317138e-01 -7.90702581e-01 -8.31847131e-01
-2.30796173e-01 -5.31157434e-01 6.49162710e-01 5.74772000e-01
-3.91810358e-01 9.45727527e-02 -7.70028830e-01 3.51697326e-01
6.98811054e-01 4.57849391e-02 1.59700501e+00 -6.91343129e-01
-5.87580323e-01 -1.27454296e-01 -3.39179516e-01 -1.48277700e+00
7.68961534e-02 -3.91633362e-01 2.70063192e-01 -1.80384028e+00
-3.18768695e-02 -5.17080367e-01 4.30724680e-01 1.10314168e-01
-3.95207256e-01 -1.36278883e-01 3.89364719e-01 1.12939134e-01
-6.24958456e-01 9.19851899e-01 1.92169523e+00 1.65501475e-01
-4.90610003e-01 7.62292817e-02 -5.48366718e-02 8.10581386e-01
5.88872015e-01 -1.64624348e-01 -7.10162222e-01 -5.50866902e-01
-1.19738661e-01 5.86794257e-01 8.52487981e-01 -1.32765019e+00
1.57929398e-03 -6.01493955e-01 7.22794771e-01 -6.48422062e-01
7.83445895e-01 -7.12421417e-01 7.24015594e-01 7.25944042e-01
1.70811731e-02 1.16486467e-01 2.82255858e-01 7.15342224e-01
2.46293306e-01 4.82321382e-01 4.96062547e-01 -4.21868831e-01
-8.62374127e-01 5.62610388e-01 -2.41801783e-01 8.24559927e-02
1.30848479e+00 -5.00270367e-01 -1.01174936e-02 -6.43064678e-01
-9.35662210e-01 4.81946707e-01 8.90981436e-01 7.49678671e-01
8.67815495e-01 -1.51478243e+00 -5.29717982e-01 2.48238027e-01
-6.69083968e-02 5.87038934e-01 7.00644374e-01 3.35769922e-01
-8.12709391e-01 1.84472576e-01 -4.73829538e-01 -9.92973983e-01
-1.06913674e+00 4.68900204e-01 2.78517932e-01 -3.68561983e-01
-8.88588011e-01 8.23427737e-01 5.16257167e-01 -7.68858612e-01
2.47201603e-02 -4.17801917e-01 1.68512926e-01 -8.47650468e-01
2.78690279e-01 3.97691548e-01 -5.05963326e-01 -7.26270139e-01
-4.01147842e-01 8.04472625e-01 5.54145753e-01 -3.16235423e-01
1.11518431e+00 -1.89893916e-01 7.23441601e-01 3.11402202e-01
7.97342837e-01 6.04889877e-02 -1.91430700e+00 2.57916033e-01
-6.17142320e-01 -9.67249811e-01 -6.53212190e-01 -4.50814664e-01
-6.72220051e-01 6.51792943e-01 4.05548632e-01 -5.09193242e-01
7.73196399e-01 -8.40014871e-03 9.86550331e-01 2.42122680e-01
1.00028086e+00 -7.97104955e-01 5.60708344e-01 5.26962936e-01
9.87265468e-01 -1.04875588e+00 -1.87976226e-01 -5.72101593e-01
-8.24706674e-01 7.36621857e-01 9.32418942e-01 -2.43676141e-01
3.58619541e-01 2.42792413e-01 -5.56193143e-02 -7.64895231e-02
-5.53291202e-01 -7.65940994e-02 1.78590491e-01 9.27845120e-01
-2.28109062e-02 2.06734568e-01 1.40638813e-01 3.68094087e-01
-4.50800925e-01 -1.14159612e-02 4.09028649e-01 1.31235182e+00
-1.95720121e-01 -1.05312288e+00 -5.56397378e-01 -1.99005634e-01
6.52875155e-02 4.16660696e-01 1.54453397e-01 9.51530397e-01
1.60714865e-01 8.00702035e-01 -1.26075402e-01 -4.41958040e-01
6.75800264e-01 -3.53708774e-01 7.27369547e-01 -5.45210719e-01
-1.89408347e-01 8.57707411e-02 6.28715307e-02 -1.02498984e+00
-4.23078895e-01 -6.14381135e-01 -1.39283657e+00 -4.91245806e-01
1.85945585e-01 -2.93978125e-01 3.44369173e-01 5.81350386e-01
3.97149414e-01 6.69896066e-01 3.82007509e-01 -1.69842982e+00
-1.19234882e-01 -4.41581666e-01 -3.42346430e-01 6.77230179e-01
4.00491327e-01 -1.06788027e+00 -1.42063266e-02 5.11702970e-02] | [7.0058417320251465, -0.6705074906349182] |
69f6a9e0-96c6-424c-a6e9-16a1f7dc655f | fast-and-accurate-intrinsic-symmetry | 1807.10162 | null | http://arxiv.org/abs/1807.10162v4 | http://arxiv.org/pdf/1807.10162v4.pdf | Fast and Accurate Intrinsic Symmetry Detection | In computer vision and graphics, various types of symmetries are extensively
studied since symmetry present in objects is a fundamental cue for
understanding the shape and the structure of objects. In this work, we detect
the intrinsic reflective symmetry in triangle meshes where we have to find the
intrinsically symmetric point for each point of the shape. We establish
correspondences between functions defined on the shapes by extending the
functional map framework and then recover the point-to-point correspondences.
Previous approaches using the functional map for this task find the functional
correspondences matrix by solving a non-linear optimization problem which makes
them slow. In this work, we propose a closed form solution for this matrix
which makes our approach faster. We find the closed-form solution based on our
following results. If the given shape is intrinsically symmetric, then the
shortest length geodesic between two intrinsically symmetric points is also
intrinsically symmetric. If an eigenfunction of the Laplace-Beltrami operator
for the given shape is an even (odd) function, then its restriction on the
shortest length geodesic between two intrinsically symmetric points is also an
even (odd) function. The sign of a low-frequency eigenfunction is the same on
the neighboring points. Our method is invariant to the ordering of the
eigenfunctions and has the least time complexity. We achieve the best
performance on the SCAPE dataset and comparable performance with the
state-of-the-art methods on the TOSCA dataset. | ['Rajendra Nagar', 'Shanmuganathan Raman'] | 2018-07-26 | fast-and-accurate-intrinsic-symmetry-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Rajendra_Nagar_Fast_and_Accurate_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Rajendra_Nagar_Fast_and_Accurate_ECCV_2018_paper.pdf | eccv-2018-9 | ['symmetry-detection'] | ['computer-vision'] | [ 1.51997998e-01 -8.39298666e-02 4.73640710e-02 -1.11286789e-01
-1.50738969e-01 -8.12904894e-01 3.40021253e-01 -6.17297813e-02
7.51874992e-04 8.40623453e-02 -6.12331070e-02 -1.08923770e-01
-2.97432929e-01 -8.96881342e-01 -8.27002704e-01 -6.82201385e-01
-6.23112954e-02 6.67199016e-01 4.31240499e-01 -3.06806952e-01
5.71286440e-01 8.14988554e-01 -1.44014740e+00 2.32017741e-01
6.49143517e-01 9.68802273e-01 -5.58819957e-02 4.87200320e-01
8.31145570e-02 -3.32050562e-01 -3.37142381e-03 -8.77988860e-02
7.45234847e-01 -4.84986961e-01 -9.61271882e-01 1.24192648e-01
6.93302214e-01 2.23693907e-01 5.37067419e-04 1.15042889e+00
3.58779430e-02 5.89238293e-02 9.28777277e-01 -1.42225313e+00
-5.01033127e-01 -1.25805795e-01 -7.54351795e-01 -3.81069541e-01
4.27043885e-01 -6.09589279e-01 1.19818008e+00 -1.18619215e+00
7.79662192e-01 1.13272738e+00 8.46411645e-01 2.78694838e-01
-1.30763578e+00 -4.09490675e-01 -2.55559027e-01 1.31358206e-01
-1.48282170e+00 -1.82167843e-01 8.42532218e-01 -5.46118319e-01
4.13660735e-01 6.63040102e-01 7.49590635e-01 2.03037962e-01
4.56073940e-01 9.99874547e-02 7.07998991e-01 -5.56126297e-01
1.01313591e-01 -2.11058199e-01 -1.07148923e-01 8.91618371e-01
1.52532727e-01 -1.86860368e-01 -2.66690522e-01 -4.23049331e-01
1.17625606e+00 1.34216115e-01 -3.26225251e-01 -8.09131980e-01
-1.50913596e+00 8.39060903e-01 4.00048465e-01 3.29889596e-01
-1.61038116e-01 1.01732962e-01 -6.83707222e-02 2.46143967e-01
-4.34092944e-03 3.95012408e-01 -1.53593779e-01 1.40048116e-02
-5.54533601e-01 4.92039993e-02 1.00041747e+00 6.90278828e-01
1.04710460e+00 -5.52659512e-01 3.41865391e-01 7.50963748e-01
3.13667178e-01 5.42165875e-01 -1.05030186e-01 -1.22533631e+00
1.37968346e-01 7.57687211e-01 1.04641290e-02 -1.59570122e+00
-6.74361348e-01 -2.22586859e-02 -8.97404909e-01 2.11611673e-01
7.48074234e-01 4.23616618e-01 -3.82296383e-01 1.39928567e+00
5.24428666e-01 1.87806129e-01 -3.11220467e-01 8.79702747e-01
5.73883474e-01 5.98907232e-01 -8.12643647e-01 -1.32540613e-01
1.54444671e+00 -5.14866292e-01 -2.74021655e-01 3.74367774e-01
2.55355239e-01 -1.24582863e+00 8.95933628e-01 2.11442232e-01
-8.02927196e-01 -1.92211285e-01 -1.03798616e+00 -2.82117724e-02
7.60382265e-02 2.01545209e-01 2.64657736e-01 3.39166462e-01
-8.12302470e-01 6.95752561e-01 -7.30339408e-01 -2.72541046e-01
-1.22050576e-01 3.31561476e-01 -5.69691002e-01 3.38965982e-01
-5.37250519e-01 5.24615765e-01 -1.29954532e-01 -4.07052152e-02
-3.09385057e-03 -9.93920922e-01 -6.34736538e-01 -2.25683078e-02
4.44417477e-01 -5.12486398e-01 8.21295440e-01 -8.17029059e-01
-1.34194338e+00 1.06694150e+00 -3.95936757e-01 1.67761087e-01
5.01128256e-01 3.96835774e-01 -7.98793212e-02 1.14895597e-01
2.77705312e-01 4.86184746e-01 8.50433409e-01 -1.34166467e+00
-3.39098126e-01 -4.70989674e-01 2.58297473e-02 6.47991104e-03
-6.85297027e-02 -4.84326072e-02 -3.75796437e-01 -4.84634995e-01
1.02401340e+00 -1.30183959e+00 1.58127636e-01 4.08168614e-01
-4.31808949e-01 -4.67296898e-01 1.16720212e+00 -5.14915824e-01
1.03341138e+00 -2.28732228e+00 7.72636235e-02 7.74567306e-01
1.43362418e-01 -2.37081483e-01 5.13391905e-02 6.13329291e-01
-2.51359463e-01 -1.23384641e-02 -3.23342294e-01 1.85231894e-01
-2.00091779e-01 1.08486027e-01 -1.28247201e-01 9.92337584e-01
-4.44690138e-02 5.62232256e-01 -7.40008414e-01 -2.05339521e-01
-2.13763818e-01 5.18373966e-01 -7.55209208e-01 -1.54380962e-01
7.08508343e-02 5.15855074e-01 -3.72554958e-01 2.94500738e-01
1.02276742e+00 -1.47284761e-01 5.48830703e-02 -7.36860335e-01
-5.28175652e-01 6.71278313e-02 -1.79043126e+00 1.29191256e+00
-2.98392326e-01 4.87646550e-01 8.15613717e-02 -1.00365794e+00
1.16832590e+00 1.33188054e-01 8.95215333e-01 -5.40923178e-01
-6.64915293e-02 5.63362539e-01 1.75067723e-01 -1.94019601e-01
1.00801051e-01 -7.26319849e-02 1.77245319e-01 4.13745701e-01
-4.39578384e-01 -1.27931744e-01 1.27098888e-01 -5.96674532e-02
8.65430117e-01 -8.69590230e-03 2.12022543e-01 -8.67216885e-01
8.71128500e-01 -3.05230707e-01 6.17587447e-01 1.72947600e-01
3.89777988e-01 1.03328705e+00 7.46669590e-01 -5.97350538e-01
-1.19522977e+00 -1.09959590e+00 -5.78118265e-01 6.00192428e-01
4.61301327e-01 -5.98445892e-01 -8.04672897e-01 -1.94261923e-01
2.32051060e-01 2.01479346e-01 -6.01867318e-01 1.37826717e-02
-7.72275209e-01 -2.63790756e-01 3.17056803e-03 3.18072975e-01
6.11860633e-01 -5.89391589e-01 -7.98284888e-01 -1.03244036e-01
-1.61432132e-01 -9.89036381e-01 -9.89445269e-01 -4.42198545e-01
-9.09345567e-01 -1.43178487e+00 -6.24499202e-01 -9.16671216e-01
1.09558666e+00 4.90367860e-01 7.51359224e-01 2.51774579e-01
-2.98339099e-01 4.12266612e-01 1.80676952e-01 -1.14096943e-02
-1.57453194e-01 -2.95756727e-01 -3.70395854e-02 6.70094192e-01
-2.63031930e-01 -5.84710002e-01 -6.46568179e-01 1.13774550e+00
-7.50358224e-01 5.34059368e-02 -9.16381925e-02 6.04262829e-01
1.00936842e+00 2.26570293e-02 1.62686557e-01 -3.99387628e-01
1.76037267e-01 -1.11827336e-01 -8.38369846e-01 2.81075299e-01
-1.11545444e-01 3.95050585e-01 4.57395405e-01 -2.35446870e-01
-5.58490574e-01 5.67553341e-01 6.32302165e-02 -3.59707326e-01
3.42498302e-01 2.41857708e-01 -1.07389435e-01 -6.10748887e-01
5.03789067e-01 7.22449794e-02 1.56111836e-01 -6.22076035e-01
-2.92328093e-02 3.48527908e-01 5.08701384e-01 -6.63477004e-01
8.17417324e-01 1.08456814e+00 8.54258001e-01 -1.38009262e+00
-5.16865432e-01 -7.93377280e-01 -7.82876074e-01 -2.76779741e-01
7.12114394e-01 -1.48909003e-01 -1.19174027e+00 3.54018360e-01
-1.44929409e+00 2.88798004e-01 -2.22426310e-01 3.76246214e-01
-7.45942652e-01 6.76140726e-01 8.28165188e-02 -6.09837830e-01
-1.66678280e-01 -1.13402069e+00 1.24752855e+00 -8.10785517e-02
-1.98992327e-01 -1.00726724e+00 2.94224024e-01 -3.63847762e-02
9.37736705e-02 4.13084865e-01 1.09964311e+00 -2.02100426e-01
-6.34344101e-01 -7.89725929e-02 -2.32857704e-01 7.48958886e-02
2.01364338e-01 4.17319089e-01 -2.78056085e-01 -2.50076771e-01
1.68313459e-01 4.39029217e-01 4.73473608e-01 2.74186105e-01
9.64334667e-01 -3.16757679e-01 -2.86928177e-01 7.76952207e-01
1.32896876e+00 3.55774015e-02 5.34254134e-01 3.80913585e-01
8.31268609e-01 8.05811882e-01 3.04182440e-01 2.73416787e-01
2.79575378e-01 1.12357092e+00 3.53490859e-01 1.05214722e-01
4.51705344e-02 -1.42855108e-01 1.99753940e-01 8.04693103e-01
-4.33689684e-01 4.58573908e-01 -9.85853493e-01 3.72989297e-01
-1.91670191e+00 -8.26350689e-01 -8.44958365e-01 2.77042913e+00
4.88288730e-01 -2.40725964e-01 2.03369424e-01 1.86373711e-01
7.51801848e-01 -2.62042642e-01 -2.37681553e-01 -5.14028907e-01
-2.36337949e-02 1.65283069e-01 4.28722113e-01 7.27024674e-01
-8.63475204e-01 2.79232383e-01 5.68590069e+00 5.28427184e-01
-1.20844686e+00 -1.34645179e-01 -1.18885137e-01 3.62340182e-01
-5.08389235e-01 3.14918041e-01 -7.40477026e-01 2.65387177e-01
1.81377828e-01 -3.49551529e-01 3.84925038e-01 5.91010988e-01
1.00287095e-01 -2.40408421e-01 -1.28650093e+00 1.03908038e+00
2.02429563e-01 -1.40478992e+00 -1.32929415e-01 3.39791954e-01
8.19964945e-01 -2.96445489e-01 -8.63028094e-02 -6.09317124e-01
-6.01832271e-01 -8.21047902e-01 5.63400984e-01 4.10775423e-01
7.32191145e-01 -7.80249596e-01 2.45186388e-01 4.06917393e-01
-1.60383403e+00 3.72203737e-01 -4.67180461e-01 7.42877573e-02
1.82634182e-02 5.83101571e-01 -6.94339395e-01 5.45441568e-01
5.80596685e-01 6.96119070e-01 -2.59483188e-01 1.18023181e+00
8.35982636e-02 1.18331112e-01 -7.58174479e-01 2.12066099e-01
8.54085162e-02 -1.01620317e+00 9.24140632e-01 8.08862805e-01
4.92177695e-01 6.13693222e-02 2.08655581e-01 9.72114265e-01
7.67098144e-02 5.04077673e-01 -6.88895762e-01 4.15368855e-01
1.08972356e-01 1.22964489e+00 -1.12779307e+00 2.26103142e-01
-4.39361781e-01 8.93943489e-01 -6.16133325e-02 2.20117241e-01
-4.77668017e-01 -3.42275441e-01 7.53594160e-01 5.85864246e-01
2.45663524e-01 -4.63441133e-01 -5.51838636e-01 -9.60425436e-01
3.83125365e-01 -6.23479187e-01 3.11787218e-01 -4.68027055e-01
-9.98368502e-01 2.01594278e-01 1.07307687e-01 -1.31254005e+00
-8.14484954e-02 -8.31878901e-01 -8.16463053e-01 7.66032338e-01
-8.64322543e-01 -7.34187484e-01 -2.45320097e-01 8.40949416e-01
6.28277808e-02 1.55685693e-01 7.70797074e-01 1.01425566e-01
1.63256794e-01 2.26760447e-01 2.05680117e-01 2.25094527e-01
3.32160801e-01 -8.83319259e-01 2.30240241e-01 5.84032059e-01
3.21489215e-01 7.78818011e-01 5.31897902e-01 -6.80848658e-01
-1.79397535e+00 -5.21986306e-01 9.62352693e-01 -2.29398206e-01
5.94010890e-01 -4.58777785e-01 -9.80784118e-01 6.49885118e-01
-2.08598554e-01 -2.70301402e-02 3.24557245e-01 -6.86972290e-02
-3.89283746e-01 -9.95940119e-02 -1.08072722e+00 6.40963376e-01
1.08485293e+00 -3.77329409e-01 -4.18114901e-01 4.63119566e-01
1.32781610e-01 -4.61840153e-01 -8.04787278e-01 6.00938439e-01
8.00475717e-01 -1.00851536e+00 9.04722750e-01 -4.40018266e-01
1.42585322e-01 -6.90298975e-01 -1.66371778e-01 -1.11881018e+00
-3.72998923e-01 -8.17341924e-01 3.45735878e-01 7.35109508e-01
2.22115189e-01 -9.17501450e-01 8.19108903e-01 2.85432786e-01
1.78913325e-02 -9.59426999e-01 -1.30790567e+00 -1.05183327e+00
-1.37534291e-01 -1.01249762e-01 3.62889498e-01 8.52110386e-01
-7.56092444e-02 2.51202941e-01 3.93613577e-02 3.23152542e-01
7.45383382e-01 7.57694423e-01 6.73194051e-01 -1.65503192e+00
6.37070239e-02 -4.02855307e-01 -8.67435455e-01 -1.06931186e+00
1.40099794e-01 -1.19636202e+00 -2.53374636e-01 -1.25958788e+00
1.27326563e-01 -6.37385368e-01 3.87866735e-01 2.81998277e-01
5.34039021e-01 3.88015270e-01 1.22659057e-01 2.11196706e-01
2.94536762e-02 3.93530279e-01 1.43601167e+00 1.62664548e-01
-2.06891894e-01 2.41955638e-01 -2.22317964e-01 9.56519186e-01
5.58156431e-01 -4.54825282e-01 1.15079209e-02 -7.78542534e-02
6.11899614e-01 -1.19158059e-01 4.44407910e-01 -7.42853045e-01
2.98095286e-01 -1.32235572e-01 -3.08649652e-02 -6.73236489e-01
2.81793803e-01 -1.13843036e+00 6.46855593e-01 6.13447547e-01
2.15803802e-01 3.84286463e-01 -5.25865704e-02 2.31188133e-01
-3.31481695e-02 -5.93749702e-01 1.04234087e+00 9.70893577e-02
-2.32173622e-01 3.37484151e-01 -3.29587143e-03 1.03880808e-01
1.13562071e+00 -5.23712873e-01 -2.42752191e-02 -3.48067373e-01
-4.42949831e-01 -1.14458520e-02 8.95724833e-01 4.46203917e-01
7.94525266e-01 -1.73355424e+00 -8.54086339e-01 6.33928120e-01
2.26517674e-02 -4.66312431e-02 -1.74699128e-01 1.07321811e+00
-7.64883816e-01 3.05227280e-01 -3.92791808e-01 -1.13053596e+00
-1.45290494e+00 8.99518579e-02 6.34174764e-01 3.28067899e-01
-6.39883637e-01 4.15228158e-01 4.89843696e-01 -5.69310248e-01
-1.75450012e-01 -3.25971544e-01 -2.70574577e-02 1.10383995e-01
2.16371596e-01 8.83346260e-01 3.51297349e-01 -1.16519952e+00
-4.95111674e-01 1.53964055e+00 4.23401803e-01 -1.42058179e-01
1.18977582e+00 3.97452831e-01 -7.75905252e-01 3.51611972e-01
1.77380717e+00 5.71725488e-01 -8.48937452e-01 -3.44723114e-03
-2.15710282e-01 -7.43144989e-01 -2.59773999e-01 -3.95122617e-02
-1.24169409e+00 6.21981680e-01 2.09826589e-01 4.46989298e-01
7.67398953e-01 2.13567212e-01 5.71401954e-01 2.94478774e-01
3.71945649e-01 -8.94413054e-01 4.59875204e-02 7.00072646e-01
1.40499938e+00 -7.80065298e-01 6.47656098e-02 -1.15068746e+00
-3.19215208e-01 1.66898429e+00 2.96386272e-01 -4.72182751e-01
9.61521745e-01 2.24808902e-02 -1.69905126e-01 -3.15011382e-01
-7.45204240e-02 1.99901983e-01 9.25528347e-01 2.67806619e-01
3.52445453e-01 2.01542117e-02 -6.35729432e-01 -1.95535019e-01
-4.64873612e-01 -3.50086033e-01 4.48594809e-01 5.89396119e-01
-4.00696605e-01 -1.25720060e+00 -7.58312523e-01 1.73749909e-01
-1.22579671e-01 3.03102553e-01 -4.25152123e-01 4.61445272e-01
-9.86742973e-02 6.45200014e-01 4.53181118e-01 -5.30777872e-02
8.36915612e-01 -1.69269294e-01 7.80070543e-01 -4.38641280e-01
-1.63198978e-01 3.14702503e-02 -3.18682104e-01 -7.67273784e-01
-3.57441396e-01 -9.90046263e-01 -1.76516461e+00 -3.42962563e-01
3.63980830e-02 2.90072896e-02 8.13427150e-01 5.77619076e-01
4.29958105e-01 -2.10775092e-01 9.76856828e-01 -7.53960013e-01
-5.03818512e-01 -2.22916886e-01 -5.45138657e-01 6.34920478e-01
1.52364045e-01 -9.13389683e-01 -3.80245864e-01 -4.37944308e-02] | [8.15269947052002, -2.3795480728149414] |
3071d261-748f-4646-857d-958029fc0154 | decomposed-meta-learning-for-few-shot-named | 2204.05751 | null | https://arxiv.org/abs/2204.05751v2 | https://arxiv.org/pdf/2204.05751v2.pdf | Decomposed Meta-Learning for Few-Shot Named Entity Recognition | Few-shot named entity recognition (NER) systems aim at recognizing novel-class named entities based on only a few labeled examples. In this paper, we present a decomposed meta-learning approach which addresses the problem of few-shot NER by sequentially tackling few-shot span detection and few-shot entity typing using meta-learning. In particular, we take the few-shot span detection as a sequence labeling problem and train the span detector by introducing the model-agnostic meta-learning (MAML) algorithm to find a good model parameter initialization that could fast adapt to new entity classes. For few-shot entity typing, we propose MAML-ProtoNet, i.e., MAML-enhanced prototypical networks to find a good embedding space that can better distinguish text span representations from different entity classes. Extensive experiments on various benchmarks show that our approach achieves superior performance over prior methods. | ['Chin-Yew Lin', 'Tiejun Zhao', 'Qianhui Wu', 'Huiqiang Jiang', 'Tingting Ma'] | 2022-04-12 | null | https://aclanthology.org/2022.findings-acl.124 | https://aclanthology.org/2022.findings-acl.124.pdf | findings-acl-2022-5 | ['few-shot-ner', 'entity-typing'] | ['natural-language-processing', 'natural-language-processing'] | [-1.42951738e-02 -8.44538510e-02 -3.25589895e-01 -3.67355675e-01
-9.81061995e-01 -4.93228853e-01 4.00027156e-01 2.63312548e-01
-6.77524269e-01 5.44247985e-01 1.03432439e-01 -9.20389965e-02
8.35031085e-03 -8.59643757e-01 -4.95576531e-01 -2.65897572e-01
-5.65277562e-02 4.74360645e-01 2.75082409e-01 -2.91689605e-01
1.44474924e-01 5.43319285e-01 -1.27050948e+00 4.90301579e-01
8.96605432e-01 4.31971937e-01 1.56603783e-01 7.99753428e-01
-8.45604300e-01 7.78962016e-01 -6.94915831e-01 -7.10948527e-01
-1.12103634e-01 -4.22590494e-01 -1.09467351e+00 -3.43414783e-01
2.99856871e-01 -1.98769756e-02 -3.32853377e-01 9.29906726e-01
7.07045078e-01 5.71323574e-01 6.76054597e-01 -1.11756194e+00
-5.58103442e-01 1.03438830e+00 -3.79463404e-01 6.28465474e-01
1.29751816e-01 4.02289778e-02 1.07486033e+00 -1.24646997e+00
7.26799548e-01 1.11244905e+00 9.36535120e-01 1.01455832e+00
-1.08076596e+00 -4.83328938e-01 1.52660951e-01 3.48203659e-01
-1.30526137e+00 -4.66893852e-01 6.31930768e-01 -1.70726016e-01
1.24431038e+00 3.17517929e-02 3.95172685e-02 1.25290370e+00
7.54291192e-03 7.96010196e-01 5.75795293e-01 -7.31827855e-01
6.14949703e-01 1.21398866e-01 7.77212620e-01 7.70490885e-01
4.60973531e-01 -6.90978393e-02 -3.28310519e-01 -4.95109111e-01
1.44328877e-01 8.87767524e-02 1.88938081e-01 5.11215515e-02
-1.11341262e+00 6.75764740e-01 1.59661360e-02 6.47898257e-01
-4.92355168e-01 -1.06532320e-01 8.55483234e-01 3.18904638e-01
4.18773383e-01 5.01864433e-01 -8.69445443e-01 -9.48754176e-02
-8.47504854e-01 -2.16679722e-01 1.06207371e+00 1.10331023e+00
7.76788414e-01 2.12988749e-01 -4.30717349e-01 1.09246755e+00
-9.20769423e-02 2.24744201e-01 6.87919796e-01 -3.83470297e-01
5.62764049e-01 6.38766944e-01 5.38082793e-02 -5.36480904e-01
-5.24815381e-01 -4.22589034e-02 -6.99593306e-01 -2.39785329e-01
1.64603829e-01 -5.67097843e-01 -1.07276547e+00 1.66248190e+00
6.14561975e-01 7.19108045e-01 2.74580568e-01 1.67872399e-01
9.11315858e-01 8.67640734e-01 6.24704719e-01 -1.75117075e-01
1.81418586e+00 -8.18084717e-01 -5.58433890e-01 -2.97491729e-01
1.16457963e+00 -2.96171129e-01 8.80551934e-01 -2.76868463e-01
-6.67118371e-01 -5.09355366e-01 -9.37098205e-01 1.79695338e-01
-9.45247114e-01 8.03561509e-03 2.89585620e-01 8.03928673e-01
-2.11291745e-01 7.55425811e-01 -8.47758949e-01 -4.99147147e-01
1.95856944e-01 4.44006436e-02 -3.00713092e-01 -7.26564899e-02
-1.60273719e+00 8.93435359e-01 1.15257907e+00 -1.75244153e-01
-5.60174286e-01 -9.74743426e-01 -1.17770648e+00 2.09369645e-01
5.36162496e-01 -7.02477276e-01 1.25995719e+00 -5.01753032e-01
-1.28002572e+00 7.38645613e-01 -1.83104262e-01 -5.48541725e-01
-1.30019113e-01 -6.38670027e-02 -1.08380628e+00 1.53477132e-01
-8.73091891e-02 1.68445900e-01 5.41192949e-01 -9.20268893e-01
-7.68766999e-01 -1.38016775e-01 -2.20607650e-02 -8.35073292e-02
-6.08990133e-01 2.84533769e-01 -1.29048452e-01 -6.22284114e-01
-4.73819435e-01 -5.69529653e-01 -3.38890672e-01 -6.86963797e-01
-3.94249350e-01 -5.29501021e-01 5.81408501e-01 -4.53311890e-01
1.68280268e+00 -2.02452803e+00 -1.72193557e-01 -1.63901195e-01
2.53289193e-01 6.20712876e-01 -5.16406298e-01 7.28471756e-01
-2.18202636e-01 1.93521217e-01 -6.61547575e-03 -2.77352750e-01
5.79228960e-02 4.40180898e-02 -3.09435248e-01 2.20478058e-01
2.88820326e-01 1.06790805e+00 -1.13840747e+00 -7.32895136e-01
-4.75152768e-02 1.33933976e-01 -8.12227279e-02 3.85285854e-01
-9.72917452e-02 -2.85504878e-01 -2.69487172e-01 7.31445134e-01
4.90085304e-01 -3.71014714e-01 2.17977539e-01 -3.70031059e-01
-1.86072588e-01 1.05687469e-01 -1.32211840e+00 1.58351779e+00
-7.07443297e-01 2.05759898e-01 -5.36514878e-01 -7.47701347e-01
8.29114914e-01 4.17793930e-01 3.07868004e-01 -2.72147059e-01
1.94267213e-01 1.36190921e-01 -2.14756623e-01 -5.38148165e-01
6.60125911e-01 -3.69917154e-01 -4.86107081e-01 4.82515097e-01
6.14237607e-01 9.59847033e-01 4.09957141e-01 1.16971992e-01
1.39269912e+00 -2.57335961e-01 6.93593442e-01 2.68580109e-01
3.87239546e-01 1.90749019e-01 9.73606408e-01 1.10745394e+00
-2.74792284e-01 1.10557057e-01 -1.73329916e-02 -5.69171667e-01
-1.13984859e+00 -8.40391159e-01 3.51681113e-02 1.78206909e+00
9.97240543e-02 -5.71559191e-01 -7.11362422e-01 -1.11018717e+00
-2.40077823e-01 1.10242391e+00 -6.94697440e-01 -3.56222808e-01
-8.10776412e-01 -8.32744479e-01 1.01794362e+00 8.84498119e-01
2.21168786e-01 -1.30390108e+00 -5.99444330e-01 5.86300611e-01
5.78957386e-02 -9.00718331e-01 -7.05563545e-01 5.31760395e-01
-7.57823229e-01 -8.70325148e-01 -8.86497319e-01 -1.11411488e+00
5.22182047e-01 5.06752618e-02 9.69975352e-01 -2.15664312e-01
-5.97445309e-01 2.99527079e-01 -6.37890279e-01 -2.68584400e-01
-5.08495450e-01 5.21296740e-01 1.31106228e-01 -1.69939175e-01
6.52966082e-01 -2.59602517e-01 -1.91070914e-01 1.30458817e-01
-8.50753546e-01 -3.27286452e-01 7.13519812e-01 1.22048020e+00
3.55685055e-01 -7.99602643e-02 7.71030962e-01 -1.46710670e+00
4.87588763e-01 -8.80728722e-01 -2.62926728e-01 8.95393252e-01
-6.49308741e-01 2.26860553e-01 1.04879045e+00 -8.74118567e-01
-1.26355064e+00 -4.44729067e-02 -1.42991856e-01 -5.56544721e-01
-3.80673885e-01 5.21475017e-01 -4.32083726e-01 1.85380697e-01
8.54718387e-01 3.47653717e-01 -7.23072052e-01 -7.43598878e-01
7.75557399e-01 7.13934541e-01 5.05415976e-01 -6.10355139e-01
8.07815731e-01 1.24099672e-01 -4.91435081e-01 -8.93008947e-01
-1.16660213e+00 -8.24012637e-01 -8.70857954e-01 1.87842138e-02
7.29944587e-01 -6.77706718e-01 -4.85452712e-01 4.17258680e-01
-1.24301553e+00 -5.47326393e-02 -5.33992052e-01 3.28518957e-01
-3.63055110e-01 4.18672740e-01 -8.08325827e-01 -8.45928907e-01
-8.12677860e-01 -3.22795957e-01 8.08128178e-01 5.37646711e-01
-9.72121209e-02 -1.26444685e+00 7.07097709e-01 -2.69036859e-01
8.17377716e-02 5.84543943e-02 1.02765369e+00 -1.55332136e+00
1.32724494e-01 -3.05923223e-01 3.49551700e-02 -8.82676989e-02
-2.33143773e-02 -2.11005628e-01 -9.54077721e-01 -2.70624846e-01
-3.29414845e-01 -1.32352069e-01 8.93434465e-01 -1.26851022e-01
8.66374791e-01 -4.59242612e-01 -5.16145527e-01 8.44883502e-01
1.60345042e+00 2.81153917e-01 4.16103601e-01 4.84212071e-01
8.07934642e-01 2.28544146e-01 7.34772921e-01 5.34108877e-01
2.53856301e-01 2.68935800e-01 -2.56398261e-01 1.37177885e-01
4.04568799e-02 -4.50689882e-01 2.29005739e-01 9.47174907e-01
2.87177861e-01 -4.32345271e-01 -1.08687532e+00 6.75266862e-01
-1.80984366e+00 -1.42862236e+00 3.53633314e-01 1.72844291e+00
7.93557346e-01 1.97031423e-01 2.62445509e-01 -2.84707546e-01
1.23918676e+00 3.02114278e-01 -8.46700728e-01 -4.24812824e-01
-3.95649783e-02 2.00422898e-01 5.42380273e-01 -7.42291436e-02
-1.27360833e+00 1.10327911e+00 5.99870443e+00 9.56818640e-01
-7.25030839e-01 3.11772734e-01 2.84130812e-01 2.79348105e-01
-4.80395369e-02 7.15639815e-03 -1.51171839e+00 5.53803027e-01
1.60189700e+00 -6.64476335e-01 -2.28712708e-02 1.15253520e+00
-4.29925501e-01 5.05482495e-01 -9.63047743e-01 8.14751804e-01
1.29172772e-01 -1.50261247e+00 3.11560966e-02 -3.25895935e-01
6.57591403e-01 1.08997472e-01 -5.22726357e-01 9.52174604e-01
5.60103774e-01 -4.39580530e-01 3.19815904e-01 7.41839409e-01
8.74446154e-01 -1.00862062e+00 7.27918506e-01 6.03014171e-01
-1.62237370e+00 -2.48195246e-01 -6.42898560e-01 6.70568407e-01
2.67712712e-01 3.73525977e-01 -9.21441317e-01 6.08936727e-01
3.37878823e-01 4.44330275e-01 -4.19573098e-01 1.29729557e+00
1.60197124e-01 6.49997115e-01 -1.76540539e-02 -2.88954616e-01
3.71234939e-02 4.50890452e-01 5.92889190e-01 1.73023212e+00
1.59608796e-01 4.26678181e-01 2.41573349e-01 5.86652517e-01
-3.93584996e-01 3.60182166e-01 -5.57045698e-01 -3.97449344e-01
1.03901446e+00 1.40320086e+00 -8.11054349e-01 -6.61669135e-01
-5.48037350e-01 1.24652863e+00 7.28685081e-01 9.23638493e-02
-6.79537177e-01 -1.15360487e+00 4.78193343e-01 -4.97128755e-01
7.39223599e-01 3.96518968e-02 -2.01519243e-02 -1.52811658e+00
-4.91014630e-01 -6.39656842e-01 1.07553899e+00 -3.29432845e-01
-1.86701274e+00 6.50355279e-01 -1.35891819e-02 -1.25015378e+00
-2.43158117e-01 -5.63131690e-01 -1.13634264e+00 5.98690748e-01
-1.34546065e+00 -1.11770165e+00 2.72311978e-02 2.31465951e-01
8.41707051e-01 -2.11614251e-01 1.05900288e+00 2.13283226e-01
-9.89470243e-01 1.08791077e+00 4.22383040e-01 7.25517452e-01
7.56325245e-01 -1.38512540e+00 9.22701657e-01 9.06784892e-01
2.29555637e-01 8.50970268e-01 4.24856007e-01 -8.02191734e-01
-1.35497975e+00 -1.44618666e+00 1.09045565e+00 -4.29263711e-01
7.47288585e-01 -1.76225170e-01 -1.28156805e+00 7.31064022e-01
-9.07309353e-02 1.49011374e-01 1.10822976e+00 2.81717718e-01
-7.22524285e-01 9.25808251e-02 -1.07346261e+00 5.26995063e-01
1.05280459e+00 -5.84691107e-01 -1.21395016e+00 3.56555767e-02
9.45903659e-01 -1.75755575e-01 -9.99603927e-01 2.45881692e-01
4.03317422e-01 -3.76489639e-01 9.98741090e-01 -1.41976798e+00
-1.06599152e-01 -9.39474732e-04 5.56100309e-02 -1.38329542e+00
-5.44504523e-01 -4.73594874e-01 -7.24067390e-01 1.73913777e+00
4.81203675e-01 -4.21431571e-01 5.21658719e-01 4.72058415e-01
-4.86703031e-02 -5.40180326e-01 -7.66111135e-01 -1.25375724e+00
8.33483413e-02 -1.86454102e-01 7.91331291e-01 1.24281061e+00
1.51278630e-01 5.81777632e-01 -5.14157712e-01 3.26082259e-01
7.49207795e-01 1.99927330e-01 4.35766280e-01 -1.27188814e+00
-2.37828448e-01 -1.92802340e-01 -3.22628736e-01 -6.05352461e-01
5.38316011e-01 -1.06967795e+00 3.05953592e-01 -1.23606944e+00
4.44191575e-01 -1.51155308e-01 -8.10930789e-01 6.22700989e-01
-5.46398282e-01 -4.37055588e-01 3.03166285e-02 5.14487475e-02
-1.05185258e+00 3.25570643e-01 3.35961431e-01 -2.34194264e-01
-3.15905035e-01 -1.07085995e-01 -5.39756954e-01 6.51205182e-01
7.13368475e-01 -8.94015789e-01 3.62165570e-02 -4.04024236e-02
1.95649564e-01 1.11733861e-01 -8.78030360e-02 -9.06292200e-01
6.56616747e-01 -3.35680276e-01 4.09709752e-01 -7.17383981e-01
-1.11150593e-01 -4.46657568e-01 -5.50539941e-02 3.85884941e-01
-5.73363483e-01 -1.06814113e-02 1.84477910e-01 7.84714401e-01
2.38915995e-01 -8.08823049e-01 8.97705734e-01 -3.13648045e-01
-1.53556514e+00 5.74936688e-01 -2.68679678e-01 5.69246531e-01
9.44638789e-01 -6.90831169e-02 -4.05931532e-01 4.20644283e-01
-8.60517144e-01 2.13713646e-01 2.26206854e-01 3.55685055e-01
5.57375312e-01 -1.28275907e+00 -6.02899015e-01 -1.68123960e-01
6.94442809e-01 -5.50852060e-01 5.67696810e-01 3.73956740e-01
-3.78918462e-02 2.83224694e-02 -2.72915065e-02 -3.43921520e-02
-1.21486676e+00 8.72751176e-01 3.49918395e-01 -5.23234665e-01
-9.01600838e-01 8.46189559e-01 -2.75288939e-01 -6.68991506e-01
2.59902239e-01 2.05600619e-01 -3.94104034e-01 2.56563932e-01
1.07571900e+00 7.69762039e-01 -7.39436671e-02 -3.66434097e-01
-3.75492483e-01 2.32995212e-01 -4.74188298e-01 1.63614154e-01
1.38358307e+00 -1.87145010e-01 2.64432609e-01 1.07804608e+00
1.30268049e+00 -2.36674070e-01 -6.87445402e-01 -6.37214601e-01
8.09071362e-01 -2.37775818e-02 -2.57256538e-01 -6.03804827e-01
-4.28236663e-01 6.82893574e-01 5.64726353e-01 2.18176637e-02
7.26808548e-01 6.03499310e-03 1.36600244e+00 9.88661826e-01
3.71016949e-01 -1.28877378e+00 -4.91501950e-02 1.01011598e+00
-5.56818284e-02 -1.22635293e+00 -3.39121848e-01 1.53558299e-01
-5.51992774e-01 1.41169775e+00 8.16030383e-01 2.05172539e-01
7.22355247e-01 1.76505089e-01 -1.56600535e-01 -2.26867795e-01
-1.08054554e+00 -4.73468184e-01 3.32217306e-01 3.95231247e-01
1.93046451e-01 -6.16674796e-02 -1.79703504e-01 1.10844862e+00
3.20108354e-01 -1.98218115e-02 2.43123263e-01 1.24980068e+00
-9.66014087e-01 -1.13645828e+00 -1.34007126e-01 6.56239569e-01
-3.67284000e-01 -3.09170663e-01 -2.07639560e-01 5.39242208e-01
5.58483303e-02 6.18321180e-01 -3.73865990e-03 -4.95039105e-01
6.40623927e-01 8.85845721e-01 1.83051541e-01 -1.10910153e+00
-7.40391970e-01 -4.14024889e-01 2.24523112e-01 -1.28254727e-01
-1.12263439e-02 -4.42343026e-01 -1.42594624e+00 -1.71817541e-01
-6.54353738e-01 3.78156811e-01 3.80508453e-01 1.06850410e+00
5.41585803e-01 3.71658117e-01 9.23760235e-01 -3.34642202e-01
-8.21543217e-01 -9.46471334e-01 -7.61363566e-01 3.53768349e-01
5.13600931e-02 -4.84948069e-01 -3.36196452e-01 4.22190465e-02] | [9.667068481445312, 9.401098251342773] |
f41080a0-aff0-416b-a573-0f4b03d53c9a | un-likelihood-training-for-interpretable | 2207.00282 | null | https://arxiv.org/abs/2207.00282v2 | https://arxiv.org/pdf/2207.00282v2.pdf | (Un)likelihood Training for Interpretable Embedding | Cross-modal representation learning has become a new normal for bridging the semantic gap between text and visual data. Learning modality agnostic representations in a continuous latent space, however, is often treated as a black-box data-driven training process. It is well-known that the effectiveness of representation learning depends heavily on the quality and scale of training data. For video representation learning, having a complete set of labels that annotate the full spectrum of video content for training is highly difficult if not impossible. These issues, black-box training and dataset bias, make representation learning practically challenging to be deployed for video understanding due to unexplainable and unpredictable results. In this paper, we propose two novel training objectives, likelihood and unlikelihood functions, to unroll semantics behind embeddings while addressing the label sparsity problem in training. The likelihood training aims to interpret semantics of embeddings beyond training labels, while the unlikelihood training leverages prior knowledge for regularization to ensure semantically coherent interpretation. With both training objectives, a new encoder-decoder network, which learns interpretable cross-modal representation, is proposed for ad-hoc video search. Extensive experiments on TRECVid and MSR-VTT datasets show the proposed network outperforms several state-of-the-art retrieval models with a statistically significant performance margin. | ['Zhijian Hou', 'Wing-Kwong Chan', 'Chong-Wah Ngo', 'Jiaxin Wu'] | 2022-07-01 | null | null | null | null | ['ad-hoc-video-search'] | ['computer-vision'] | [ 3.50716561e-01 -1.97189406e-01 -8.17545712e-01 -5.08558095e-01
-9.41528261e-01 -6.11919641e-01 6.80909872e-01 2.75193844e-02
-3.32689047e-01 4.55375075e-01 4.91405785e-01 -2.45897487e-01
-1.55279428e-01 -2.23099977e-01 -7.63043463e-01 -5.92539132e-01
3.56391184e-02 3.48969162e-01 -3.11585933e-01 9.01594833e-02
3.12762409e-02 3.05546098e-03 -1.64367127e+00 6.87310815e-01
5.84940314e-01 1.24706757e+00 5.45187555e-02 3.42994601e-01
-1.70850813e-01 9.11317766e-01 -2.43290842e-01 -2.55629957e-01
1.30895823e-01 -1.83847308e-01 -6.20593965e-01 2.88454473e-01
7.09259927e-01 -2.86707550e-01 -6.09156370e-01 9.63736475e-01
2.32881576e-01 6.17228486e-02 1.09635389e+00 -1.49929452e+00
-1.22447681e+00 4.26673472e-01 -7.71709502e-01 1.49410382e-01
3.43218654e-01 -2.81503946e-02 1.37038827e+00 -1.21033692e+00
5.05410373e-01 1.32563782e+00 4.68851149e-01 6.28535628e-01
-1.20032024e+00 -6.30280018e-01 4.05335188e-01 2.45455250e-01
-1.55083060e+00 -4.05337811e-01 6.94056332e-01 -7.02654123e-01
7.46796429e-01 2.18177214e-01 3.63314062e-01 1.51985049e+00
-1.21756583e-01 1.08392012e+00 8.54014337e-01 -4.41366971e-01
1.15187757e-01 3.51884544e-01 4.99056168e-02 6.27181351e-01
1.96455434e-01 8.80099759e-02 -7.54135728e-01 -6.67174719e-03
6.69872582e-01 4.05702263e-01 -4.14897859e-01 -8.75976920e-01
-1.03669989e+00 1.05676937e+00 5.72335899e-01 1.66873448e-02
-2.06112325e-01 4.83205169e-01 7.05039799e-01 2.92495519e-01
4.58409339e-01 2.52613664e-01 -3.44655633e-01 -4.22188826e-02
-7.90140450e-01 -3.58944088e-02 4.00499135e-01 9.78958189e-01
7.72601128e-01 1.46012902e-01 -2.43112668e-01 8.92367721e-01
6.52948320e-01 4.68791723e-01 6.24264777e-01 -6.71174109e-01
5.22906661e-01 7.77375817e-01 -1.49523646e-01 -1.02487636e+00
-2.01285686e-02 -2.86654651e-01 -7.91640103e-01 -1.14957236e-01
1.71485811e-01 1.80865496e-01 -1.32749534e+00 1.80445433e+00
1.41713709e-01 1.89077467e-01 1.81268960e-01 1.15382719e+00
9.61853623e-01 7.33813405e-01 4.03493881e-01 -2.81541254e-02
1.54546416e+00 -9.58409548e-01 -7.07206845e-01 -4.05158907e-01
6.55268073e-01 -5.99170089e-01 1.27358806e+00 1.46183401e-01
-6.95923209e-01 -4.30504739e-01 -1.08995330e+00 -3.36048126e-01
-3.96744281e-01 2.20874220e-01 7.79350996e-01 4.08847451e-01
-6.93527162e-01 6.49410263e-02 -7.60612071e-01 -4.25063610e-01
6.09437704e-01 2.33138308e-01 -6.40507936e-01 -4.55993801e-01
-1.03372252e+00 6.55507267e-01 5.48334181e-01 -2.01022234e-02
-1.07147801e+00 -6.89879775e-01 -1.03081739e+00 1.28979459e-01
5.50009906e-01 -5.64323127e-01 8.39349389e-01 -1.03063381e+00
-8.58240068e-01 1.00460124e+00 -1.10973075e-01 -2.97470570e-01
1.59166679e-01 -2.94388115e-01 -5.14180362e-01 2.15440810e-01
2.18989804e-01 9.15091872e-01 1.31977904e+00 -1.48739874e+00
-3.93140256e-01 -4.08989102e-01 1.40601724e-01 2.66478390e-01
-7.59533644e-01 -2.99727619e-01 -8.25634360e-01 -9.71261859e-01
1.64675370e-01 -9.29460406e-01 1.20824166e-01 1.91886753e-01
-1.81659892e-01 -2.83550858e-01 1.05702937e+00 -5.63872933e-01
1.27246273e+00 -2.32359767e+00 3.43602985e-01 1.55035019e-01
3.01456094e-01 4.75358125e-03 -3.05746734e-01 2.83945084e-01
-3.11656803e-01 1.27862677e-01 -5.94216436e-02 -2.94629872e-01
6.13002889e-02 4.47456896e-01 -6.34386301e-01 4.83571470e-01
3.10549378e-01 9.42994475e-01 -8.89496565e-01 -5.25868714e-01
1.71102867e-01 7.08364367e-01 -7.43624330e-01 2.97417343e-01
-3.04293096e-01 3.50135893e-01 -4.02335554e-01 9.41315174e-01
3.53460342e-01 -6.74772382e-01 1.42256707e-01 -5.51012635e-01
2.60330349e-01 -1.68854762e-02 -1.02215028e+00 2.11826086e+00
-4.49095517e-01 7.63260543e-01 -2.64751732e-01 -1.32628691e+00
5.51792383e-01 3.74356270e-01 4.04191345e-01 -7.54626870e-01
1.05884135e-01 9.05560777e-02 -4.41319287e-01 -7.45999396e-01
5.29858291e-01 -2.20536083e-01 -3.24606687e-01 4.79783297e-01
2.80362815e-01 3.75645965e-01 2.80906633e-02 3.63406003e-01
7.73946226e-01 1.90494984e-01 1.45053253e-01 3.01917773e-02
3.76014322e-01 -8.59749988e-02 5.82845390e-01 6.24327242e-01
1.95381977e-02 6.96903765e-01 4.31161940e-01 -5.09272397e-01
-8.59123826e-01 -1.04620111e+00 -1.28997743e-01 1.34701025e+00
2.00522095e-01 -6.17501557e-01 -3.01324785e-01 -8.92077208e-01
1.49587154e-01 4.40501362e-01 -9.56806064e-01 -4.70870733e-01
-3.04962099e-01 -3.48347872e-01 3.59635115e-01 5.74767828e-01
9.15220529e-02 -8.64043534e-01 -4.23695117e-01 -2.90083230e-01
-1.83153227e-01 -1.31201386e+00 -6.16505384e-01 2.55652964e-01
-6.30015969e-01 -9.62098598e-01 -6.24766767e-01 -7.52830982e-01
9.05584514e-01 5.61752856e-01 1.06376910e+00 6.93211854e-02
-3.87138605e-01 7.67213762e-01 -5.95920146e-01 -1.56672344e-01
-1.18237890e-01 -1.42705888e-01 8.71009678e-02 1.83644101e-01
6.71134651e-01 -2.83964187e-01 -6.46888673e-01 1.96541026e-01
-1.24834275e+00 1.35514364e-02 6.23410702e-01 1.07329834e+00
7.83641100e-01 -2.30049804e-01 4.24875051e-01 -8.88316989e-01
4.26147163e-01 -7.71895230e-01 -3.04914802e-01 5.21095276e-01
-5.74130714e-01 3.90310913e-01 4.00854379e-01 -6.27262354e-01
-7.18138814e-01 5.94794266e-02 4.69052821e-01 -1.21489429e+00
1.10886000e-01 8.28577042e-01 -1.41104802e-01 1.99525476e-01
5.46884656e-01 2.29110882e-01 1.67519245e-02 -4.79468942e-01
7.03184187e-01 6.68301821e-01 5.14158487e-01 -5.84344923e-01
8.23491931e-01 5.72413206e-01 -2.66800582e-01 -8.04909527e-01
-1.12006247e+00 -7.27786362e-01 -5.31677842e-01 -1.06224589e-01
1.03071988e+00 -1.32882071e+00 -2.89108932e-01 -1.80689201e-01
-9.46772397e-01 2.36935616e-01 -3.72762501e-01 5.08681178e-01
-4.98144507e-01 3.07856888e-01 1.81337874e-02 -5.64584374e-01
-1.37298569e-01 -1.24060476e+00 1.26099241e+00 3.68009955e-02
-2.78741658e-01 -1.14780164e+00 -7.36746341e-02 6.36174560e-01
1.68168575e-01 7.93304145e-02 1.02045858e+00 -6.05152011e-01
-7.16859102e-01 -4.02680397e-01 -4.84961629e-01 4.11817640e-01
2.15051338e-01 -3.71350855e-01 -1.13743925e+00 -5.76742709e-01
-1.81713283e-01 -8.81799340e-01 1.18382108e+00 2.28183478e-01
1.51542151e+00 -3.90411228e-01 -2.70014375e-01 7.85082579e-01
1.41902208e+00 -2.18985841e-01 3.32235277e-01 2.56565452e-01
9.53934729e-01 4.44668740e-01 6.55524015e-01 5.47154129e-01
4.19328660e-01 6.70310974e-01 4.73427445e-01 6.32055551e-02
-1.48243204e-01 -6.00303054e-01 3.35918248e-01 6.80252969e-01
2.44293675e-01 -3.26341420e-01 -8.23371291e-01 5.00722587e-01
-1.99542630e+00 -9.35609877e-01 4.28117245e-01 2.13311625e+00
8.28452170e-01 -8.78011063e-02 -7.83228502e-02 -5.11829183e-02
4.71290201e-01 3.07965875e-01 -6.72137618e-01 -5.69805615e-02
4.33149599e-02 -2.31522337e-01 3.54097337e-01 7.78881386e-02
-1.28585136e+00 9.14458930e-01 5.58267403e+00 8.08481038e-01
-1.30162454e+00 1.73851684e-01 4.02161062e-01 -2.01734290e-01
-6.18909419e-01 4.63836975e-02 -6.24322295e-01 3.84400398e-01
7.50952840e-01 7.06910193e-02 2.96668530e-01 8.40200186e-01
3.56493052e-03 2.12860867e-01 -1.50904953e+00 1.49197257e+00
4.90865082e-01 -1.43440676e+00 5.40931165e-01 -1.67983305e-02
7.27619529e-01 1.74526609e-02 4.87488687e-01 5.29449582e-01
-2.10772199e-03 -1.28018737e+00 7.02030003e-01 3.14357758e-01
1.18934774e+00 -3.94059509e-01 3.93179178e-01 6.14102290e-04
-1.25818181e+00 -2.89660037e-01 -3.50315601e-01 2.76071608e-01
1.48036614e-01 1.12874329e-01 -6.38511360e-01 2.73534894e-01
6.42937958e-01 1.11666536e+00 -6.20504498e-01 7.91049898e-01
-1.55468330e-01 3.90124023e-01 -6.88315630e-02 4.34833050e-01
3.97856951e-01 -6.14594482e-03 3.47812504e-01 1.12870216e+00
7.66277462e-02 -8.83871838e-02 2.72926569e-01 7.31367052e-01
-4.17500496e-01 -1.13518290e-01 -7.76768029e-01 -4.81398761e-01
4.05415267e-01 1.07455385e+00 -4.42552477e-01 -1.81279644e-01
-6.68647766e-01 8.88033211e-01 4.86573428e-01 6.53422952e-01
-9.19259310e-01 1.86759368e-01 7.56311893e-01 1.03095807e-01
4.15638477e-01 -1.38870701e-01 -2.33945519e-01 -1.44837320e+00
9.94295701e-02 -1.01438951e+00 7.55967140e-01 -6.36950970e-01
-1.46825004e+00 3.30137968e-01 1.39140025e-01 -1.61558068e+00
-2.67184377e-01 -7.04805791e-01 -1.56542853e-01 4.98231560e-01
-1.81688964e+00 -1.46959746e+00 -3.47180486e-01 6.19721770e-01
8.26665640e-01 -4.74228173e-01 8.47000301e-01 4.83246714e-01
-5.09736896e-01 8.77262414e-01 2.27482438e-01 1.59922808e-01
7.91859806e-01 -9.01092887e-01 -3.09169471e-01 4.78212476e-01
6.51151001e-01 7.70618260e-01 5.50614655e-01 -3.66448998e-01
-1.77361095e+00 -1.10659683e+00 4.50807601e-01 -4.76616710e-01
7.94742167e-01 -3.65349770e-01 -9.84299600e-01 8.80943835e-01
2.64310353e-02 3.24688494e-01 1.13004327e+00 3.00115198e-01
-9.53933418e-01 2.57045049e-02 -6.11684978e-01 3.42157871e-01
8.21346700e-01 -1.05507076e+00 -6.67496443e-01 4.25255001e-01
7.16951787e-01 -2.37836763e-01 -6.95088923e-01 3.52753997e-01
7.06507981e-01 -4.53798503e-01 1.17730772e+00 -1.00533020e+00
5.75665057e-01 -2.51720428e-01 -5.43272078e-01 -1.07258868e+00
-6.60460889e-02 -1.98374435e-01 -4.23078567e-01 1.13774347e+00
3.14934134e-01 -1.02187566e-01 7.10955381e-01 6.86846793e-01
3.16265002e-02 -7.66821742e-01 -8.74682724e-01 -7.51472116e-01
-1.15693569e-01 -6.02553368e-01 1.51652917e-01 1.11520815e+00
1.81507238e-03 6.83907211e-01 -6.64036870e-01 2.29992375e-01
6.90301299e-01 9.57343206e-02 6.12353802e-01 -1.01390779e+00
-1.86771169e-01 -1.53736174e-01 -5.20283282e-01 -1.21230114e+00
4.42346215e-01 -1.13922596e+00 -2.31837749e-01 -1.43736339e+00
5.11677504e-01 -4.33958977e-01 -5.75011551e-01 5.44118583e-01
-3.33185233e-02 3.89267892e-01 1.86591819e-01 5.47688246e-01
-9.69210804e-01 8.72328579e-01 9.53926623e-01 -4.49538052e-01
6.77557439e-02 -4.27351981e-01 -7.91034997e-01 8.03547561e-01
2.73225933e-01 -5.16333699e-01 -8.74819219e-01 -8.72962713e-01
3.41830045e-01 -5.88228069e-02 4.71309930e-01 -6.27196670e-01
2.36007676e-01 -1.50838673e-01 3.88830245e-01 -5.24506748e-01
6.43684566e-01 -1.15080738e+00 -1.58803061e-01 -7.34796468e-03
-7.13131607e-01 -5.73973618e-02 1.11471172e-02 1.02038336e+00
-5.66260695e-01 -6.97909370e-02 4.55684423e-01 1.45613533e-02
-1.15139270e+00 5.48770368e-01 3.33326869e-02 4.25743639e-01
1.01364255e+00 -3.82043660e-01 -2.85597414e-01 -3.94096047e-01
-6.51476800e-01 4.72949237e-01 3.16186309e-01 9.47022498e-01
8.95808399e-01 -1.61868310e+00 -4.80122536e-01 2.80087411e-01
7.64921486e-01 -7.63914734e-02 3.74702781e-01 5.52172661e-01
-1.23894438e-01 5.58305383e-01 -1.07000679e-01 -8.90481055e-01
-1.22663963e+00 6.80656135e-01 -8.71982276e-02 -8.95600691e-02
-5.13040602e-01 9.75348830e-01 5.53947806e-01 -1.96515322e-01
5.37999272e-01 -1.01292036e-01 -2.79924244e-01 3.76818210e-01
5.55480778e-01 -5.38035706e-02 -2.81302094e-01 -8.19520593e-01
-3.32513958e-01 5.71050525e-01 -3.34886104e-01 1.98980898e-01
1.32717490e+00 -2.76034832e-01 1.19588144e-01 6.11244202e-01
1.66209304e+00 -5.46490312e-01 -1.29193187e+00 -4.80372876e-01
4.87715714e-02 -7.29498386e-01 2.22585797e-01 -6.07423723e-01
-1.00610721e+00 1.03823400e+00 7.20813870e-01 -7.65328929e-02
9.20087159e-01 2.39341348e-01 5.59602439e-01 2.41693661e-01
4.59131449e-02 -1.09963691e+00 5.38843870e-01 3.71927083e-01
9.32572782e-01 -1.78321183e+00 2.39663899e-01 -2.54482239e-01
-9.08352673e-01 9.56008911e-01 7.27657616e-01 1.48510367e-01
6.19667768e-01 -2.32208118e-01 -5.42155616e-02 -4.06062514e-01
-8.90573561e-01 -5.55242877e-03 8.43320966e-01 4.11152303e-01
4.79909092e-01 -1.36541501e-01 1.64846867e-01 3.99779797e-01
2.23787665e-01 -1.02170862e-01 5.33015504e-02 8.97438407e-01
-2.59922177e-01 -9.55519021e-01 -1.65056229e-01 4.95675027e-01
-3.57622623e-01 -1.03972733e-01 -1.22150876e-01 7.59034336e-01
-3.50229777e-02 6.28119111e-01 1.40312955e-01 -2.90888727e-01
1.05178326e-01 1.59652799e-01 2.50043541e-01 -6.84235394e-01
9.02260393e-02 4.32738997e-02 -1.05358854e-01 -4.63435829e-01
-5.78015685e-01 -4.19612080e-01 -1.21412170e+00 6.06831945e-02
-2.10751727e-01 1.27530843e-01 6.14745915e-01 1.04368854e+00
4.69703704e-01 3.37153435e-01 4.56736833e-01 -5.92609167e-01
-7.35689640e-01 -6.84090793e-01 -4.29582387e-01 7.55049944e-01
5.14907360e-01 -1.01323831e+00 -4.11578834e-01 2.97026068e-01] | [10.507333755493164, 1.3317723274230957] |
d2e449af-552d-4ea6-aa72-cd23593bd243 | csclog-a-component-subsequence-correlation | 2307.03359 | null | https://arxiv.org/abs/2307.03359v1 | https://arxiv.org/pdf/2307.03359v1.pdf | CSCLog: A Component Subsequence Correlation-Aware Log Anomaly Detection Method | Anomaly detection based on system logs plays an important role in intelligent operations, which is a challenging task due to the extremely complex log patterns. Existing methods detect anomalies by capturing the sequential dependencies in log sequences, which ignore the interactions of subsequences. To this end, we propose CSCLog, a Component Subsequence Correlation-Aware Log anomaly detection method, which not only captures the sequential dependencies in subsequences, but also models the implicit correlations of subsequences. Specifically, subsequences are extracted from log sequences based on components and the sequential dependencies in subsequences are captured by Long Short-Term Memory Networks (LSTMs). An implicit correlation encoder is introduced to model the implicit correlations of subsequences adaptively. In addition, Graph Convolution Networks (GCNs) are employed to accomplish the information interactions of subsequences. Finally, attention mechanisms are exploited to fuse the embeddings of all subsequences. Extensive experiments on four publicly available log datasets demonstrate the effectiveness of CSCLog, outperforming the best baseline by an average of 7.41% in Macro F1-Measure. | ['Feifei Li', 'Dachao Fu', 'Xu Wang', 'Chaodu Song', 'Ling Chen'] | 2023-07-07 | null | null | null | null | ['anomaly-detection'] | ['methodology'] | [-2.29213014e-02 -4.11239058e-01 -1.78757645e-02 -2.55381376e-01
5.52457012e-02 -2.55381733e-01 3.56356949e-01 7.08720386e-01
-2.39125401e-01 6.43649548e-02 4.20913219e-01 -4.78704274e-01
3.22780833e-02 -4.58639473e-01 -9.39361870e-01 -4.95403647e-01
-6.37420058e-01 -3.89288515e-02 2.97006458e-01 -2.40107253e-01
4.47251290e-01 3.91639918e-01 -1.29296756e+00 2.71726608e-01
7.56298304e-01 1.23394907e+00 -2.53595505e-03 8.14554632e-01
-5.86249948e-01 1.12969708e+00 -5.30743420e-01 3.31284013e-03
-1.27765238e-01 -5.43569207e-01 -3.62141132e-01 -2.69218031e-02
4.30577435e-02 -5.49849391e-01 -8.94203484e-01 1.08987343e+00
5.03633954e-02 1.70681864e-01 2.70416290e-01 -9.98990953e-01
-7.58072257e-01 6.99872673e-01 -6.50490940e-01 8.45826507e-01
2.30029166e-01 3.61355752e-01 1.20705771e+00 -7.63438523e-01
-1.19769201e-01 1.07315707e+00 4.52411503e-01 2.58805342e-02
-8.00053000e-01 -6.10438824e-01 6.84691072e-01 7.48439431e-01
-1.12048733e+00 -1.59341395e-01 8.68208706e-01 -2.23295718e-01
1.50238121e+00 -2.89763547e-02 5.02177119e-01 9.67716575e-01
6.02169394e-01 7.63188899e-01 -1.32521465e-01 -3.10770452e-01
-7.66118541e-02 -4.68859047e-01 6.13333046e-01 7.88189173e-01
2.94090509e-01 -2.30322719e-01 -7.25889862e-01 -3.71558815e-01
4.24553335e-01 8.14141572e-01 -1.67809412e-01 2.34722137e-01
-1.01983380e+00 5.33276081e-01 3.19201052e-01 4.81310874e-01
-5.90113342e-01 3.35440844e-01 8.87115061e-01 6.79754496e-01
5.15316427e-01 2.08415672e-01 -5.26110888e-01 -4.70302939e-01
-2.11955264e-01 -2.71270275e-01 7.11288154e-01 1.13822436e+00
6.15133464e-01 1.71232805e-01 -2.95699954e-01 5.99255621e-01
3.74996424e-01 8.19480270e-02 9.27120566e-01 -3.59135538e-01
5.34239769e-01 1.08045018e+00 -3.52322787e-01 -1.18099439e+00
-1.55251876e-01 -6.04133606e-01 -8.90401661e-01 -5.54920554e-01
-1.41905427e-01 1.71361342e-01 -6.95246816e-01 1.44822407e+00
-1.62433293e-02 8.69442940e-01 -3.49237263e-01 3.32141757e-01
1.86373010e-01 6.37158811e-01 -1.25116378e-01 -3.78250808e-01
1.12033617e+00 -1.05432010e+00 -1.01396930e+00 -2.52600938e-01
9.81999815e-01 -5.65794408e-01 1.21443176e+00 1.09582193e-01
-6.27373755e-01 -5.44060528e-01 -1.18842876e+00 1.92700937e-01
-2.08797634e-01 -2.68677801e-01 2.39246115e-01 -4.11606096e-02
-6.30674779e-01 8.23453784e-01 -1.36941206e+00 -1.00946032e-01
2.87705392e-01 3.01770777e-01 -8.34755716e-04 4.60937321e-02
-1.13817883e+00 8.96007270e-02 7.76291311e-01 3.92843872e-01
-6.39883816e-01 -5.45406938e-01 -1.05120218e+00 3.26620907e-01
5.89509428e-01 3.20930667e-02 1.10962045e+00 -8.54801059e-01
-1.12013996e+00 9.69116911e-02 -4.42679077e-01 -7.22984672e-01
2.39520833e-01 -7.82568634e-01 -8.75738978e-01 -1.60290569e-01
-2.52952427e-01 -6.51709676e-01 6.72795057e-01 -4.77784574e-01
-7.15031326e-01 -2.33740181e-01 -3.43093574e-01 -1.04620971e-01
-1.01531732e+00 -7.39889443e-02 -6.76355183e-01 -7.02594101e-01
-7.43465498e-02 -5.73776126e-01 -2.51385659e-01 -4.82427627e-01
-5.95328987e-01 -4.95916247e-01 1.06605661e+00 -7.95646966e-01
2.22321224e+00 -2.56989217e+00 -2.77456939e-01 3.62478822e-01
2.93920428e-01 3.79261494e-01 -2.59311497e-01 7.90498972e-01
-1.65723357e-02 -2.80713700e-02 -2.91929454e-01 -4.76264536e-01
-4.93607074e-02 4.79008883e-01 -5.58152497e-01 3.78154397e-01
4.36480194e-01 9.36904907e-01 -9.54111397e-01 -3.51048231e-01
2.17416382e-04 4.51465473e-02 -3.21273148e-01 6.18802369e-01
-4.45689112e-01 2.85548419e-01 -4.53088790e-01 4.74739462e-01
4.54701364e-01 -4.18731838e-01 1.35899469e-01 4.79315631e-02
6.00770377e-02 4.30940956e-01 -6.21968687e-01 1.77292275e+00
-3.90342236e-01 4.14881140e-01 -6.18906915e-01 -1.04414165e+00
7.04163432e-01 2.11624548e-01 4.24827009e-01 -8.67888689e-01
-9.51655954e-02 3.03269684e-01 2.66840875e-01 -6.70829415e-01
2.28069663e-01 3.70283961e-01 -1.77282561e-02 8.10693562e-01
8.48019868e-03 8.34056616e-01 3.62316996e-01 3.90606850e-01
1.80655766e+00 -1.65656492e-01 2.70529509e-01 1.73876569e-01
7.61071622e-01 -4.73785520e-01 6.66215122e-01 5.36161304e-01
-5.38390316e-03 4.13107246e-01 6.91251457e-01 -6.36858881e-01
-7.75264502e-01 -9.66709018e-01 5.23350894e-01 9.83941615e-01
1.60516664e-01 -7.54788399e-01 -4.17733580e-01 -1.07933366e+00
1.06690355e-01 7.84737229e-01 -7.15164661e-01 -8.81454766e-01
-1.04375887e+00 -6.06563807e-01 4.24943388e-01 9.07726705e-01
1.98497027e-01 -1.39356828e+00 -6.56881705e-02 6.60517633e-01
-1.00295156e-01 -1.34577227e+00 -1.14431989e+00 2.22783893e-01
-9.90758657e-01 -1.34404671e+00 8.68192315e-02 -5.75427651e-01
5.75485229e-01 2.04377457e-01 9.44076538e-01 5.06513834e-01
-3.00340176e-01 8.94779861e-02 -5.49884737e-01 -3.80224615e-01
-3.20014238e-01 2.51385495e-02 9.82941389e-02 3.17809314e-01
1.02962554e+00 -9.96005416e-01 -6.46984041e-01 2.34505713e-01
-1.19564605e+00 -5.91351867e-01 9.21896577e-01 8.65690410e-01
6.69398010e-01 1.84147388e-01 4.38595682e-01 -1.02591622e+00
8.40323389e-01 -9.25268769e-01 -4.39708322e-01 1.75348714e-01
-7.80289292e-01 2.77445197e-01 1.35793638e+00 -5.28887570e-01
-8.86640608e-01 -4.42044824e-01 4.00230624e-02 -8.30633342e-01
-2.19099298e-01 7.10255623e-01 -1.78025216e-01 5.18356562e-01
2.67725736e-01 6.71605289e-01 -2.82436106e-02 -6.98453069e-01
-1.61195070e-01 6.58847868e-01 5.23379743e-01 -3.21896404e-01
5.94574213e-01 3.94651294e-01 -2.67059177e-01 -7.78296411e-01
-8.78704786e-01 -7.70755529e-01 -6.28459871e-01 2.61078000e-01
5.29893637e-01 -6.71604395e-01 -6.67863667e-01 6.79920256e-01
-1.20470655e+00 -1.36608824e-01 -2.54328132e-01 5.35231173e-01
-1.95345700e-01 8.37296665e-01 -1.08117473e+00 -6.62109792e-01
-4.83029336e-01 -7.14189351e-01 7.78311968e-01 1.30848557e-01
-2.30414361e-01 -1.12335813e+00 6.32098690e-02 -2.08291367e-01
3.19242656e-01 1.44168451e-01 1.32166588e+00 -1.33939266e+00
-5.43151855e-01 -6.88297212e-01 -2.01686248e-01 6.91401482e-01
5.17754614e-01 -1.25079647e-01 -5.69015145e-01 -4.86745328e-01
8.34044665e-02 2.21506983e-01 8.44508290e-01 -1.45230010e-01
1.58589804e+00 -4.18727875e-01 -3.27427983e-01 6.10737801e-01
1.02558362e+00 5.04282653e-01 4.07275736e-01 1.38146117e-01
1.19483757e+00 1.62075907e-01 5.10589123e-01 6.54410303e-01
1.03785679e-01 1.36694998e-01 7.05055535e-01 4.71893191e-01
1.74721465e-01 -5.77016771e-01 7.84674227e-01 1.65902078e+00
2.39356935e-01 -2.91142136e-01 -7.33726084e-01 5.27323186e-01
-2.02824426e+00 -6.58783376e-01 -3.46735001e-01 2.15549159e+00
3.97494733e-01 6.05710685e-01 -1.31925210e-01 1.05886750e-01
7.44418025e-01 3.59477997e-01 -8.86195898e-01 -4.20469761e-01
2.56072879e-01 -1.90636348e-02 3.52696508e-01 1.08868495e-01
-9.82587397e-01 5.62813580e-01 4.81132460e+00 7.75716543e-01
-6.82645857e-01 -6.87203035e-02 2.33677596e-01 -1.62201971e-01
-8.63140747e-02 5.56762181e-02 -5.60889840e-01 1.01399803e+00
1.32117248e+00 -2.74240315e-01 3.28666270e-01 8.38388205e-01
2.62730837e-01 3.50257725e-01 -1.24766374e+00 6.14718616e-01
1.70177490e-01 -8.39108288e-01 2.70052612e-01 2.02307358e-01
4.51378971e-01 1.86278418e-01 -1.77037150e-01 3.08006316e-01
1.74203545e-01 -8.06302607e-01 3.27962190e-01 6.70333087e-01
4.51334029e-01 -1.06601393e+00 1.04859793e+00 4.25690353e-01
-1.52076876e+00 -3.60165060e-01 -3.99962753e-01 -2.10868657e-01
1.43088460e-01 7.93124735e-01 -6.52675867e-01 6.68575525e-01
7.10814297e-01 1.30154335e+00 -4.56251323e-01 9.33309138e-01
-3.02034348e-01 1.11147845e+00 -2.81733274e-01 -3.19510475e-02
3.67060691e-01 -3.10279697e-01 5.57964802e-01 1.34844160e+00
1.94383368e-01 -2.88816869e-01 2.24329397e-01 6.15878820e-01
-2.51712322e-01 3.89027894e-02 -4.82017010e-01 -5.47434807e-01
4.41496581e-01 9.44723248e-01 -3.35391164e-01 -1.84519321e-01
-8.10684621e-01 1.26644325e+00 3.45514655e-01 4.22851980e-01
-7.87255406e-01 -8.86332095e-01 9.47777092e-01 -1.56840030e-02
3.41506392e-01 -4.77713108e-01 2.67371804e-01 -1.18083405e+00
7.17889547e-01 -7.06409097e-01 6.21242166e-01 9.29802060e-02
-1.60964155e+00 5.43095052e-01 -3.64474028e-01 -1.39235008e+00
-2.44727820e-01 -3.97548318e-01 -1.25064874e+00 7.18607783e-01
-1.48463726e+00 -7.71206856e-01 -4.98382539e-01 5.34082353e-01
7.03689933e-01 -1.29361063e-01 6.18051648e-01 5.04775882e-01
-9.42811370e-01 6.65930212e-01 2.69027144e-01 4.92966771e-01
5.14725029e-01 -1.24106991e+00 1.16584480e+00 1.14615095e+00
2.57724106e-01 8.68406475e-01 3.02794367e-01 -8.50726306e-01
-1.42849386e+00 -1.35667896e+00 6.76759779e-01 -1.16227888e-01
1.18802941e+00 -3.84240896e-01 -1.65168488e+00 1.05222166e+00
-6.67619333e-02 4.51389819e-01 8.38701904e-01 -7.58759901e-02
-3.62684578e-01 3.61794271e-02 -2.23252371e-01 4.48475718e-01
1.55189264e+00 -6.97290719e-01 -4.69683349e-01 1.81203932e-01
1.15555274e+00 -1.15805671e-01 -4.55822557e-01 3.14863026e-01
2.93809563e-01 -9.51545596e-01 5.95018268e-01 -9.41353083e-01
3.54786903e-01 -2.93702871e-01 4.29698899e-02 -1.25319517e+00
-1.12086438e-01 -7.25358546e-01 -1.06294727e+00 1.11930978e+00
1.72637358e-01 -9.53729510e-01 6.95346832e-01 -1.57066109e-03
-5.95277131e-01 -8.11531901e-01 -5.73806882e-01 -1.09118259e+00
-4.74933326e-01 -6.57034695e-01 7.68497884e-01 7.44257510e-01
1.07100904e-01 2.53955722e-01 -3.06304991e-01 3.54563475e-01
4.31751013e-01 7.61302337e-02 6.01591051e-01 -1.00904453e+00
-5.82437873e-01 -3.69495600e-01 -5.91512680e-01 -1.37169111e+00
4.12218839e-01 -7.71505117e-01 5.70507310e-02 -1.06992614e+00
-8.45155586e-03 1.27543032e-03 -1.11472237e+00 2.14278728e-01
-5.72261691e-01 -5.13734758e-01 -3.58361542e-01 3.68637472e-01
-9.26694334e-01 9.10228968e-01 8.02886367e-01 -7.15845376e-02
-3.70930344e-01 8.83737206e-02 -3.97562027e-01 8.64810705e-01
8.92963648e-01 -4.55676883e-01 -4.85782593e-01 -5.08644342e-01
2.72318050e-02 -2.87415177e-01 5.68952877e-03 -1.01956534e+00
5.38731575e-01 2.45637055e-02 1.61123246e-01 -5.89622974e-01
-4.22889180e-03 -7.73381948e-01 -2.93742329e-01 6.49252713e-01
-2.09595621e-01 6.78120911e-01 1.69266731e-01 1.37237263e+00
-6.07042849e-01 2.76312437e-02 6.78025633e-02 8.99401456e-02
-7.58766532e-01 7.76383877e-01 -3.03260863e-01 1.59039095e-01
9.31646883e-01 1.89775694e-02 -4.15148363e-02 -2.75154173e-01
-4.01279032e-01 5.42253315e-01 1.76432416e-01 6.71482205e-01
7.76934087e-01 -1.43145216e+00 -4.57640380e-01 6.51783645e-01
5.39278984e-01 2.02968881e-01 3.55516493e-01 9.33935463e-01
-5.51449120e-01 2.31181458e-01 2.46177673e-01 -3.42065066e-01
-1.18740678e+00 7.96870530e-01 1.79985836e-01 -5.10341585e-01
-8.53069901e-01 7.72778928e-01 1.10359676e-01 8.96020755e-02
3.45056862e-01 -3.19699883e-01 -1.04586445e-01 -2.43088990e-01
7.93583274e-01 3.75266969e-01 1.83576912e-01 -3.46761316e-01
-3.28940362e-01 2.94870138e-01 -7.91144609e-01 5.60412109e-01
1.22324848e+00 -1.46195754e-01 -3.69975567e-01 7.59531081e-01
1.50225663e+00 -1.54199777e-02 -1.23993242e+00 -7.71752119e-01
6.75279558e-01 -4.63914007e-01 -3.87614369e-01 -9.15017053e-02
-1.13959384e+00 1.02848268e+00 1.59464598e-01 2.50826210e-01
1.12183201e+00 -2.20658213e-01 1.43450284e+00 5.15603602e-01
5.01618758e-02 -8.92133772e-01 6.45119846e-01 8.33601117e-01
4.76185828e-01 -9.79296625e-01 -4.09917086e-01 -1.52524799e-01
-3.13165337e-01 1.26865041e+00 7.94608474e-01 -3.59541565e-01
7.83743382e-01 2.12436274e-01 -1.55239508e-01 -2.45920151e-01
-8.76104176e-01 7.23669603e-02 1.78481013e-01 1.50193989e-01
4.64250326e-01 -2.75345206e-01 -1.41468227e-01 6.54290080e-01
3.02241564e-01 -2.61300355e-01 3.08301747e-01 1.18952811e+00
-3.09285879e-01 -1.04743171e+00 1.86353743e-01 8.63112628e-01
-6.51228428e-01 -3.24001312e-01 -3.32385331e-01 3.46492201e-01
-1.56037971e-01 8.57012093e-01 5.24209082e-01 -7.88327098e-01
6.16746068e-01 1.97632670e-01 -3.04830372e-01 -6.99727774e-01
-5.48606396e-01 -1.07634336e-01 -3.60923886e-01 -9.16565776e-01
4.24404651e-01 -5.54808974e-01 -1.53003097e+00 -2.44886145e-01
-3.66486430e-01 3.61228079e-01 2.29245767e-01 1.06018162e+00
7.43953645e-01 1.11259413e+00 8.79409909e-01 -2.37194821e-01
-6.48616076e-01 -1.19752514e+00 -6.82577252e-01 7.01461673e-01
6.24676347e-01 -3.30723703e-01 -7.04064250e-01 -4.78951335e-02] | [7.346531867980957, 2.6302378177642822] |
e24167cb-f8f3-49ee-9548-2c38b06a8448 | rf-based-fall-monitoring-using-convolutional | null | null | https://doi.org/10.1145/3264947 | http://people.csail.mit.edu/yonglong/yonglong/rffall.pdf | RF-Based Fall Monitoring Using Convolutional Neural Networks | Falls are the top reason for fatal and non-fatal injuries among seniors. Existing solutions are based on wearable fall-alert sensors, but medical research has shown that they are ineffective, mostly because seniors do not wear them. These revelations have led to new passive sensors that infer falls by analyzing Radio Frequency (RF) signals in homes. Seniors can go about their lives as usual without the need to wear any device. While passive monitoring has made major advances, current approaches still cannot deal with the complexities of real-world scenarios. They typically train and test their classifiers on the same people in the same environments, and cannot generalize to new people or new environments. Further, they cannot separate motions from different people and can easily miss a fall in the presence of other motions.
To overcome these limitations, we introduce Aryokee, an RF-based fall detection system that uses convolutional neural networks governed by a state machine. Aryokee works with new people and environments unseen in the training set. It also separates different sources of motion to increase robustness. Results from testing Aryokee with over 140 people performing 40 types of activities in 57 different environments show a recall of 94% and a precision of 92% in detecting falls. | ['Dina Katabi', 'Chen-Yu Hsu', 'Guang-He Lee', 'Yonglong Tian', 'Hao He'] | 2018-09-01 | null | null | null | proceedings-of-the-acm-on-interactive-mobile | ['rf-based-pose-estimation'] | ['computer-vision'] | [ 3.80331054e-02 -3.09270173e-01 -6.32843226e-02 -3.58702302e-01
-5.45344472e-01 -1.67539805e-01 -2.45566994e-01 -1.17684109e-02
-8.31705093e-01 1.01128292e+00 5.47908604e-01 -1.47447854e-01
-3.11169773e-02 -9.97981071e-01 -4.23561573e-01 -2.66015917e-01
-2.18495235e-01 2.00092316e-01 8.61674726e-01 -4.87348527e-01
-3.68478268e-01 3.33666503e-01 -1.70906687e+00 5.35122693e-01
6.69882238e-01 5.29608011e-01 -1.16546877e-01 6.89980090e-01
5.04946887e-01 4.30714905e-01 -9.48200583e-01 1.97676599e-01
1.24880418e-01 -3.06796998e-01 -4.06675220e-01 -5.96289039e-01
4.19460803e-01 -8.85905147e-01 -8.11182261e-01 2.87171453e-01
9.41099524e-01 2.38657385e-01 3.37438166e-01 -1.16591573e+00
-2.52513289e-01 1.57177851e-01 -2.51672417e-01 7.12133110e-01
9.79466796e-01 1.04120083e-01 2.20326379e-01 -4.20071661e-01
-6.31595328e-02 1.07889771e+00 1.31577444e+00 9.11217153e-01
-1.03693020e+00 -8.03766489e-01 1.61241636e-01 2.41328895e-01
-1.34661305e+00 -4.88875210e-01 2.06872687e-01 -3.58964205e-01
1.20895791e+00 3.91015023e-01 1.02482390e+00 1.44054484e+00
4.66813117e-01 4.88124073e-01 4.86960441e-01 -2.11675882e-01
4.39982802e-01 -3.68438631e-01 5.40877819e-01 3.07986259e-01
1.11940825e+00 4.37928438e-02 -9.66379106e-01 -2.71732599e-01
1.67846963e-01 6.79813087e-01 -6.20361209e-01 1.03876829e-01
-9.04859245e-01 4.07869518e-01 6.82720721e-01 5.22528768e-01
-2.15533629e-01 1.66384459e-01 2.07516477e-01 2.03995571e-01
-6.02354743e-02 -2.21348166e-01 -1.50462508e-01 -3.03340673e-01
-8.82329583e-01 5.53390026e-01 7.33357310e-01 6.76314771e-01
3.05894464e-01 -2.01928884e-01 -4.87205610e-02 2.17142791e-01
5.45094788e-01 9.33441877e-01 7.97078073e-01 -4.30736125e-01
7.44560361e-01 7.65877783e-01 4.17153746e-01 -8.97008121e-01
-1.01043689e+00 -2.24551484e-01 -6.26359701e-01 3.32461208e-01
5.30804336e-01 -4.50299740e-01 -9.25851285e-01 1.45149267e+00
1.33178040e-01 -1.68392792e-01 -2.62474090e-01 9.38550293e-01
6.85494244e-01 -9.69835892e-02 1.20615512e-01 1.49833441e-01
1.17298555e+00 -3.04736435e-01 -6.36981547e-01 -1.00234532e+00
5.49110115e-01 8.72751884e-03 9.53482628e-01 4.77521002e-01
-4.63780582e-01 -6.16363645e-01 -1.56397319e+00 3.46982837e-01
-3.00259054e-01 -2.51316786e-01 3.77833068e-01 1.24886298e+00
-8.91929924e-01 6.23912096e-01 -1.42577958e+00 -6.68701112e-01
5.26877046e-01 6.13567531e-01 -3.11531425e-01 -1.39624327e-01
-1.37993872e+00 1.00570130e+00 -4.15781140e-02 3.54244202e-01
-2.79821277e-01 -4.03546751e-01 -6.97540045e-01 -2.79229373e-01
-1.24628268e-01 -6.73302233e-01 1.13718510e+00 -4.43596065e-01
-8.81509662e-01 2.74054021e-01 -2.56150097e-01 -6.06673896e-01
7.68484890e-01 -1.06162763e+00 -9.47418213e-01 -6.09722435e-02
3.96151185e-01 -1.33865759e-01 3.51484895e-01 -5.96076846e-01
-9.16028798e-01 -8.25700700e-01 -3.33719820e-01 -1.09295376e-01
-5.47344029e-01 -1.57601163e-01 3.81372690e-01 -1.95708260e-01
3.80119622e-01 -8.74411881e-01 -1.07585348e-01 -2.54113544e-02
-2.11164311e-01 1.98137805e-01 8.55509102e-01 -6.35990083e-01
1.50516450e+00 -1.86783385e+00 -6.27736807e-01 1.53605729e-01
2.73654282e-01 3.46967787e-01 6.49342358e-01 2.57340699e-01
2.02465326e-01 -2.24005014e-01 -9.34995264e-02 4.18980233e-02
-2.78550655e-01 2.12495640e-01 8.05807784e-02 5.25281429e-01
-2.62981713e-01 5.15409768e-01 -9.83639419e-01 -8.42067599e-02
2.89196879e-01 6.40794337e-01 -2.96871036e-01 -2.10274041e-01
6.72397733e-01 4.46240038e-01 -4.71504450e-01 5.62248886e-01
3.32810670e-01 2.32277006e-01 -7.97246769e-02 1.88911743e-02
9.48236361e-02 2.59501278e-01 -1.43246567e+00 1.29186618e+00
1.99434236e-02 6.21264994e-01 -4.17579502e-01 -8.91169012e-01
6.69672012e-01 3.54494274e-01 5.19633710e-01 -7.13591099e-01
1.59618109e-01 4.11654770e-01 -8.66459161e-02 -9.87290263e-01
9.02048200e-02 -2.75719494e-01 -2.32533589e-01 3.71418715e-01
-5.49103618e-01 7.02251732e-01 1.34447232e-01 -1.92068636e-01
2.03985524e+00 1.58072352e-01 1.75013334e-01 1.16021158e-02
-4.52480987e-02 -9.47783515e-02 6.12715900e-01 1.05196118e+00
-5.96777081e-01 7.08348751e-01 -7.30392575e-01 -8.94073188e-01
-2.23822638e-01 -1.57815576e+00 5.74705489e-02 9.87379193e-01
1.52403623e-01 -3.37642848e-01 -6.60110533e-01 -5.00508130e-01
3.38372052e-01 4.03017402e-01 -4.63988513e-01 -7.48156011e-01
-8.80036771e-01 -9.42800224e-01 9.10501361e-01 1.19815671e+00
8.46480370e-01 -9.04237449e-01 -1.68816090e+00 5.03887773e-01
-5.95567226e-01 -7.77882099e-01 -1.59655958e-01 3.48158747e-01
-7.98788428e-01 -1.32088184e+00 -8.13331962e-01 -5.31675935e-01
2.87869751e-01 4.62596476e-01 6.81030989e-01 1.66147128e-01
-5.03998756e-01 5.05549848e-01 -2.84021676e-01 -7.40627050e-01
1.69809729e-01 8.28894079e-02 5.68291008e-01 -2.51274437e-01
1.08326781e+00 -8.28793406e-01 -8.32470119e-01 4.85599190e-01
-2.55163997e-01 -6.14219487e-01 2.84548581e-01 3.43875885e-01
7.24254968e-03 8.19009095e-02 6.15485072e-01 -1.36801824e-01
5.23662746e-01 -4.77613807e-01 3.23893428e-01 1.21265939e-02
-4.16276395e-01 -1.07542545e-01 6.83121979e-02 -6.21047139e-01
-7.74220943e-01 2.78493404e-01 -1.57211155e-01 4.83239770e-01
-4.31019634e-01 -6.23205025e-03 -2.82525480e-01 3.07146132e-01
1.37774968e+00 -1.14104636e-01 -2.30659544e-01 -4.49241906e-01
-4.30019140e-01 1.09435606e+00 8.27609122e-01 -2.31095672e-01
5.55217624e-01 7.74171889e-01 -2.55895674e-01 -9.28211331e-01
-8.17579091e-01 -7.15203643e-01 -8.76988411e-01 -5.86948633e-01
8.28600228e-01 -9.08742309e-01 -9.17388797e-01 6.72856450e-01
-7.03563333e-01 -4.21839386e-01 -1.15840413e-01 8.07085156e-01
-2.87561089e-01 1.49064913e-01 -2.98458308e-01 -1.05577826e+00
-3.39496106e-01 -5.38035512e-01 8.17885816e-01 5.58865607e-01
-1.03995824e+00 -3.59909654e-01 2.45694607e-01 3.86964470e-01
5.65339744e-01 7.22533762e-01 5.24615124e-02 -1.58496469e-01
1.69902191e-01 -7.36066401e-01 3.18170875e-01 -6.74580038e-02
6.81740046e-01 -7.39437759e-01 -1.00073719e+00 -4.18356627e-01
-3.73792313e-02 6.51736557e-02 8.51538062e-01 2.97907352e-01
2.77112305e-01 2.98044831e-02 -9.93291378e-01 2.23595142e-01
9.58252549e-01 3.40097874e-01 8.83436382e-01 8.60443711e-01
6.09841168e-01 3.16922784e-01 1.99344784e-01 2.64278919e-01
4.29416090e-01 7.03047991e-01 2.58915365e-01 1.07523829e-01
-1.51283786e-01 -1.06700882e-01 6.51468754e-01 -1.52060211e-01
-5.12028217e-01 -3.02477032e-01 -1.01476610e+00 4.33340311e-01
-1.95248127e+00 -1.18430519e+00 -4.83355343e-01 2.32509160e+00
4.08763468e-01 4.54876304e-01 5.05856276e-01 9.97881472e-01
6.08120978e-01 -5.00657737e-01 -8.30643415e-01 8.53191838e-02
8.89729336e-02 3.39167655e-01 7.43784666e-01 2.09581539e-01
-1.14955032e+00 2.14257807e-01 6.43584013e+00 -4.34392750e-01
-8.86388242e-01 7.07248449e-02 -4.68683392e-02 -5.38774073e-01
4.00271535e-01 -5.10066450e-01 -8.44868898e-01 6.25655890e-01
1.18682027e+00 3.45063001e-01 -2.28373244e-01 7.20687807e-01
3.99237841e-01 -5.10623991e-01 -1.19002926e+00 9.79459107e-01
1.00827038e-01 -4.43794191e-01 -5.37412703e-01 -9.72417593e-02
-5.60127646e-02 6.31253794e-02 -5.40045261e-01 2.78559119e-01
5.33402525e-02 -1.02791250e+00 9.10360515e-01 8.24312866e-01
5.14341891e-01 -6.02183938e-01 8.91328514e-01 3.42388064e-01
-1.37571192e+00 -4.86940533e-01 1.39371492e-02 -7.55392373e-01
4.79616135e-01 3.73367697e-01 -4.43695545e-01 1.22619316e-01
1.63094211e+00 3.61131638e-01 -6.09657645e-01 1.13555741e+00
-2.23801211e-01 6.92922115e-01 -8.82999539e-01 -1.47177905e-01
-4.74473834e-01 5.09293556e-01 2.97399163e-01 9.90684450e-01
5.60138226e-01 1.69291034e-01 1.48701534e-01 1.35577008e-01
5.95832765e-01 -4.30538744e-01 -5.80347776e-01 7.27767527e-01
2.67564386e-01 3.55766475e-01 -6.72481596e-01 4.03277501e-02
-5.42671382e-01 1.13556302e+00 -2.03605041e-01 4.00437176e-01
-5.84761739e-01 -4.90976959e-01 7.18786716e-01 1.04846382e+00
-1.82327852e-01 -3.82538021e-01 -3.33721310e-01 -8.25557590e-01
5.29182017e-01 -5.26926160e-01 5.44151902e-01 -5.45814097e-01
-8.37926149e-01 2.48799220e-01 6.93799257e-02 -1.41189790e+00
-1.58899173e-01 -4.89765674e-01 -4.57549483e-01 5.78197360e-01
-7.71717370e-01 -6.08217537e-01 -9.17203248e-01 7.62089550e-01
2.76006758e-01 1.50677755e-01 9.82229829e-01 6.59104466e-01
-4.10293281e-01 8.51701796e-01 -6.92118704e-02 3.43727291e-01
9.23386157e-01 -8.43857169e-01 5.05608678e-01 7.98409164e-01
-3.82839948e-01 6.48149312e-01 8.64116967e-01 -1.03211713e+00
-1.02367365e+00 -8.53214145e-01 9.92512703e-01 -7.89298356e-01
1.46468416e-01 -3.16649795e-01 -8.15534592e-01 6.46090865e-01
-7.15497375e-01 1.48861241e-02 7.29577780e-01 -4.08044225e-03
-2.06200555e-02 -2.82143861e-01 -1.33322787e+00 5.88072777e-01
1.77259207e+00 -1.47696152e-01 -9.10353541e-01 3.73226255e-02
2.39161015e-01 -2.17728227e-01 -3.52032810e-01 4.14346069e-01
1.22466147e+00 -1.01242208e+00 1.15182149e+00 -4.19459164e-01
-3.95090044e-01 -3.97926003e-01 -2.33138427e-01 -8.52620363e-01
-3.47560495e-01 -1.66356891e-01 -2.23602459e-01 7.68420756e-01
1.18298069e-01 -8.48342359e-01 9.12561536e-01 9.30689871e-01
-1.02638490e-01 -2.06958532e-01 -1.12355089e+00 -9.54651475e-01
-3.44979018e-01 -7.92867422e-01 6.02907360e-01 2.69773632e-01
2.47037902e-01 2.76885837e-01 -3.47264528e-01 1.63944706e-01
4.80221629e-01 -7.29538083e-01 6.43669784e-01 -1.75517118e+00
-1.41490012e-01 -3.70506607e-02 -1.04270196e+00 -5.90938151e-01
-5.13370633e-01 -2.38111749e-01 4.84628320e-01 -1.88525486e+00
-6.97091147e-02 -1.41880661e-01 -2.32731760e-01 9.14093196e-01
-1.69415891e-01 4.37281251e-01 -4.45001513e-01 8.27286318e-02
-2.28490725e-01 1.12437584e-01 4.60627437e-01 -3.16743493e-01
-5.51812232e-01 6.03672802e-01 -6.77597523e-01 8.60926926e-01
9.38191175e-01 -5.22526801e-01 -4.13433820e-01 -3.39237779e-01
3.76819521e-01 -4.65134025e-01 5.97788453e-01 -2.04677534e+00
7.58098736e-02 1.21402316e-01 1.02589512e+00 -4.82202142e-01
2.02390894e-01 -7.33178198e-01 3.01460862e-01 1.09905612e+00
2.49540493e-01 -2.69849529e-03 3.44316512e-01 5.94531596e-01
4.19553876e-01 3.02878488e-02 5.00134945e-01 -1.85792614e-02
-4.99718934e-01 -3.79718319e-02 -8.82246435e-01 4.04368453e-02
7.99315929e-01 -8.01566064e-01 -2.47529939e-01 -3.74874502e-01
-9.38684046e-01 1.19667329e-01 1.67993352e-01 7.15529025e-01
5.90663970e-01 -1.27495563e+00 -3.63557667e-01 3.35827887e-01
1.42341763e-01 2.57585756e-02 1.05066538e-01 6.25513494e-01
-3.84158015e-01 1.78833142e-01 -3.01194936e-01 -5.04204869e-01
-1.34689009e+00 1.16079263e-01 6.98041439e-01 2.85445929e-01
-1.34324360e+00 5.14863431e-01 -4.04993147e-01 4.46074493e-02
4.62634921e-01 -4.81574923e-01 -2.36104444e-01 -6.03429712e-02
1.17289996e+00 7.11795151e-01 4.64470059e-01 -6.49519682e-01
-9.95292246e-01 4.34918404e-01 1.88181788e-01 -1.24585100e-01
1.23172045e+00 -1.09923080e-01 7.25043476e-01 6.49514258e-01
5.46279311e-01 -2.29823351e-01 -1.10010576e+00 4.31518257e-02
1.80111483e-01 -2.00166568e-01 -2.71911114e-01 -6.61091685e-01
-6.61048234e-01 5.90359747e-01 1.54915059e+00 -2.01034267e-02
9.37674701e-01 -1.18557654e-01 1.17602265e+00 7.05956161e-01
8.43022346e-01 -1.09810078e+00 -5.45834331e-03 1.79232568e-01
4.83647346e-01 -9.74748671e-01 2.82506440e-02 8.31334293e-03
2.77147237e-02 1.07108140e+00 5.32334208e-01 -1.94226101e-01
8.49916339e-01 4.73687440e-01 1.89702064e-01 -5.34302220e-02
1.67447120e-01 -2.78594613e-01 -4.96320613e-02 1.28635037e+00
5.01847804e-01 2.04515800e-01 -3.30171317e-01 1.13807416e+00
-5.84554791e-01 4.65928286e-01 3.21095645e-01 1.72204685e+00
-9.27511215e-01 -7.55616665e-01 -7.62522101e-01 6.85487747e-01
-3.81877452e-01 6.39083743e-01 -2.69318104e-01 7.67756879e-01
4.37017143e-01 1.51746356e+00 -1.58177793e-01 -7.76829600e-01
9.53974664e-01 2.52338290e-01 5.19572735e-01 -5.74201524e-01
-4.35074419e-01 -5.11267602e-01 1.22653462e-01 -6.78559840e-01
-4.07755136e-01 -9.08057332e-01 -1.68813121e+00 -1.80437773e-01
1.48470688e-03 -1.05243273e-01 1.86709568e-01 1.24299395e+00
1.94346197e-02 8.81280959e-01 -4.09429334e-02 -8.65746319e-01
-1.80446252e-01 -1.05779541e+00 -4.47654039e-01 3.27015877e-01
6.27047777e-01 -9.04390514e-01 -3.20998847e-01 3.35968584e-02] | [7.197625160217285, 0.5851206183433533] |
fcc7c05f-1b16-4836-99da-5ae63cc00f4a | cascaded-deep-monocular-3d-human-pose-1 | 2006.07778 | null | https://arxiv.org/abs/2006.07778v3 | https://arxiv.org/pdf/2006.07778v3.pdf | Cascaded deep monocular 3D human pose estimation with evolutionary training data | End-to-end deep representation learning has achieved remarkable accuracy for monocular 3D human pose estimation, yet these models may fail for unseen poses with limited and fixed training data. This paper proposes a novel data augmentation method that: (1) is scalable for synthesizing massive amount of training data (over 8 million valid 3D human poses with corresponding 2D projections) for training 2D-to-3D networks, (2) can effectively reduce dataset bias. Our method evolves a limited dataset to synthesize unseen 3D human skeletons based on a hierarchical human representation and heuristics inspired by prior knowledge. Extensive experiments show that our approach not only achieves state-of-the-art accuracy on the largest public benchmark, but also generalizes significantly better to unseen and rare poses. Code, pre-trained models and tools are available at this HTTPS URL. | ['Kwang-Ting Cheng', 'Chi-Keung Tang', 'Yu-Wing Tai', 'Lei Ke', 'Shichao Li', 'Kevin Pratama'] | 2020-06-14 | cascaded-deep-monocular-3d-human-pose | http://openaccess.thecvf.com/content_CVPR_2020/html/Li_Cascaded_Deep_Monocular_3D_Human_Pose_Estimation_With_Evolutionary_Training_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Li_Cascaded_Deep_Monocular_3D_Human_Pose_Estimation_With_Evolutionary_Training_CVPR_2020_paper.pdf | cvpr-2020-6 | ['monocular-3d-human-pose-estimation', 'weakly-supervised-3d-human-pose-estimation'] | ['computer-vision', 'computer-vision'] | [ 4.99725789e-02 1.48735285e-01 -2.88600236e-01 -1.97695211e-01
-8.02347720e-01 -3.08539987e-01 3.58874232e-01 -5.93068659e-01
-3.52051765e-01 7.61594951e-01 6.44427657e-01 6.32199571e-02
3.69238853e-01 -3.98153514e-01 -1.00154030e+00 -1.15067407e-01
-1.30196661e-01 1.08095348e+00 1.04733348e-01 -1.56712070e-01
-3.51311982e-01 5.96498132e-01 -1.43631828e+00 6.79560900e-02
1.80495352e-01 7.84335911e-01 -3.10045838e-01 6.51955187e-01
5.08970976e-01 4.40613180e-01 -4.73381490e-01 -4.31511194e-01
9.35638785e-01 -3.58089179e-01 -7.67158151e-01 2.92609960e-01
9.26552594e-01 -8.22052896e-01 -9.60815430e-01 5.44301271e-01
1.08804977e+00 7.46316612e-02 5.61996698e-01 -1.16091800e+00
-4.36393529e-01 1.19596414e-01 -6.69211447e-01 5.24666309e-02
8.00677240e-01 6.13436341e-01 5.35746753e-01 -9.15864587e-01
9.39474106e-01 1.40278924e+00 9.39063668e-01 1.03043091e+00
-9.82022047e-01 -4.94905293e-01 3.24775949e-02 -1.91046610e-01
-1.28506231e+00 -1.96842760e-01 6.42459810e-01 -4.39956099e-01
1.02006364e+00 -9.91248637e-02 1.18497324e+00 1.87440526e+00
2.39381846e-03 1.05340612e+00 7.90548265e-01 -1.64139926e-01
-2.18255743e-01 -6.79371893e-01 -1.48032978e-01 9.48150277e-01
5.79602718e-01 2.99336314e-01 -6.87998235e-01 -9.71689224e-02
1.30853879e+00 -1.15274332e-01 -8.88654292e-02 -9.65949297e-01
-1.49413478e+00 6.32428706e-01 5.18132448e-01 -3.87135774e-01
-6.14727437e-01 3.99235219e-01 4.59387273e-01 5.78505173e-02
1.74304351e-01 2.73200691e-01 -6.82508588e-01 -1.29407808e-01
-6.14720941e-01 8.46797764e-01 4.59668666e-01 1.15694404e+00
3.85705918e-01 2.24851947e-02 -1.50282726e-01 6.21759832e-01
5.61477393e-02 6.87859476e-01 4.91615206e-01 -1.03584719e+00
7.05694735e-01 5.59980512e-01 2.12966248e-01 -6.70154810e-01
-8.52931917e-01 -4.66425419e-01 -7.13682890e-01 1.06578380e-01
5.88722527e-01 -2.25860134e-01 -1.29568386e+00 1.85678935e+00
7.36514032e-01 -1.91199005e-01 -2.29518637e-01 1.28453338e+00
1.11507046e+00 1.66726679e-01 -7.77370259e-02 3.71125907e-01
1.27452314e+00 -9.47952628e-01 -2.53020376e-01 -6.21177435e-01
2.91163146e-01 -5.23222506e-01 1.20979869e+00 3.96412939e-01
-1.27126443e+00 -8.03251266e-01 -1.00801373e+00 -5.14534056e-01
-8.91536176e-02 3.40312243e-01 7.48069525e-01 6.49618685e-01
-7.85819650e-01 4.79869336e-01 -1.00236583e+00 -5.83409131e-01
5.83867669e-01 4.51486230e-01 -7.77829945e-01 -1.21233530e-01
-9.83858287e-01 7.17230678e-01 4.49767292e-01 -4.01580967e-02
-1.16276753e+00 -4.62969035e-01 -1.06864321e+00 -5.36612809e-01
5.83357036e-01 -1.58476114e+00 1.26050150e+00 -2.74670869e-01
-1.54326320e+00 1.28300238e+00 1.18510060e-01 -5.45152605e-01
1.07066524e+00 -1.08983004e+00 1.94362774e-01 2.75158048e-01
3.21076035e-01 1.22890425e+00 7.50954568e-01 -1.15845764e+00
-2.54014105e-01 -6.91899836e-01 -1.09755047e-01 4.22551244e-01
-7.64562488e-02 -3.13964099e-01 -9.21242476e-01 -8.58687818e-01
4.29571122e-01 -1.11406696e+00 -3.66072863e-01 1.93197027e-01
-6.08572364e-01 -1.36190400e-01 3.97548944e-01 -7.88683474e-01
6.30236745e-01 -1.53395700e+00 5.11869669e-01 -7.59383217e-02
1.24596961e-01 1.52559280e-01 -9.61547419e-02 1.44551903e-01
5.19546270e-02 -3.40947211e-01 -8.66172016e-02 -4.27521020e-01
2.22844064e-01 1.48956895e-01 1.04211830e-01 6.14866316e-01
1.11828476e-01 1.19541287e+00 -8.17555666e-01 -5.83618879e-01
4.64760691e-01 4.54405010e-01 -8.09187114e-01 3.60373288e-01
-1.48730978e-01 7.16228783e-01 -1.96700856e-01 8.86234522e-01
3.51065427e-01 -4.66839433e-01 7.86366910e-02 -2.50458330e-01
4.99313086e-01 1.58413991e-01 -1.21032298e+00 2.26098728e+00
1.11661823e-02 1.07799292e-01 -4.41662312e-01 -4.62097257e-01
7.78347194e-01 2.26415977e-01 6.40739858e-01 -2.59185106e-01
3.59130234e-01 9.32907611e-02 -4.48962122e-01 -4.41518217e-01
2.65069366e-01 9.50437114e-02 -3.26553375e-01 2.48009905e-01
2.65312105e-01 -4.83939759e-02 1.03301428e-01 -9.54553206e-03
1.04488170e+00 8.81269813e-01 4.66422349e-01 1.22496083e-01
2.11929306e-01 2.70428956e-02 6.82107687e-01 6.05149508e-01
-2.95997888e-01 1.02714491e+00 2.40961969e-01 -9.99219000e-01
-1.42954445e+00 -1.45027268e+00 2.03252122e-01 9.21615183e-01
5.89002483e-02 -3.09322536e-01 -6.55582428e-01 -8.67436349e-01
3.00407767e-01 1.11887291e-01 -7.89863050e-01 -3.28626931e-02
-8.98286462e-01 -5.21938562e-01 8.39003503e-01 1.06674969e+00
6.42822564e-01 -9.47189331e-01 -8.91321540e-01 -5.51319681e-02
-4.53925043e-01 -1.44788742e+00 -3.90807837e-01 -2.17663899e-01
-9.61483061e-01 -1.26776361e+00 -1.10075366e+00 -6.85827255e-01
6.14706159e-01 1.14781767e-01 1.28223753e+00 -2.12955967e-01
-5.06371260e-01 6.35680735e-01 -2.38773733e-01 -3.00452381e-01
-2.40454804e-02 2.12867603e-01 4.80014920e-01 -4.53649610e-01
4.80237275e-01 -4.02351022e-01 -7.35157549e-01 4.39915687e-01
-3.65450859e-01 7.44986460e-02 8.42999041e-01 8.09320271e-01
7.93694735e-01 -5.40827870e-01 2.46832415e-01 -5.29903412e-01
2.00436592e-01 -1.51763439e-01 -2.75735945e-01 -6.56769052e-02
-1.49103597e-01 -7.70631805e-02 2.45295510e-01 -3.30057979e-01
-9.82003927e-01 6.54762268e-01 -2.52607137e-01 -7.28474796e-01
-5.40373862e-01 -1.17374606e-01 -1.76071763e-01 7.00249001e-02
1.13344395e+00 3.01458891e-02 1.29065707e-01 -7.58270800e-01
4.93948758e-01 2.64244407e-01 1.00490618e+00 -8.34026158e-01
1.06024182e+00 7.32249677e-01 2.60366023e-01 -5.41187108e-01
-9.53769326e-01 -4.42438573e-01 -1.24539137e+00 -2.20775276e-01
9.83325720e-01 -1.33595419e+00 -4.48733121e-01 4.75607276e-01
-9.78428423e-01 -4.52561200e-01 -2.80878276e-01 7.60467887e-01
-9.86022949e-01 6.13849401e-01 -8.32429528e-01 -4.93991107e-01
-4.68840092e-01 -8.81423235e-01 1.55726159e+00 -9.39815268e-02
-5.59984148e-01 -3.71762455e-01 1.95957616e-01 7.30238378e-01
-2.81051546e-01 9.44589972e-01 4.92255718e-01 -3.13177407e-01
-4.61955518e-01 -4.89767820e-01 6.24157023e-03 2.04128399e-01
-7.17333630e-02 -5.17996013e-01 -7.18362987e-01 -5.24870932e-01
-5.10732710e-01 -1.01181674e+00 6.85706794e-01 4.07559395e-01
1.13760650e+00 -1.66107431e-01 -2.46288598e-01 7.30129957e-01
7.96422601e-01 -6.65750325e-01 6.33150518e-01 3.84823799e-01
1.02622926e+00 4.32600111e-01 6.42206728e-01 6.94283128e-01
3.50243717e-01 7.43875682e-01 4.04104143e-01 -9.38776135e-03
-4.91770715e-01 -6.96746647e-01 8.64198357e-02 3.85162324e-01
-5.71860194e-01 1.69391394e-01 -8.33197355e-01 3.97143096e-01
-1.81195903e+00 -8.74369442e-01 1.37424007e-01 2.23585033e+00
7.36268282e-01 3.18826199e-01 9.40750659e-01 1.93408981e-01
6.06628835e-01 6.27516806e-02 -8.40570629e-01 4.42390621e-01
-1.48421064e-01 3.89050394e-01 4.90170300e-01 6.57513067e-02
-1.30585063e+00 1.14981604e+00 6.68010521e+00 3.26367021e-01
-6.31169140e-01 -2.64278706e-02 2.42420822e-01 -5.08057952e-01
2.71240026e-01 -4.84927475e-01 -9.25143957e-01 5.89934327e-02
2.90552735e-01 3.83369952e-01 1.26254454e-01 1.13105893e+00
-1.57935038e-01 3.71045500e-01 -1.25995719e+00 1.42361188e+00
2.54843593e-01 -1.04022145e+00 4.23854321e-01 2.41697356e-01
7.84239709e-01 7.12640584e-02 -3.14817056e-02 2.79886305e-01
2.43317127e-01 -1.00613046e+00 8.67580771e-01 3.31527561e-01
8.60581577e-01 -8.29802573e-01 5.93158901e-01 3.33814055e-01
-1.10557866e+00 -7.61935934e-02 -5.43539941e-01 -2.21483052e-01
2.44468987e-01 1.78828627e-01 -8.16798270e-01 5.58844626e-01
9.02937889e-01 7.84644186e-01 -7.58642018e-01 1.05963945e+00
-6.01541698e-01 3.31975259e-02 -4.55060482e-01 1.76801831e-02
7.28792027e-02 4.46164638e-01 5.34321189e-01 9.25392747e-01
1.38552606e-01 7.95889199e-02 4.01518553e-01 4.07275766e-01
-2.03798637e-01 -7.16370344e-02 -5.72732031e-01 1.31533965e-01
3.18606436e-01 9.28304136e-01 -5.76062560e-01 -2.40164712e-01
-2.08524570e-01 1.18110383e+00 3.74604255e-01 1.65544987e-01
-9.20625508e-01 4.60881414e-03 5.87967932e-01 3.08743447e-01
2.47542724e-01 -4.51061815e-01 -3.66177589e-01 -1.26121509e+00
2.36183256e-01 -1.08271289e+00 7.49723434e-01 -8.06166410e-01
-1.45059931e+00 4.22824979e-01 1.87974080e-01 -1.26348829e+00
-4.94826287e-01 -9.43024576e-01 1.23623319e-01 3.67929280e-01
-8.84308040e-01 -1.49817765e+00 -6.71339273e-01 8.67292762e-01
5.18023133e-01 -2.90351033e-01 6.77029848e-01 2.81848371e-01
-3.33108932e-01 7.72122920e-01 -6.54143631e-01 4.99457091e-01
8.01069260e-01 -1.19386792e+00 9.82236683e-01 7.34918773e-01
2.37078324e-01 4.40295905e-01 6.11359060e-01 -9.74187851e-01
-1.33345234e+00 -9.76279676e-01 6.50253475e-01 -1.06497002e+00
1.08037874e-01 -3.80411118e-01 -3.41083705e-01 1.10644996e+00
-2.45541826e-01 9.57216993e-02 6.75214410e-01 2.40058675e-01
-5.01112521e-01 2.26604745e-01 -1.14858902e+00 6.92970276e-01
1.99391818e+00 -1.62281573e-01 -9.25828576e-01 6.78549945e-01
5.74236810e-01 -1.11088324e+00 -8.77843857e-01 8.17073226e-01
8.85702550e-01 -8.50911379e-01 1.52485478e+00 -1.09281170e+00
4.84163642e-01 -2.53792018e-01 -2.69779056e-01 -8.31576109e-01
-4.27329332e-01 -5.96189320e-01 -6.07576311e-01 3.09919536e-01
7.64132589e-02 -1.03613466e-01 1.42433202e+00 2.13216484e-01
-4.86990921e-02 -9.44759786e-01 -8.89272332e-01 -9.31164980e-01
-3.14029194e-02 -2.47229621e-01 6.34907961e-01 5.43068945e-01
-4.23107803e-01 2.74482161e-01 -8.93949687e-01 2.59113193e-01
1.03056133e+00 -8.21185336e-02 1.63081181e+00 -1.29499352e+00
-4.46075559e-01 -1.85875714e-01 -7.78702974e-01 -1.56908655e+00
2.95544695e-03 -6.36319220e-01 -1.46734610e-01 -1.46428668e+00
2.17126548e-01 4.04348271e-03 1.51943147e-01 6.12297952e-01
-2.53444284e-01 7.24358976e-01 1.51739061e-01 1.97514787e-01
-6.38718188e-01 6.13384485e-01 1.38866091e+00 1.87790588e-01
-5.85010685e-02 6.10800423e-02 -3.99521261e-01 8.27511549e-01
6.51440561e-01 -2.08995000e-01 -3.23853135e-01 -5.82252026e-01
7.78013393e-02 -2.42750734e-01 8.61567080e-01 -1.43492436e+00
-1.89313889e-01 2.55793463e-02 1.24198139e+00 -9.85691130e-01
7.04510331e-01 -3.97321254e-01 4.76061106e-02 7.68704236e-01
-1.36649504e-01 7.86641166e-02 6.44780323e-02 6.67554855e-01
3.38686675e-01 2.64354825e-01 7.90656030e-01 -6.20691240e-01
-8.17471743e-01 6.85630739e-01 2.11060777e-01 4.84996527e-01
9.09114301e-01 -3.83078128e-01 2.50615906e-02 -4.47260827e-01
-1.02483296e+00 1.12082914e-01 5.99740863e-01 6.51656926e-01
7.16344059e-01 -1.72950697e+00 -7.98328161e-01 1.60347760e-01
1.55271664e-01 3.24673861e-01 1.65926918e-01 2.10268110e-01
-9.07478333e-01 4.71566230e-01 -6.41719460e-01 -8.26812506e-01
-1.08460855e+00 4.94422168e-01 3.24981868e-01 -1.07061647e-01
-9.10778046e-01 1.07706106e+00 -1.50870934e-01 -9.14097011e-01
4.98931020e-01 -1.81117002e-02 4.01931882e-01 -4.48844731e-01
2.61748701e-01 6.54750764e-01 -1.30651325e-01 -8.09657872e-01
-3.70046109e-01 8.03494275e-01 1.82738602e-01 -3.61702628e-02
1.29896545e+00 1.20431237e-01 5.57102740e-01 1.19150236e-01
1.13447642e+00 -3.58736426e-01 -1.58894312e+00 -2.67555118e-01
-5.08850217e-01 -7.55393982e-01 -4.95217294e-01 -7.56478727e-01
-8.82599413e-01 8.70404482e-01 5.01478732e-01 -6.46559477e-01
8.36350679e-01 2.30131671e-01 1.29774058e+00 6.47582591e-01
7.01192677e-01 -1.19509161e+00 6.47801995e-01 3.76825422e-01
9.30586040e-01 -1.22780335e+00 3.39621872e-01 -4.47321057e-01
-5.54346681e-01 1.05579793e+00 1.05686581e+00 -2.71824419e-01
2.94841886e-01 -1.53551474e-01 5.63777350e-02 -3.74294788e-01
-3.17243546e-01 -4.00126189e-01 4.46536452e-01 8.22902083e-01
2.66140282e-01 -2.47858907e-03 -1.00931451e-01 5.41322470e-01
-6.51657879e-01 8.83515552e-02 2.04534307e-02 1.01743579e+00
-3.08964491e-01 -9.11331117e-01 -5.41870058e-01 2.14678586e-01
-2.33225301e-01 4.26846206e-01 -6.93181992e-01 1.19031835e+00
2.36082405e-01 3.68062526e-01 -2.07000077e-01 -4.75008667e-01
6.80006742e-01 8.05389658e-02 1.04308176e+00 -5.96697986e-01
-2.85656065e-01 7.16658533e-02 1.52144507e-01 -9.14848149e-01
-2.20732152e-01 -7.78168082e-01 -1.15279424e+00 -2.15115204e-01
2.41130441e-01 -5.06645918e-01 3.10738951e-01 7.43807554e-01
4.99329478e-01 2.98961669e-01 8.79622772e-02 -1.44352937e+00
-8.11565936e-01 -9.16356087e-01 -2.29815096e-01 8.34175289e-01
-4.42614183e-02 -1.10243249e+00 -7.33078793e-02 -2.45223269e-02] | [6.9748406410217285, -0.886874794960022] |
8259321d-c68f-4cd1-ae81-2c50f1dd91eb | object-driven-active-mapping-for-more | 2012.01788 | null | https://arxiv.org/abs/2012.01788v3 | https://arxiv.org/pdf/2012.01788v3.pdf | Object SLAM-Based Active Mapping and Robotic Grasping | This paper presents the first active object mapping framework for complex robotic manipulation and autonomous perception tasks. The framework is built on an object SLAM system integrated with a simultaneous multi-object pose estimation process that is optimized for robotic grasping. Aiming to reduce the observation uncertainty on target objects and increase their pose estimation accuracy, we also design an object-driven exploration strategy to guide the object mapping process, enabling autonomous mapping and high-level perception. Combining the mapping module and the exploration strategy, an accurate object map that is compatible with robotic grasping can be generated. Additionally, quantitative evaluations also indicate that the proposed framework has a very high mapping accuracy. Experiments with manipulation (including object grasping and placement) and augmented reality significantly demonstrate the effectiveness and advantages of our proposed framework. | ['Zhiqiang Deng', 'Xinggang Hu', 'Wenkai Sun', 'Sonya Coleman', 'Xin Chen', 'Delong Zhu', 'Yunzhou Zhang', 'Yanmin Wu'] | 2020-12-03 | null | null | null | null | ['object-slam'] | ['computer-vision'] | [-1.02777466e-01 7.25988373e-02 -1.65414855e-01 -3.97821993e-01
-2.80569315e-01 -4.94228512e-01 3.68868172e-01 -1.65890902e-02
-1.45257249e-01 4.09583360e-01 -2.46728197e-01 3.07208419e-01
-8.01086247e-01 -8.41425121e-01 -6.99520767e-01 -4.27432209e-01
-2.11284563e-01 9.46752131e-01 3.76041740e-01 -1.17735989e-01
6.07143581e-01 9.70606267e-01 -1.75171280e+00 -7.02062473e-02
9.62297142e-01 1.11852622e+00 1.44068849e+00 2.73879766e-01
-2.37004578e-01 3.81358057e-01 -2.17319161e-01 3.58504057e-01
5.30431628e-01 5.62996805e-01 -6.16528332e-01 1.58167064e-01
-9.62281302e-02 -8.21557641e-01 -4.46875542e-02 1.02474177e+00
2.89140940e-01 3.26319113e-02 3.37019056e-01 -1.36977303e+00
-4.19331372e-01 7.17157364e-01 -1.33406341e-01 -5.56230128e-01
4.06835556e-01 2.73554260e-03 6.02628052e-01 -1.04446340e+00
7.82444000e-01 1.57835209e+00 9.02547315e-02 2.20756203e-01
-6.12393379e-01 -4.56834346e-01 2.30182454e-01 3.42640400e-01
-1.32754350e+00 -6.28711730e-02 8.46452415e-01 -3.80594611e-01
5.21457732e-01 2.33051077e-01 5.94016552e-01 6.69413209e-01
5.14758885e-01 6.32042587e-01 1.00139785e+00 -4.14174229e-01
2.20342889e-01 4.79095161e-01 9.93253514e-02 3.88568431e-01
5.33685386e-01 1.51353255e-01 -4.46867734e-01 8.86728708e-03
9.62006390e-01 3.53827387e-01 -1.71001464e-01 -1.28454340e+00
-1.56915140e+00 3.66292357e-01 7.51266956e-01 9.01924148e-02
-8.19977105e-01 2.23441318e-01 1.07172512e-01 -1.38039082e-01
-1.49335846e-01 4.63302374e-01 -1.06913731e-01 5.29114194e-02
1.15873717e-01 3.80754381e-01 7.69249260e-01 1.88528669e+00
8.57369959e-01 -2.27203876e-01 -1.23024389e-01 6.08255148e-01
8.26662302e-01 7.72757530e-01 -7.95130506e-02 -1.10247457e+00
3.33735436e-01 9.33467090e-01 6.55917048e-01 -1.00050092e+00
-5.87600410e-01 -1.12436317e-01 -2.27027744e-01 6.41762912e-01
1.00500938e-02 4.46353614e-01 -7.37909794e-01 1.24341762e+00
5.63600779e-01 -6.19288862e-01 8.25647116e-02 1.09884536e+00
6.17511988e-01 3.96416754e-01 1.72108337e-02 -3.72891799e-02
1.30884874e+00 -8.70257854e-01 -1.06113851e+00 -1.08997405e-01
1.01051293e-01 -7.40239799e-01 7.51415431e-01 5.41886866e-01
-8.53014290e-01 -6.10192537e-01 -1.21928787e+00 -5.38678095e-02
-2.73877740e-01 5.06629109e-01 8.90199959e-01 2.89106686e-02
-5.40837228e-01 4.29287106e-01 -8.95469069e-01 -5.26530921e-01
2.42971063e-01 5.46372712e-01 -4.88928556e-01 -1.58800080e-01
-4.66924667e-01 1.55784464e+00 1.12816966e+00 6.09505773e-01
-1.16506803e+00 -3.41770589e-01 -5.28532565e-01 1.34856105e-02
5.89994252e-01 -1.77826449e-01 1.04765737e+00 -1.16233289e-01
-1.64931226e+00 4.92730528e-01 2.25519434e-01 3.20200995e-02
5.12298524e-01 -6.97135746e-01 1.07182026e-01 3.21755648e-01
1.40603736e-01 6.56300128e-01 4.92044836e-01 -1.83509040e+00
-5.60886562e-01 -6.71017289e-01 1.75606251e-01 6.38060093e-01
-2.68829107e-01 -1.77079871e-01 -2.64253020e-01 -3.22018377e-02
1.00524104e+00 -1.00579357e+00 -1.76131576e-01 4.92169231e-01
-1.23790681e-01 -2.83080876e-01 1.32823682e+00 -5.01049161e-01
3.10005575e-01 -1.91593778e+00 4.18968856e-01 2.62739331e-01
-1.28126949e-01 -1.93059698e-01 -4.36098352e-02 6.05874538e-01
5.75986147e-01 -5.27119994e-01 2.90414959e-01 2.55676098e-02
2.07252145e-01 3.15101683e-01 -5.01539052e-01 3.93782794e-01
-7.33915344e-02 9.07288611e-01 -8.61717045e-01 -3.86126071e-01
5.90843976e-01 1.28891468e-01 -2.34410584e-01 5.39498150e-01
-4.44130003e-01 5.42312920e-01 -8.45656872e-01 1.19854999e+00
9.77455139e-01 1.80793792e-01 3.59447092e-01 -3.48016739e-01
-5.69276989e-01 -1.14412881e-01 -1.49500895e+00 2.07625055e+00
-4.91563201e-01 -1.22680627e-02 5.57184577e-01 -4.97191280e-01
1.64396632e+00 3.90766375e-02 6.80090308e-01 -2.53457040e-01
2.84701407e-01 5.67354500e-01 -9.30665508e-02 -7.81464756e-01
7.26289213e-01 6.00885928e-01 2.43389100e-01 1.38427213e-01
5.03474325e-02 -6.05396569e-01 -1.78868741e-01 -1.65693939e-01
5.61060250e-01 9.52228129e-01 2.22438812e-01 -6.72361970e-01
4.13155496e-01 4.19743657e-01 3.15477341e-01 6.19506657e-01
-5.43416478e-03 -3.09810732e-02 -2.99046129e-01 -2.38709152e-01
-9.84055042e-01 -1.36271143e+00 -3.81547123e-01 5.77668011e-01
1.14887416e+00 1.52707487e-01 -2.78781444e-01 -2.68449247e-01
4.17589575e-01 5.26571453e-01 -2.60634452e-01 -1.43469889e-02
-6.92023933e-01 -2.59588540e-01 -3.89037371e-01 5.21805048e-01
6.70850456e-01 -1.26333272e+00 -1.08971596e+00 3.04872036e-01
1.20353578e-02 -1.11850214e+00 2.26071879e-01 -1.81754176e-02
-9.44969356e-01 -9.77778614e-01 -2.20415816e-01 -9.46429193e-01
8.96450460e-01 5.72440624e-01 2.30600908e-01 -4.66629453e-02
-4.17085797e-01 5.55689812e-01 -8.03067863e-01 -7.16428518e-01
-4.59104836e-01 -1.14264652e-01 1.84314847e-01 -3.11211497e-01
7.76633695e-02 -4.65623140e-01 -4.19616133e-01 5.70759773e-01
-5.10755122e-01 -4.24627401e-03 1.01463413e+00 4.40588504e-01
4.39846426e-01 -1.12367429e-01 6.49789155e-01 1.24836795e-01
3.41340631e-01 -2.53977865e-01 -1.03882515e+00 4.41824168e-01
-5.50215602e-01 -4.32361662e-01 -7.48211816e-02 -6.08746886e-01
-1.09061086e+00 4.83060241e-01 3.70211512e-01 -4.67243582e-01
1.78836193e-02 1.80382535e-01 -5.40352345e-01 -4.97185946e-01
3.22797179e-01 9.23961028e-02 5.92917740e-01 -6.46629930e-01
3.70277286e-01 1.00876856e+00 5.67688823e-01 -7.01009631e-01
6.45268798e-01 5.69945693e-01 1.70722321e-01 -4.37152117e-01
-2.65578717e-01 -3.65467906e-01 -1.19701064e+00 -6.42274678e-01
7.42398620e-01 -9.61419880e-01 -1.36087954e+00 3.76194656e-01
-1.46206367e+00 3.73360552e-02 -4.41679470e-02 1.00566721e+00
-9.95499671e-01 2.76083440e-01 -1.50726900e-01 -1.14032543e+00
-3.90782714e-01 -1.51615214e+00 1.28066480e+00 1.37252316e-01
2.15044562e-02 -2.22495675e-01 -5.11336982e-01 6.18757308e-02
4.59983438e-01 2.55451858e-01 5.18223584e-01 -3.82198542e-01
-1.69069195e+00 -2.11804822e-01 -3.63207906e-01 -1.91294551e-01
3.67858171e-01 -2.31060982e-01 -4.54882383e-01 -3.60895455e-01
-2.69705933e-02 -2.78970659e-01 1.43164679e-01 -1.48722436e-02
9.26509857e-01 4.23023589e-02 -7.41090298e-01 1.65658325e-01
1.58255696e+00 5.88223040e-01 5.73089123e-01 6.53410137e-01
4.50114995e-01 9.85397160e-01 1.82442880e+00 4.42263097e-01
2.17554763e-01 1.15728676e+00 1.28234267e+00 5.57976782e-01
1.43913969e-01 -1.67293400e-01 4.61069159e-02 5.34923315e-01
1.02952383e-01 -1.38389328e-02 -8.21917295e-01 3.78881514e-01
-2.15419483e+00 -4.64318871e-01 -1.70260206e-01 2.01366830e+00
2.44384110e-01 -1.08532891e-01 -3.88810873e-01 -1.19181834e-01
7.47464955e-01 -4.55584735e-01 -5.72016597e-01 -1.05693616e-01
3.34001541e-01 -3.72860372e-01 4.71745998e-01 5.94284117e-01
-8.11973393e-01 1.04048336e+00 5.87423468e+00 3.79546881e-01
-9.74110425e-01 -1.78275688e-03 -8.04234266e-01 3.98452461e-01
-8.76036659e-02 1.51293427e-01 -8.93096209e-01 1.55145988e-01
-1.12117141e-01 -2.68645249e-02 3.48581105e-01 1.51072752e+00
-5.26556224e-02 -3.81651074e-01 -1.02480376e+00 7.57226765e-01
-1.00595094e-01 -1.25382149e+00 1.26517296e-01 3.61928120e-02
2.11954623e-01 -1.66656688e-01 -3.76738012e-01 7.43129402e-02
1.06464259e-01 -4.73231524e-01 1.25711215e+00 7.50096023e-01
3.18545818e-01 -3.93082857e-01 6.19988501e-01 7.15502203e-01
-1.07397377e+00 -6.58025444e-01 -5.91432095e-01 -9.26269591e-02
5.05557120e-01 2.93402612e-01 -1.23568106e+00 8.62894773e-01
8.33946347e-01 3.31350803e-01 -6.57890588e-02 1.20669603e+00
-1.57719210e-01 -5.54717541e-01 -3.15160692e-01 -5.24275959e-01
-1.47152036e-01 -5.05082846e-01 1.05388820e+00 4.21917707e-01
4.42479879e-01 1.73765123e-01 6.41659558e-01 1.34167349e+00
5.73086560e-01 2.23137707e-01 -7.08891213e-01 7.47888610e-02
1.01814449e+00 1.42456210e+00 -7.58662581e-01 -1.72108352e-01
2.01403528e-01 6.87577188e-01 2.73025632e-01 -2.23408230e-02
-6.99764788e-01 -3.42021734e-01 2.94618905e-01 -1.26239359e-01
5.98657764e-02 -7.98574626e-01 -4.32873815e-01 -6.39350414e-01
4.88279432e-01 -2.30253413e-01 -4.64849859e-01 -1.13165081e+00
-5.94451189e-01 4.50152189e-01 4.90918159e-01 -1.38770759e+00
-1.77898765e-01 -8.90124261e-01 -8.03443789e-02 8.02295864e-01
-1.26255834e+00 -1.35245776e+00 -9.98016953e-01 6.51516691e-02
4.57427472e-01 -2.28008628e-01 8.07839870e-01 -9.12327915e-02
-8.42849351e-03 -3.37687701e-01 -1.42704636e-01 -4.28632677e-01
3.30793530e-01 -7.98231840e-01 -4.08220291e-01 4.74474967e-01
-5.10902822e-01 8.70916486e-01 6.01471305e-01 -1.02976203e+00
-2.16194391e+00 -8.54747355e-01 -2.71230880e-02 -5.67487836e-01
3.25289816e-01 -6.66270733e-01 -7.31774271e-01 8.46374094e-01
-2.99973309e-01 -2.59058088e-01 -4.64605242e-01 -2.26250395e-01
2.05940425e-01 -2.85628229e-01 -1.45070648e+00 4.26113337e-01
1.05413878e+00 -3.23251300e-02 -7.52034366e-01 6.70170307e-01
7.86545932e-01 -8.05464566e-01 -1.11950910e+00 1.22130787e+00
9.81907308e-01 -4.80786771e-01 9.71542358e-01 -7.49523118e-02
-7.15733990e-02 -4.72117037e-01 -4.56781179e-01 -7.06186771e-01
-4.26342875e-01 -2.49491110e-01 -1.53056398e-01 1.14485025e+00
-1.59943536e-01 -7.80229092e-01 6.85390890e-01 3.12801927e-01
-4.74558324e-01 -5.44100940e-01 -8.93622458e-01 -1.09566963e+00
-6.53499186e-01 6.92743212e-02 6.98714793e-01 3.66030693e-01
-1.52575495e-02 -3.43644142e-01 -1.19862735e-01 8.73889148e-01
7.71169245e-01 5.30636132e-01 1.12797093e+00 -1.49253929e+00
2.74318606e-01 -1.29085243e-01 -5.34060717e-01 -1.04157329e+00
2.17859447e-01 -6.67938292e-01 6.70980573e-01 -1.82536936e+00
2.01610610e-01 -1.04106736e+00 1.29977345e-01 3.73837620e-01
2.95993507e-01 -1.42407537e-01 3.32606971e-01 5.23821771e-01
-1.86810970e-01 6.86481833e-01 1.46078861e+00 2.27816910e-01
-2.80811280e-01 -1.67081177e-01 -2.36562118e-02 5.08603513e-01
6.28807664e-01 -1.87711433e-01 -3.64262104e-01 -5.45996964e-01
-4.14157361e-01 2.39626437e-01 6.44576967e-01 -9.46149707e-01
2.48638108e-01 -3.95256698e-01 4.61886674e-02 -1.06870127e+00
8.61870110e-01 -1.59404993e+00 4.20914352e-01 9.07025576e-01
-8.84509236e-02 -2.49130845e-01 8.08512419e-02 5.85261822e-01
8.22947025e-02 -4.27933544e-01 5.39606273e-01 -5.08206375e-02
-1.08245826e+00 3.04649740e-01 1.53918073e-01 -1.18349099e+00
1.72588444e+00 -3.01733762e-01 -2.76015401e-01 4.52048518e-02
-7.95752883e-01 4.54495490e-01 5.13312697e-01 9.97078359e-01
8.79383087e-01 -1.46708214e+00 -3.65245968e-01 2.23787710e-01
3.83460283e-01 2.59873480e-01 9.94514003e-02 6.35674715e-01
-8.53286922e-01 4.64291662e-01 -7.99970865e-01 -1.15386283e+00
-1.04737878e+00 6.55505657e-01 -1.49381384e-01 4.33519125e-01
-6.83804512e-01 3.63765061e-01 -1.30329698e-01 -9.31507111e-01
3.40569407e-01 -2.13754535e-01 -1.92788526e-01 -3.81272674e-01
3.20953399e-01 6.78564072e-01 -1.58436805e-01 -3.71438116e-01
-5.09759128e-01 6.42512083e-01 1.50603577e-01 -4.69130054e-02
1.33317029e+00 -4.40021068e-01 -6.27080142e-01 3.61938298e-01
6.26395166e-01 -7.18385354e-03 -1.26707816e+00 2.63112653e-02
1.68235898e-01 -7.86749363e-01 -2.47491360e-01 -8.43984485e-01
-3.18644702e-01 5.43358982e-01 7.71671295e-01 -1.81708157e-01
5.99832118e-01 5.44828475e-02 1.72960714e-01 8.00047994e-01
1.30164313e+00 -1.09760010e+00 1.91073239e-01 1.87400058e-01
1.77498066e+00 -1.34116352e+00 2.83712357e-01 -1.14882576e+00
-2.77299613e-01 1.32891476e+00 1.14754999e+00 -1.41955376e-01
5.05236804e-01 3.42808038e-01 -6.33282736e-02 -2.68828034e-01
-2.91157335e-01 5.75883454e-03 1.21917196e-01 6.82623565e-01
-4.48787451e-01 1.66125074e-02 -3.25891882e-01 1.85176656e-01
6.74245358e-02 1.55049160e-01 3.41677159e-01 1.34380817e+00
-1.28246105e+00 -8.81766856e-01 -6.10337317e-01 -7.02711120e-02
3.62314075e-01 6.68189824e-01 3.69566269e-02 1.10441470e+00
1.06001683e-02 7.77864039e-01 1.76715180e-02 -2.56344199e-01
4.48992223e-01 -2.47819096e-01 9.52172518e-01 -7.19241321e-01
6.41264990e-02 -2.12593332e-01 -1.27090558e-01 -9.24524486e-01
-2.87657738e-01 -4.46413934e-01 -1.55745780e+00 4.23373282e-01
-7.53629208e-01 -9.57490876e-02 1.55961239e+00 7.76548684e-01
3.54480475e-01 3.93026352e-01 6.16168201e-01 -1.22013628e+00
-1.17659295e+00 -1.14854169e+00 -6.22474074e-01 -4.61735912e-02
-6.08121082e-02 -1.25634384e+00 3.31023633e-02 -5.51318049e-01] | [5.897825717926025, -0.9387659430503845] |
1903a1a4-5863-4c2f-9738-5eaf997bcb04 | meta-voice-fast-few-shot-style-transfer-for | 2111.07218 | null | https://arxiv.org/abs/2111.07218v1 | https://arxiv.org/pdf/2111.07218v1.pdf | Meta-Voice: Fast few-shot style transfer for expressive voice cloning using meta learning | The task of few-shot style transfer for voice cloning in text-to-speech (TTS) synthesis aims at transferring speaking styles of an arbitrary source speaker to a target speaker's voice using very limited amount of neutral data. This is a very challenging task since the learning algorithm needs to deal with few-shot voice cloning and speaker-prosody disentanglement at the same time. Accelerating the adaptation process for a new target speaker is of importance in real-world applications, but even more challenging. In this paper, we approach to the hard fast few-shot style transfer for voice cloning task using meta learning. We investigate the model-agnostic meta-learning (MAML) algorithm and meta-transfer a pre-trained multi-speaker and multi-prosody base TTS model to be highly sensitive for adaptation with few samples. Domain adversarial training mechanism and orthogonal constraint are adopted to disentangle speaker and prosody representations for effective cross-speaker style transfer. Experimental results show that the proposed approach is able to conduct fast voice cloning using only 5 samples (around 12 second speech data) from a target speaker, with only 100 adaptation steps. Audio samples are available online. | ['Dong Yu', 'Dan Su', 'Songxiang Liu'] | 2021-11-14 | null | null | null | null | ['voice-cloning'] | ['speech'] | [ 5.52625656e-01 7.59155676e-02 -1.19581111e-01 -2.72632360e-01
-1.33381212e+00 -4.51906592e-01 5.62180340e-01 -4.53681886e-01
-3.79378766e-01 7.98224330e-01 2.75858968e-01 -2.31637537e-01
4.62338239e-01 -2.30750412e-01 -5.86847842e-01 -7.79572964e-01
3.61123651e-01 5.41877151e-01 9.38642547e-02 -4.00455832e-01
-1.12238742e-01 4.67702955e-01 -1.38249242e+00 3.84969920e-01
7.54229069e-01 7.57848084e-01 4.78939414e-01 1.10803783e+00
-3.62635374e-01 4.81420904e-01 -8.35288107e-01 -3.22137922e-01
4.17790301e-02 -8.97549450e-01 -7.50864983e-01 4.30248119e-02
3.41074556e-01 -1.04059927e-01 1.30066514e-01 1.04708052e+00
1.17409790e+00 3.49719405e-01 8.00190330e-01 -1.20738530e+00
-3.46295863e-01 6.87204421e-01 -2.42019653e-01 3.40381145e-01
1.58116922e-01 9.06256363e-02 6.45873308e-01 -1.00928664e+00
3.14321548e-01 1.50045753e+00 4.82485622e-01 1.19775486e+00
-1.17157257e+00 -1.03097713e+00 -7.20763877e-02 1.24563955e-01
-1.20333838e+00 -1.06011975e+00 1.03972316e+00 -1.57304540e-01
6.39751017e-01 5.21301806e-01 3.08301181e-01 1.64212763e+00
5.48813120e-02 6.30596817e-01 1.04574418e+00 -5.59131861e-01
4.21719700e-01 6.69851899e-01 -3.57255727e-01 1.28484398e-01
-5.74506283e-01 1.90531552e-01 -6.14430249e-01 -2.11190924e-01
3.46531272e-01 -3.58468503e-01 -1.44998312e-01 -3.28305699e-02
-1.01702809e+00 7.66288817e-01 -5.18635660e-02 4.99650091e-01
-1.33762106e-01 -2.42449284e-01 6.88647807e-01 8.43759954e-01
6.65991306e-01 2.47124255e-01 -4.89352822e-01 -3.66357118e-01
-9.87837374e-01 1.12533800e-01 7.52876222e-01 1.22075677e+00
3.48766893e-01 9.33090210e-01 -1.01074897e-01 1.30588520e+00
-8.33728388e-02 6.29752755e-01 1.07319140e+00 -4.98871863e-01
5.38492858e-01 -2.02557623e-01 -1.69881359e-02 -2.91380107e-01
1.02846503e-01 -3.91495734e-01 -9.60935116e-01 4.95818168e-01
9.74873304e-02 -5.27315438e-01 -7.68683314e-01 1.85900176e+00
5.05957365e-01 5.43053031e-01 2.36535773e-01 6.42141044e-01
6.49909019e-01 1.13755810e+00 5.26553355e-02 -7.46464193e-01
1.23192537e+00 -1.04214072e+00 -8.89867425e-01 -3.41208339e-01
3.56407553e-01 -1.20428193e+00 1.56361532e+00 1.33199900e-01
-1.24355245e+00 -9.85541284e-01 -1.01971352e+00 2.51211494e-01
-2.15586543e-01 -1.96260020e-01 -1.51663169e-01 9.74477708e-01
-6.22452378e-01 4.47647423e-01 -4.27046895e-01 -2.73444384e-01
1.31341815e-01 5.16583204e-01 -2.71179080e-01 2.63758510e-01
-1.26189244e+00 7.54680812e-01 3.40353519e-01 -3.94206583e-01
-9.22653139e-01 -9.55426931e-01 -6.46127224e-01 2.97970064e-02
1.88322887e-01 -5.36113083e-01 1.45729458e+00 -1.58281481e+00
-2.40959883e+00 7.54399300e-01 -2.05321789e-01 -3.96679640e-01
4.50264484e-01 -1.62171423e-01 -9.05954182e-01 -2.84310967e-01
-3.02683502e-01 3.33559394e-01 1.66915619e+00 -1.11953354e+00
-6.77434504e-01 -1.48494214e-01 -6.80266619e-01 4.86638904e-01
-6.06751382e-01 5.16627014e-01 1.74702168e-01 -9.77792919e-01
-3.34382325e-01 -9.14158344e-01 2.64969975e-01 -3.58290762e-01
-1.96590677e-01 6.68072328e-03 1.07715726e+00 -8.52415085e-01
9.71310854e-01 -2.20289445e+00 5.27200699e-01 -4.04422224e-01
-4.43508923e-01 6.83540583e-01 -3.22268963e-01 3.65001917e-01
-2.86950469e-01 -2.11630538e-01 -3.43293935e-01 -7.66198397e-01
-2.05750212e-01 -6.91506043e-02 -7.17966855e-01 3.13021064e-01
4.56962623e-02 4.80438352e-01 -5.77148914e-01 -5.42225599e-01
2.15349391e-01 4.16994005e-01 -6.06911778e-01 7.80918419e-01
-6.22174107e-02 8.39321971e-01 3.69097404e-02 4.30081993e-01
5.47731280e-01 8.46866310e-01 -2.01435104e-01 -7.13356137e-02
-1.90920278e-03 1.68183908e-01 -1.05795276e+00 1.79406106e+00
-9.73420024e-01 3.74803334e-01 3.28777611e-01 -7.09343255e-01
1.05919099e+00 8.79702568e-01 1.05200842e-01 -3.07349712e-01
3.20397288e-01 3.24005574e-01 2.80125499e-01 -2.16271684e-01
2.25130260e-01 -1.02774417e+00 -1.20525151e-01 4.81755853e-01
5.22997618e-01 -5.95294535e-01 -7.13202000e-01 -3.88965130e-01
5.16834438e-01 -1.64991379e-01 3.66015792e-01 -1.62119284e-01
8.42201948e-01 -5.35587192e-01 7.34875679e-01 3.34056258e-01
-2.75579661e-01 6.41353905e-01 -2.16543362e-01 9.67567116e-02
-1.28998590e+00 -1.20441055e+00 2.36171231e-01 1.46954095e+00
-4.37562287e-01 1.64630339e-01 -9.93174016e-01 -3.05724025e-01
-3.69795293e-01 1.29618549e+00 -2.43244693e-01 -6.20765507e-01
-9.02549148e-01 -2.79287368e-01 7.80425668e-01 2.32694522e-01
6.90890523e-03 -1.29661345e+00 1.23173818e-01 5.04849374e-01
5.56736952e-04 -9.38826740e-01 -1.04881346e+00 1.11012988e-01
-8.63457203e-01 -1.89656869e-01 -1.26049638e+00 -1.20857918e+00
1.67587399e-01 -5.09055927e-02 6.16825938e-01 -7.29571998e-01
-9.05588791e-02 1.99206829e-01 -2.53881097e-01 -5.73604286e-01
-1.35722113e+00 3.24226141e-01 5.69501281e-01 4.42089111e-01
-4.05518785e-02 -9.32641029e-01 -3.25855613e-02 4.13619757e-01
-6.87468112e-01 -1.81652568e-02 3.83540541e-01 1.15343499e+00
2.56908059e-01 -9.11741629e-02 1.17356932e+00 -6.67932808e-01
8.16744566e-01 -3.58760953e-01 -2.94808060e-01 2.15005577e-01
-1.87743232e-01 -1.68266341e-01 1.16415048e+00 -1.07066262e+00
-1.48213315e+00 -6.61380813e-02 -4.48872536e-01 -8.39862406e-01
-2.51131177e-01 -2.41187274e-01 -7.96617806e-01 7.85999447e-02
6.39357746e-01 5.72224796e-01 1.32843122e-01 -6.94336236e-01
6.72233999e-01 1.21146560e+00 7.02156305e-01 -5.57705700e-01
8.95225227e-01 4.52513024e-02 -4.46358413e-01 -1.18158627e+00
-4.05907869e-01 -4.53568548e-01 -7.35786080e-01 -6.42041638e-02
6.66457891e-01 -8.80177379e-01 -4.12761092e-01 7.11622000e-01
-1.19472384e+00 -3.15122932e-01 -5.01825213e-01 5.82793057e-01
-1.01961148e+00 1.25579759e-01 -3.77091914e-01 -1.00049937e+00
-6.67655349e-01 -1.01219738e+00 7.66892672e-01 -1.07443549e-01
-3.54536384e-01 -9.78163421e-01 3.12388539e-01 3.45731050e-01
5.87175786e-01 -1.52192637e-01 9.74460542e-01 -1.01652265e+00
3.84082571e-02 -1.40487058e-02 5.40630877e-01 8.29736114e-01
6.93326354e-01 -3.71929199e-01 -1.40154934e+00 -6.86103761e-01
6.72728360e-01 -3.35943311e-01 3.69805634e-01 4.71740849e-02
6.29150093e-01 -5.54061532e-01 7.86718577e-02 4.90188211e-01
8.72364759e-01 4.65141147e-01 3.37972909e-01 -2.88813263e-01
8.79824221e-01 5.89506328e-01 6.74568236e-01 2.73784846e-01
-3.19689482e-01 8.89158309e-01 -9.32630002e-02 7.18261302e-02
-5.09017169e-01 -2.15942726e-01 7.47868240e-01 1.58705938e+00
2.66619086e-01 -2.21657977e-01 -4.61600482e-01 5.88740528e-01
-1.22971916e+00 -1.08713043e+00 5.36415040e-01 2.48570418e+00
1.22766364e+00 1.74961984e-01 3.70160848e-01 3.26000422e-01
1.18785441e+00 2.06781358e-01 -6.47288442e-01 -7.99941421e-01
9.54349786e-02 5.00991940e-01 -1.51835242e-03 8.42715919e-01
-7.22771645e-01 1.33135331e+00 5.16829967e+00 1.30874681e+00
-1.51838017e+00 5.64257026e-01 3.81541491e-01 -4.29159135e-01
-1.51154488e-01 -2.60329962e-01 -8.84569526e-01 5.93389094e-01
1.39332497e+00 -5.36129594e-01 6.92896485e-01 8.10192704e-01
1.50281385e-01 6.15556777e-01 -1.24740958e+00 1.11265194e+00
2.52490401e-01 -8.22335124e-01 1.49824888e-01 -2.36690611e-01
5.73834479e-01 -2.57710725e-01 3.22925150e-01 5.83899438e-01
-2.01127946e-01 -8.88920665e-01 7.73704946e-01 2.07127705e-01
1.16868532e+00 -1.13213980e+00 2.22699329e-01 5.54996014e-01
-1.08680391e+00 -3.69481812e-03 -4.95241255e-01 2.53452271e-01
2.63540685e-01 1.36058107e-01 -1.22016001e+00 3.35419208e-01
1.25430554e-01 1.41014174e-01 6.68854639e-02 6.21033728e-01
3.10617000e-01 8.62640083e-01 -1.42922420e-02 -6.03033230e-03
-1.16706803e-01 1.24700159e-01 1.07601702e+00 1.16668916e+00
6.45240784e-01 -1.78265378e-01 -1.51449785e-01 6.92462087e-01
-1.09742656e-01 4.60334748e-01 -6.44873977e-01 -6.85376599e-02
5.66034555e-01 8.94702733e-01 -8.68593007e-02 -4.04234886e-01
-1.08894348e-01 1.56186819e+00 1.46526799e-01 2.34524786e-01
-7.68982351e-01 -5.03850937e-01 9.05559957e-01 7.06019253e-03
2.11519435e-01 8.68019555e-03 -1.61123015e-02 -1.00142908e+00
-2.38275468e-01 -1.21171272e+00 7.33335987e-02 -6.52499974e-01
-1.30312610e+00 8.38475108e-01 -2.67590463e-01 -1.52406085e+00
-6.36810660e-01 -2.45369613e-01 -1.03942132e+00 1.34638882e+00
-1.20206320e+00 -1.34676385e+00 2.45137170e-01 8.03007185e-01
1.36482525e+00 -9.11946654e-01 1.19339156e+00 3.11706543e-01
-4.14307088e-01 1.11142528e+00 2.45174617e-01 -1.80457920e-01
1.14350140e+00 -9.35659230e-01 7.57960379e-01 5.17066836e-01
3.10378037e-02 3.01342309e-01 9.83367205e-01 -4.89337564e-01
-1.26769924e+00 -1.14353025e+00 7.30900347e-01 -1.33054554e-01
3.98419917e-01 -6.23803258e-01 -1.14879560e+00 5.00732839e-01
4.09585804e-01 -6.59186319e-02 8.80145967e-01 -3.31079625e-02
-1.21994466e-01 -3.01737875e-01 -1.23737729e+00 8.21499825e-01
7.37214506e-01 -8.55786860e-01 -1.00395179e+00 4.46690954e-02
1.09792745e+00 -2.60945052e-01 -7.00946391e-01 6.59761801e-02
4.09022599e-01 -7.06135035e-01 9.21474755e-01 -9.28281069e-01
-1.63656563e-01 4.56827581e-02 -2.55416840e-01 -1.78024828e+00
-1.82714462e-01 -1.09205508e+00 9.94055122e-02 1.80071092e+00
2.22077087e-01 -5.60703874e-01 4.84850854e-01 1.47740617e-01
-3.08942944e-01 -2.87938505e-01 -1.32288527e+00 -1.13680315e+00
4.83505309e-01 -2.54683137e-01 6.33243859e-01 1.01213956e+00
-8.46532136e-02 7.56610096e-01 -9.53902543e-01 1.78054839e-01
5.57328403e-01 -1.23554148e-01 8.45794439e-01 -8.57254326e-01
-5.85791469e-01 -2.65036076e-01 -1.01445287e-01 -5.85195124e-01
5.26409030e-01 -8.81579101e-01 3.18948515e-02 -5.99181652e-01
-3.21165562e-01 -2.04034984e-01 -3.78375709e-01 -2.67475136e-02
-2.46219724e-01 -1.63976207e-01 3.11470598e-01 6.40172735e-02
1.41251922e-01 1.05849838e+00 1.17076552e+00 -2.48662069e-01
-4.29832280e-01 6.96659267e-01 -3.39864463e-01 6.08396411e-01
8.97506952e-01 -7.69932389e-01 -6.54953361e-01 1.27884880e-01
-7.28322387e-01 4.26175922e-01 -2.51864661e-02 -1.07497466e+00
9.13595110e-02 -1.46214470e-01 7.81210326e-03 -1.96744338e-01
7.86759436e-01 -8.19984138e-01 1.18946925e-01 4.98977214e-01
-4.34764028e-01 -3.71775657e-01 4.76961941e-01 4.31060851e-01
-2.55860686e-01 -4.32187229e-01 1.36508000e+00 -1.25118211e-01
-4.09190357e-01 1.85770646e-01 -3.32533717e-01 2.24406347e-01
8.81059289e-01 -1.45691395e-01 3.35385591e-01 -4.86938089e-01
-9.14075434e-01 -4.80991155e-01 1.48899451e-01 7.63380587e-01
4.48504239e-01 -1.61338806e+00 -9.77649629e-01 4.76132542e-01
3.90743837e-02 -4.46697146e-01 4.79689479e-01 2.40274385e-01
2.04698384e-01 1.11700341e-01 -4.78330195e-01 -4.42617834e-01
-1.63289678e+00 7.92039752e-01 4.41897929e-01 2.49870732e-01
-3.91874254e-01 1.00127041e+00 3.65876555e-01 -5.75493693e-01
3.75660956e-01 -1.46786809e-01 1.12501003e-01 1.48191467e-01
6.44296944e-01 4.94010299e-01 2.16057291e-03 -7.48766422e-01
-1.35675907e-01 5.05939007e-01 -2.53757417e-01 -6.19070113e-01
1.02041757e+00 -2.59011090e-01 3.62531573e-01 1.07205129e+00
1.47142649e+00 3.58256727e-01 -1.10682857e+00 -4.65697676e-01
-2.74385870e-01 -3.32314402e-01 -7.60936365e-02 -6.05706811e-01
-5.75020432e-01 1.21291447e+00 8.62748861e-01 -8.92317444e-02
9.26762044e-01 -3.19517910e-01 1.16486037e+00 1.10667460e-01
1.82465836e-01 -1.19675243e+00 1.95599571e-01 3.47910732e-01
1.28516996e+00 -1.12382853e+00 -4.88589257e-01 -2.79107183e-01
-9.48110342e-01 9.68647599e-01 4.65502441e-01 1.00894883e-01
5.90469301e-01 2.03579798e-01 3.29560786e-01 7.15574682e-01
-7.28041470e-01 2.00060382e-01 3.58196229e-01 7.03032076e-01
2.60317862e-01 2.60038511e-03 2.31678650e-01 7.06369400e-01
-4.60769355e-01 -3.24076593e-01 2.49406010e-01 5.58433831e-01
-5.78200281e-01 -1.41048348e+00 -6.30683541e-01 -7.34669790e-02
-4.66110617e-01 -2.34790176e-01 -2.63307691e-01 2.54300714e-01
-1.75349489e-01 8.52733254e-01 -6.33106083e-02 -3.78343016e-01
3.11380446e-01 7.55731463e-01 4.03522909e-01 -7.92927563e-01
-7.92662203e-01 3.52535486e-01 -6.73365518e-02 5.96878603e-02
3.64894606e-02 -6.35890722e-01 -8.67874265e-01 -1.78663686e-01
-2.83293217e-01 1.18155867e-01 7.20467925e-01 8.39139581e-01
2.13566035e-01 5.98309457e-01 1.17755091e+00 -9.01316047e-01
-9.63058710e-01 -1.21796072e+00 -7.67134607e-01 2.82687783e-01
5.45662880e-01 -3.17191094e-01 -5.15689969e-01 2.69201994e-01] | [14.929698944091797, 6.614534854888916] |
c1b7cb3d-d19d-49ab-8c5d-1ab4daa102a7 | safe-exploration-in-linear-equality | null | null | https://openreview.net/forum?id=5vjyt5JHmaU | https://openreview.net/pdf?id=5vjyt5JHmaU | Safe Exploration in Linear Equality Constraint | With the extensive research and application, some shortcomings of reinforcement learning methods are gradually revealed. One of the considerable problems is that it is difficult for reinforcement learning methods to strictly satisfy the constraints. In this paper, a Singular Value Decomposition-based non-training method called 'Action Decomposition Regular' is proposed to achieve safe exploration. By adopting linear dynamics model, our method decomposes the action space into a constraint dimension and a free dimension for separate control, making policy strictly satisfy the linear equality constraint without limiting the exploration region. In addition, we show how our method should be used when the action space is limited and convex, which makes the method more suitable for real-world scenarios. Finally, we show the effectiveness of our method in a physically-based environment and prevail where reward shaping fails. | ['Jinwei Liu', 'Wang Yao', 'Zijia Niu', 'Xiaohu Jia'] | 2021-09-29 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-6.58722371e-02 1.34189636e-01 -5.65103590e-01 2.96522379e-01
-1.26738427e-02 -3.92778248e-01 2.57577360e-01 -3.76951396e-01
-5.36739349e-01 1.35181999e+00 3.76760960e-02 -3.30190569e-01
-4.15642083e-01 -7.32397199e-01 -4.01386440e-01 -1.10029233e+00
-1.48127764e-01 -9.66542438e-02 2.24208593e-01 -6.86394393e-01
4.46890354e-01 2.95239568e-01 -1.37121809e+00 -5.19607902e-01
1.22762704e+00 7.97611415e-01 4.45516288e-01 8.37547109e-02
1.32577628e-01 5.54891467e-01 -4.23236370e-01 4.37694728e-01
5.65634429e-01 -5.85660100e-01 -5.93608499e-01 3.52106988e-01
-5.09433627e-01 -5.61816692e-01 -2.77683675e-01 1.15839696e+00
6.09565437e-01 6.32735848e-01 3.50519896e-01 -1.14637876e+00
-4.41417605e-01 3.09258789e-01 -7.76882291e-01 1.60416976e-01
1.32016331e-01 4.18670326e-01 6.14944756e-01 -1.94201082e-01
4.03487623e-01 1.23501658e+00 3.21971148e-01 7.75669396e-01
-9.31838155e-01 -6.07730925e-01 7.11298168e-01 1.09507322e-01
-9.95549500e-01 -9.61839929e-02 7.56934643e-01 -1.28305271e-01
7.77626991e-01 1.74795240e-01 1.01222491e+00 1.02381182e+00
4.11524296e-01 7.19685316e-01 1.47709107e+00 -2.62336105e-01
7.96894670e-01 -1.43711448e-01 -4.42811012e-01 4.83438492e-01
3.34445864e-01 6.86884880e-01 -3.42390649e-02 -1.74279325e-02
1.29684508e+00 -4.16711085e-02 -3.29887927e-01 -7.41673589e-01
-1.14359665e+00 9.86785114e-01 1.77364320e-01 1.24123976e-01
-4.23386484e-01 1.05386518e-01 2.56499559e-01 2.74980634e-01
1.99409291e-01 5.78682899e-01 -4.29353327e-01 -4.70204175e-01
-4.76048380e-01 6.73435867e-01 6.28647685e-01 7.31067002e-01
2.85550535e-01 6.41035438e-01 -6.44639283e-02 6.85094059e-01
4.00990337e-01 3.11424822e-01 6.33635700e-01 -1.16066408e+00
5.42911351e-01 4.98922169e-01 6.08188927e-01 -8.49136353e-01
-2.77998298e-01 -5.42924404e-01 -7.29784846e-01 8.76498044e-01
2.93048173e-01 -6.40090227e-01 -6.07278347e-01 1.71740222e+00
6.18807852e-01 1.00201227e-01 2.49967933e-01 1.18884873e+00
1.18426606e-01 5.83582222e-01 -7.61960521e-02 -8.30402672e-01
9.02138412e-01 -8.41798663e-01 -9.97037053e-01 -1.64840132e-01
4.01172072e-01 -2.50426441e-01 1.18354058e+00 6.28643692e-01
-1.02268159e+00 -3.28176200e-01 -1.23989868e+00 7.50714540e-01
-9.69453007e-02 -8.01029205e-02 9.50954735e-01 4.44155991e-01
-4.65198129e-01 9.62010741e-01 -8.74791145e-01 -2.05509424e-01
-6.93939328e-02 4.74250078e-01 -4.82023805e-02 5.90165198e-01
-1.37638628e+00 1.07696092e+00 8.61651123e-01 2.19002530e-01
-1.05467844e+00 -2.02197701e-01 -5.73106289e-01 -1.08360335e-01
7.94356048e-01 -2.64494449e-01 1.27070773e+00 -6.08491540e-01
-2.06036711e+00 -5.87978177e-02 3.85065109e-01 -2.99413145e-01
8.20865810e-01 -1.52795136e-01 -1.75806612e-01 -6.82433844e-02
-6.02888055e-02 1.35015935e-01 8.56565952e-01 -1.23729968e+00
-5.99071801e-01 -1.59899086e-01 4.68588173e-01 7.88337946e-01
-3.47852558e-01 -3.96513671e-01 2.29654349e-02 -6.33442521e-01
8.38653594e-02 -8.81220996e-01 -9.25135791e-01 -3.26576054e-01
4.81374562e-04 -3.63068879e-01 8.63210857e-01 -3.25150818e-01
1.43086791e+00 -2.07249022e+00 4.44567055e-01 1.27939701e-01
-7.56063238e-02 2.50257254e-01 8.08240548e-02 6.24031425e-01
1.22359253e-01 8.82795430e-04 -3.57627779e-01 2.55912215e-01
-9.80175883e-02 7.43638575e-01 -5.08892894e-01 4.83125210e-01
-2.05104113e-01 5.09707093e-01 -1.07357383e+00 -3.97567898e-01
2.42637873e-01 6.56352639e-02 -6.50984168e-01 1.99824065e-01
-3.85579973e-01 9.44715142e-01 -1.12567592e+00 3.89066696e-01
7.39830196e-01 3.39316905e-01 1.35474056e-01 2.69917876e-01
-4.95381266e-01 -1.03346251e-01 -1.77578700e+00 1.49440551e+00
-1.82340577e-01 -2.26903245e-01 4.87675965e-01 -1.40375292e+00
1.00859845e+00 1.08961068e-01 7.78662026e-01 -7.04006672e-01
2.95480251e-01 7.12858140e-02 2.36803740e-02 -7.56587684e-01
1.98709026e-01 -3.54946285e-01 1.03969075e-01 2.87350088e-01
-2.20433995e-01 -3.05692464e-01 2.06031844e-01 -1.52680159e-01
7.29369164e-01 7.07409203e-01 4.84503657e-01 -5.09575605e-01
6.80734158e-01 -7.10521340e-02 1.12188292e+00 5.49825847e-01
-4.87652957e-01 -4.25745845e-02 6.43722057e-01 -2.37798527e-01
-8.63473296e-01 -7.29361176e-01 -2.73736566e-01 8.41005683e-01
5.23909211e-01 -9.05112177e-02 -5.97754836e-01 -7.66510010e-01
2.93336809e-01 5.33249497e-01 -5.72204113e-01 -3.13640773e-01
-6.67274833e-01 -8.80530477e-01 6.31633848e-02 4.86863226e-01
8.53369534e-01 -1.02860904e+00 -8.92064273e-01 3.90974998e-01
1.80299133e-01 -5.53259969e-01 -2.78227657e-01 3.00152123e-01
-1.16700613e+00 -9.34042454e-01 -6.26604676e-01 -4.91010845e-01
5.46852112e-01 1.56463653e-01 3.66351664e-01 3.05636339e-02
3.81543562e-02 2.48702377e-01 -4.11770433e-01 -3.14237863e-01
9.23303701e-03 -3.06709111e-01 5.92025936e-01 -1.94619626e-01
-1.37367576e-01 -7.12529242e-01 -7.11571991e-01 5.18636405e-01
-7.79575109e-01 3.72205339e-02 4.78355020e-01 9.09948051e-01
5.72485447e-01 4.40617383e-01 8.97096038e-01 -1.87325835e-01
1.02279556e+00 -3.80837440e-01 -8.40549350e-01 -2.35563926e-02
-8.81664634e-01 2.54734695e-01 8.46933544e-01 -7.64983773e-01
-1.24399412e+00 -5.68800420e-03 3.76480259e-02 -2.61876076e-01
8.38467628e-02 4.15728956e-01 -1.60189241e-01 -3.58311236e-01
6.39297903e-01 4.79330063e-01 2.73429573e-01 -4.73257840e-01
1.62785500e-01 5.71799994e-01 1.77397504e-02 -8.96785736e-01
7.35915363e-01 2.94632196e-01 3.52327645e-01 -5.62652707e-01
-5.62630475e-01 -1.16355374e-01 -3.56709063e-01 -3.45718205e-01
5.61456740e-01 -5.65021753e-01 -1.19634080e+00 9.88522470e-02
-4.96723354e-01 -4.56988722e-01 -3.44713032e-01 8.32667649e-01
-9.81087685e-01 6.53226972e-01 -4.37127113e-01 -1.31700671e+00
-4.33439612e-02 -1.05495882e+00 3.05640221e-01 5.46207726e-01
4.32539076e-01 -8.80901456e-01 3.97192001e-01 -3.60513031e-02
3.54084402e-01 3.97736192e-01 4.95458603e-01 -6.71677142e-02
-3.30138057e-01 3.78769159e-01 1.56448916e-01 4.04306799e-01
9.79411677e-02 -2.90104806e-01 -2.98825204e-01 -6.98307395e-01
4.71562237e-01 -5.28282166e-01 6.54161632e-01 3.25849831e-01
1.18617463e+00 -4.50687498e-01 -3.03262174e-01 4.87903029e-01
1.47101307e+00 8.60085666e-01 4.91252720e-01 7.29391217e-01
2.97267079e-01 4.38063622e-01 1.12347770e+00 9.37968612e-01
-1.09179690e-03 4.70999807e-01 8.14519703e-01 8.52883309e-02
5.19382834e-01 -4.00041044e-01 3.68514717e-01 6.29625201e-01
-5.06051302e-01 1.47918269e-01 -4.88194525e-01 3.28656733e-01
-2.40823698e+00 -1.00716078e+00 2.12048143e-01 2.33478975e+00
9.21433270e-01 1.32477731e-01 2.30794102e-01 -1.07466385e-01
4.43493307e-01 6.93389550e-02 -8.75876427e-01 -5.58064938e-01
1.81163400e-01 -2.42713019e-01 6.11694217e-01 5.15191197e-01
-1.06538868e+00 8.25643897e-01 6.91474676e+00 1.02148497e+00
-1.02539313e+00 -2.35140771e-01 2.25421667e-01 -1.26238735e-02
-1.28048390e-01 2.82849520e-01 -7.39007115e-01 6.69860542e-01
4.73962128e-01 -2.32315615e-01 7.03746200e-01 1.08988595e+00
7.93368876e-01 -4.28877354e-01 -4.36903477e-01 7.59691298e-01
-4.51595694e-01 -7.38552213e-01 -3.24007750e-01 2.80250460e-01
6.40706539e-01 -5.53619564e-01 -1.11212516e-02 6.27288043e-01
3.46763134e-01 -9.36760426e-01 4.13885027e-01 3.29948932e-01
4.88612235e-01 -1.09470284e+00 4.33176816e-01 8.00326049e-01
-1.04770434e+00 -5.29451072e-01 -7.07188487e-01 -7.30406642e-01
1.12766907e-01 2.76339561e-01 -2.54493326e-01 6.15779817e-01
3.94668162e-01 6.95232749e-01 -7.11225495e-02 1.07933390e+00
-2.95465112e-01 4.16631162e-01 -3.53917778e-01 -3.89640182e-01
6.36766911e-01 -8.09044242e-01 7.17053652e-01 5.10406137e-01
1.85182571e-01 4.03918564e-01 7.01754868e-01 7.71403313e-01
7.46957004e-01 8.67275447e-02 -8.60445797e-01 -6.97830468e-02
1.85864449e-01 1.05052710e+00 -6.94221675e-01 6.27418533e-02
-1.23663977e-01 6.84450924e-01 3.30602050e-01 5.61544001e-01
-1.06754124e+00 -4.89068896e-01 5.13022542e-01 -2.39327237e-01
1.95741266e-01 -5.36508858e-01 -2.88564056e-01 -1.36483598e+00
1.13856117e-03 -8.50170910e-01 1.66152149e-01 -1.65245771e-01
-9.42615390e-01 1.96450621e-01 2.39360929e-01 -1.65041518e+00
-3.45614970e-01 -5.36325276e-01 -6.61562622e-01 6.48042083e-01
-1.22881699e+00 -6.35089695e-01 2.06183389e-01 6.61818504e-01
4.90895569e-01 -3.96231204e-01 5.66111803e-01 -1.95973851e-02
-7.86441505e-01 1.44120589e-01 6.35748506e-01 -3.61781359e-01
2.24284813e-01 -1.37651157e+00 -3.51567775e-01 6.02716446e-01
-5.70101798e-01 5.18971801e-01 1.07895815e+00 -7.91206419e-01
-1.46704817e+00 -5.73406339e-01 -9.92444344e-03 9.73768383e-02
7.03719795e-01 -2.00358436e-01 -8.09089065e-01 3.42012823e-01
2.53418356e-01 -2.16715500e-01 1.61280051e-01 1.51489722e-02
4.36613560e-01 -1.60314724e-01 -1.10565901e+00 9.38098609e-01
1.01800990e+00 2.02946812e-01 -6.47778809e-01 3.83542597e-01
6.80353343e-01 -5.53561211e-01 -7.97662616e-01 6.50923908e-01
3.48152488e-01 -6.88449979e-01 8.41092944e-01 -9.10917878e-01
1.52290076e-01 -5.14584899e-01 9.19740051e-02 -1.54275823e+00
-4.48999792e-01 -9.24955368e-01 -3.52062851e-01 8.23403001e-01
-1.04588822e-01 -7.67602801e-01 8.62031162e-01 4.44559753e-01
-3.34067866e-02 -1.27348840e+00 -1.20127523e+00 -1.29730189e+00
3.66157055e-01 4.19966727e-02 3.29761565e-01 9.50231194e-01
6.54194891e-01 2.79311351e-02 -6.97537482e-01 -1.03813887e-01
4.97310102e-01 1.44197434e-01 4.63551611e-01 -7.15920210e-01
-5.88101208e-01 -3.55297416e-01 7.72385001e-02 -1.17228782e+00
5.82991168e-02 -2.82859981e-01 8.60974789e-02 -1.49920428e+00
-4.49813791e-02 -5.17593503e-01 -4.45760816e-01 4.20445919e-01
-5.34941442e-02 -5.93577445e-01 1.45175889e-01 2.35016152e-01
-4.24854368e-01 1.29269290e+00 2.00004053e+00 1.52793691e-01
-7.07181811e-01 2.13314682e-01 -5.32764733e-01 6.32140994e-01
1.12022293e+00 -7.18814358e-02 -9.26255405e-01 -2.93344557e-02
2.33635139e-02 4.14901555e-01 1.71601474e-02 -9.17671978e-01
-1.80001289e-01 -1.13722587e+00 1.19160376e-01 -3.31034362e-01
1.98703408e-01 -8.99251580e-01 -1.73811167e-01 8.46106708e-01
-1.57720178e-01 -1.66659765e-02 1.40262574e-01 8.10113609e-01
-9.12056398e-03 -4.05237049e-01 7.90944159e-01 -2.97851652e-01
-6.35228455e-01 1.83526129e-01 -4.12162215e-01 6.04453310e-02
1.42753816e+00 -2.53191054e-01 6.15826920e-02 -3.42710614e-01
-6.96136653e-01 6.54509306e-01 2.49085233e-01 1.41561821e-01
5.78633487e-01 -1.49113584e+00 -3.61674249e-01 8.71046185e-02
-4.00287807e-01 -3.64985242e-02 2.81417757e-01 8.50307822e-01
-3.22249919e-01 4.18143928e-01 -4.81899321e-01 -1.65582821e-01
-7.53151059e-01 8.72260928e-01 4.16815668e-01 -3.08922201e-01
-8.43764484e-01 3.06652784e-01 1.81816205e-01 -4.28207457e-01
3.81494433e-01 -3.00739795e-01 -5.15162826e-01 -2.22035691e-01
3.97581398e-01 4.64363813e-01 -4.93734926e-01 -1.32145941e-01
-1.46453112e-01 4.62332636e-01 1.73112571e-01 -4.19046223e-01
1.27475631e+00 -1.69379085e-01 1.39475733e-01 1.41458541e-01
6.03054583e-01 -1.59786895e-01 -1.80094874e+00 1.59008130e-01
-3.37660044e-01 -5.95467210e-01 2.65108317e-01 -6.41301513e-01
-9.01217818e-01 6.83519840e-01 5.71563900e-01 3.12960684e-01
1.10823798e+00 -7.84935117e-01 5.14517665e-01 5.69947362e-01
6.32648408e-01 -1.70209908e+00 1.66739434e-01 6.44607067e-01
1.10519063e+00 -1.06861937e+00 2.11254433e-01 -1.64550275e-01
-9.70352829e-01 1.08657622e+00 1.19913137e+00 -4.52046156e-01
4.75373566e-01 7.61781037e-02 -3.47931206e-01 1.64309517e-01
-6.40291274e-01 -2.21703082e-01 -1.29540756e-01 6.40042663e-01
2.88221501e-02 -2.31513921e-02 -1.11488938e+00 6.75301254e-01
1.19871631e-01 2.22352389e-02 5.26677489e-01 1.13837028e+00
-8.73960674e-01 -1.23044610e+00 -5.77854991e-01 2.29300663e-01
-2.95241177e-01 3.64662349e-01 5.04468158e-02 1.02516484e+00
1.48149818e-01 8.34911704e-01 -4.15120989e-01 -1.88131571e-01
3.87378067e-01 -2.74616957e-01 4.65804964e-01 -4.55390841e-01
-2.07001522e-01 3.06622088e-01 -1.02568671e-01 -7.41212845e-01
-2.83699363e-01 -6.53041422e-01 -1.56847799e+00 -1.67903379e-01
-3.95775467e-01 3.91308397e-01 2.66773522e-01 9.91312623e-01
4.56961393e-02 5.44318080e-01 8.44225883e-01 -5.19383430e-01
-1.32365954e+00 -8.49277318e-01 -8.81399512e-01 6.02735355e-02
2.68100560e-01 -1.33306527e+00 -4.35879409e-01 -4.65077013e-01] | [4.291292190551758, 2.15521502494812] |
b39dc31b-1996-4d5d-bec8-8d892e9bb2cc | perceiving-the-world-question-guided | 2204.09597 | null | https://arxiv.org/abs/2204.09597v2 | https://arxiv.org/pdf/2204.09597v2.pdf | Perceiving the World: Question-guided Reinforcement Learning for Text-based Games | Text-based games provide an interactive way to study natural language processing. While deep reinforcement learning has shown effectiveness in developing the game playing agent, the low sample efficiency and the large action space remain to be the two major challenges that hinder the DRL from being applied in the real world. In this paper, we address the challenges by introducing world-perceiving modules, which automatically decompose tasks and prune actions by answering questions about the environment. We then propose a two-phase training framework to decouple language learning from reinforcement learning, which further improves the sample efficiency. The experimental results show that the proposed method significantly improves the performance and sample efficiency. Besides, it shows robustness against compound error and limited pre-training data. | ['Chengqi Zhang', 'Joey Tianyi Zhou', 'Yali Du', 'Ling Chen', 'Meng Fang', 'Yunqiu Xu'] | 2022-03-20 | null | https://aclanthology.org/2022.acl-long.41 | https://aclanthology.org/2022.acl-long.41.pdf | acl-2022-5 | ['text-based-games'] | ['playing-games'] | [-6.40298799e-02 -1.27686873e-01 5.19077405e-02 -1.38596511e-02
-6.66970372e-01 -5.60161233e-01 5.91352761e-01 -9.48412642e-02
-9.10914302e-01 6.60173595e-01 -9.00482107e-03 -4.93348420e-01
-2.86644381e-02 -1.11874676e+00 -3.62147689e-01 -4.94950533e-01
1.21363133e-01 3.65941554e-01 5.08940399e-01 -4.12523091e-01
3.04712206e-01 1.74707934e-01 -1.64779449e+00 8.76965076e-02
1.11646438e+00 7.32314110e-01 6.14732921e-01 6.97767794e-01
-3.22645158e-01 1.35011125e+00 -6.62104070e-01 -9.64493230e-02
2.42576659e-01 -4.79508430e-01 -7.51205564e-01 -1.51995644e-01
-4.35698986e-01 -6.14288151e-01 -2.96472877e-01 1.13451147e+00
5.11886716e-01 4.47179675e-01 1.11363098e-01 -1.10980892e+00
-3.37537706e-01 5.24412513e-01 -3.50897014e-01 8.28811601e-02
4.46998864e-01 5.11738002e-01 9.58582401e-01 -5.87608635e-01
3.27601433e-01 1.38770032e+00 1.34323984e-01 7.83058167e-01
-7.37982750e-01 -7.52186179e-01 4.73545134e-01 3.27963024e-01
-1.13008666e+00 -3.11633855e-01 6.81059659e-01 -2.43422344e-01
1.26103818e+00 4.28923452e-03 6.29674137e-01 9.55092132e-01
1.23613015e-01 1.21819055e+00 1.13520312e+00 -4.68796611e-01
6.18378937e-01 -7.32235610e-02 -1.95939735e-01 8.58163238e-01
9.29640457e-02 3.27552855e-01 -4.20763969e-01 1.44154355e-01
8.86029363e-01 -8.03890601e-02 8.48851353e-02 -3.36632818e-01
-9.35914576e-01 9.50653374e-01 1.15532272e-01 1.94735318e-01
-2.63569981e-01 1.61051586e-01 5.97091019e-01 2.91826427e-01
3.05303544e-01 6.47718132e-01 -5.52367508e-01 -5.75305581e-01
-4.75544810e-01 3.68298322e-01 7.04673231e-01 8.47000837e-01
6.23481095e-01 1.89313188e-01 -9.98079777e-03 8.09359610e-01
4.02695477e-01 6.51789129e-01 6.15608692e-01 -8.70815754e-01
6.14638329e-01 9.13709044e-01 2.29439050e-01 -8.77529383e-01
-5.16679525e-01 -5.53678349e-02 -5.37384868e-01 3.12299192e-01
4.16589022e-01 -4.20227736e-01 -5.92972636e-01 1.81287920e+00
3.29568356e-01 1.61545724e-01 2.84224510e-01 8.22630048e-01
7.03914225e-01 8.11559975e-01 3.65991652e-01 -2.00338349e-01
1.43311429e+00 -1.04201126e+00 -7.88004756e-01 -8.45727921e-01
6.26871705e-01 -3.23437840e-01 1.55151999e+00 5.44468522e-01
-1.05175531e+00 -6.62915707e-01 -1.01397789e+00 -3.27555798e-02
-3.72663766e-01 2.57603377e-02 7.66212046e-01 6.45791471e-01
-8.12180638e-01 2.11130008e-01 -1.00541162e+00 -1.74567893e-01
3.20681781e-01 4.15028572e-01 -4.51456495e-02 5.74681796e-02
-1.49866760e+00 8.20967495e-01 8.12036991e-01 -6.50356337e-02
-8.75984609e-01 -2.90552586e-01 -1.08009696e+00 3.04561496e-01
8.75246465e-01 -3.24363679e-01 1.59688580e+00 -5.38907170e-01
-2.21200609e+00 4.99822855e-01 7.13952184e-02 -3.63057166e-01
5.54624081e-01 -2.72292972e-01 -2.37335682e-01 8.62510204e-02
5.16405664e-02 5.83551228e-01 4.69701380e-01 -8.69784176e-01
-1.03816450e+00 -2.04815403e-01 6.21289790e-01 5.18848658e-01
-3.43940288e-01 3.01658940e-02 -6.87882543e-01 -3.53163868e-01
-2.25333646e-01 -5.84015191e-01 -6.04379892e-01 -3.26243490e-01
1.65002555e-01 -4.98014778e-01 3.50303233e-01 -4.53903258e-01
1.25546968e+00 -1.98246825e+00 -1.13958307e-01 -7.54686743e-02
1.87975198e-01 4.64423448e-01 -3.97016138e-01 3.92487556e-01
3.61559570e-01 8.48150533e-03 3.12957391e-02 -1.08590610e-01
7.71913771e-03 3.81481528e-01 -2.52641171e-01 -4.26000692e-02
3.44065040e-01 1.09935772e+00 -1.36187184e+00 -4.64463472e-01
4.14401382e-01 -1.18932605e-01 -8.27312350e-01 5.19950151e-01
-5.29281557e-01 3.92167509e-01 -6.98502541e-01 1.60265386e-01
5.57406783e-01 -4.86749969e-02 3.00743997e-01 4.53583807e-01
-2.26320457e-02 5.27741253e-01 -1.39405525e+00 1.80622971e+00
-7.05956280e-01 2.29715943e-01 4.02115397e-02 -1.08865905e+00
9.39094245e-01 1.22704335e-01 1.52505741e-01 -1.39073467e+00
1.84206769e-01 -8.71158466e-02 2.17210963e-01 -8.98006499e-01
5.10880947e-01 -2.91382283e-01 -4.09154981e-01 6.42902792e-01
-1.44495033e-02 -1.78196222e-01 5.05343318e-01 -4.88686226e-02
1.04069245e+00 4.37440962e-01 6.90740824e-01 -1.08619981e-01
7.05533803e-01 3.24789584e-02 8.52269053e-01 9.00295317e-01
-4.28764611e-01 -8.06390215e-03 6.85965836e-01 -4.09935266e-01
-6.78955495e-01 -9.03694272e-01 6.06720269e-01 1.50466669e+00
2.97527105e-01 -4.60112780e-01 -7.97450066e-01 -7.50945628e-01
-5.81949234e-01 8.58232379e-01 -2.91957974e-01 -4.02396053e-01
-6.39098048e-01 -7.18211055e-01 4.54161286e-01 6.37363493e-01
8.56253088e-01 -1.54202497e+00 -1.05690134e+00 3.68950278e-01
-3.01913708e-01 -1.17832160e+00 -1.66462034e-01 1.31272823e-01
-6.24188244e-01 -1.11554027e+00 -7.73559660e-02 -9.03729081e-01
4.55399752e-01 3.37444454e-01 1.09773529e+00 5.23406044e-02
-5.57834804e-02 2.81230718e-01 -4.02360737e-01 -5.23789406e-01
-4.64941770e-01 1.22354805e-01 3.50662656e-02 -4.17611986e-01
5.97779036e-01 -3.71831834e-01 -4.86422598e-01 1.32012457e-01
-8.38456690e-01 3.45967174e-01 6.23316884e-01 8.53240371e-01
3.07487488e-01 4.44442421e-01 6.56686246e-01 -1.02983832e+00
1.10702872e+00 -2.53827751e-01 -8.54264677e-01 1.87358245e-01
-5.48745632e-01 3.00006360e-01 9.91899490e-01 -4.43470985e-01
-1.21405196e+00 1.12091757e-01 -2.21994132e-01 2.01350898e-01
-3.63673836e-01 5.96340537e-01 -4.27294016e-01 3.59389752e-01
5.18848479e-01 5.08257210e-01 4.96673398e-02 -2.05964684e-01
2.59611428e-01 5.55591226e-01 2.43876651e-01 -8.79391491e-01
7.89194047e-01 1.60863280e-01 -4.66551691e-01 -6.76042259e-01
-7.31025517e-01 -3.96499008e-01 -2.26206407e-01 -6.56979978e-02
8.23449314e-01 -1.01318192e+00 -1.11518717e+00 3.44079167e-01
-9.58364904e-01 -6.62365198e-01 -1.94153920e-01 2.76807696e-01
-8.18169177e-01 3.48537505e-01 -5.95539749e-01 -1.01041949e+00
-4.08828378e-01 -1.34774828e+00 7.35201001e-01 4.86381739e-01
-1.06875129e-01 -7.36429632e-01 2.05244556e-01 3.50084543e-01
3.21940899e-01 -3.08457673e-01 7.43174851e-01 -6.05568051e-01
-5.24828434e-01 -3.17482799e-02 -1.17163159e-01 1.77159667e-01
1.81586131e-01 -4.42440867e-01 -8.26911092e-01 -2.20036805e-01
1.90780059e-01 -5.42770088e-01 3.71177047e-01 5.83137386e-03
1.24109161e+00 -2.38187343e-01 9.02426168e-02 2.92349219e-01
1.25616848e+00 6.73215270e-01 6.61177635e-01 6.58208609e-01
4.06099916e-01 5.44450760e-01 8.80346954e-01 5.82441151e-01
4.52920467e-01 3.73821139e-01 3.35695446e-01 -8.90247673e-02
2.65202492e-01 -5.73809385e-01 7.02405870e-01 9.05834138e-01
5.46564087e-02 -1.71784297e-01 -1.06161165e+00 3.98711711e-01
-2.14221191e+00 -1.01169860e+00 4.25348580e-01 1.86208725e+00
8.36172938e-01 4.53661978e-01 1.53453991e-01 2.84178788e-03
2.79855281e-01 1.29813910e-01 -7.76158452e-01 -3.84098798e-01
2.24293232e-01 5.01170576e-01 1.10489078e-01 7.99874485e-01
-1.11997163e+00 1.50395036e+00 6.15498447e+00 9.24992681e-01
-9.90355611e-01 -1.15666455e-02 3.03152174e-01 4.72603068e-02
4.63127438e-03 -2.96213031e-01 -4.81588066e-01 3.34957510e-01
9.79048610e-01 -2.91096836e-01 6.67407930e-01 9.40390050e-01
5.57069778e-01 -3.25177282e-01 -8.83149147e-01 8.91856134e-01
-1.98484078e-01 -9.39046800e-01 1.04687907e-01 -1.62454098e-01
2.95843065e-01 6.49336949e-02 -1.60189137e-01 8.93778324e-01
7.70347118e-01 -8.85047197e-01 6.58303261e-01 1.81380078e-01
4.94215548e-01 -9.79246497e-01 6.25790954e-01 9.36369061e-01
-1.24495244e+00 -2.44692862e-01 -5.26429594e-01 -7.61552453e-01
-2.55259633e-01 1.24736615e-02 -8.90489221e-01 3.79265517e-01
4.82476443e-01 4.71204847e-01 -4.90182996e-01 8.05443883e-01
-7.41968989e-01 5.88684440e-01 -1.23847097e-01 -7.06158817e-01
3.11754763e-01 -4.01417196e-01 1.63802251e-01 1.03832912e+00
5.45059294e-02 3.11221063e-01 5.68528533e-01 7.71683693e-01
2.23316029e-01 9.44823548e-02 -6.46135449e-01 -6.39901310e-02
3.58031899e-01 1.12286878e+00 -8.29820693e-01 -1.80881664e-01
-6.78322673e-01 7.79769838e-01 5.58438540e-01 3.75662535e-01
-8.68360937e-01 -4.44714069e-01 8.41417074e-01 -9.03719664e-02
1.29037023e-01 -3.66046011e-01 -2.76672721e-01 -1.34053397e+00
-9.94176865e-02 -1.30829108e+00 2.89876550e-01 -5.39134681e-01
-9.51165378e-01 3.60407740e-01 -1.87557489e-01 -1.17012942e+00
-3.84862721e-01 -9.11680639e-01 -5.18079996e-01 5.94910622e-01
-1.44199896e+00 -7.42302179e-01 -6.13351986e-02 4.96751249e-01
8.85708451e-01 -3.44581574e-01 9.63648736e-01 9.85696465e-02
-8.54859591e-01 4.53119516e-01 -1.21028330e-02 3.19382966e-01
2.82770485e-01 -1.22597671e+00 4.31243479e-01 1.00225139e+00
6.30152524e-02 6.28031909e-01 5.12234390e-01 -3.48195553e-01
-1.43139446e+00 -8.78038883e-01 4.68650132e-01 -3.04000258e-01
7.06915736e-01 -8.46055329e-01 -6.20119452e-01 3.43409300e-01
1.20730571e-01 -4.03982610e-01 5.87375402e-01 1.91272900e-01
-1.16843462e-01 -1.27903491e-01 -1.00390482e+00 1.07240021e+00
9.50641990e-01 -5.55936754e-01 -8.74845028e-01 -2.27014422e-02
6.91254795e-01 -4.44435537e-01 -2.58331418e-01 1.22364247e-04
1.47680342e-01 -7.82740951e-01 8.33808184e-01 -6.97787464e-01
5.90676904e-01 -3.46447110e-01 1.61996037e-01 -1.30048501e+00
-3.79714280e-01 -5.73558867e-01 -3.93890031e-02 1.06380343e+00
3.44553739e-01 -6.36050344e-01 9.78201687e-01 6.36356175e-01
2.30679870e-01 -6.05145097e-01 -7.70578861e-01 -6.29687607e-01
1.80980027e-01 -7.38759279e-01 6.26958847e-01 6.66140139e-01
3.63791257e-01 7.03565717e-01 -3.53888303e-01 3.02958619e-02
2.22099572e-01 2.19644755e-01 8.36890578e-01 -1.07535136e+00
-5.33650219e-01 -4.86097306e-01 2.11714767e-02 -1.33623803e+00
2.05479518e-01 -6.82331324e-01 4.23338592e-01 -1.54193759e+00
2.28784084e-01 -3.17590028e-01 -4.22222674e-01 6.38657212e-01
-5.25910914e-01 -3.65729898e-01 1.76041439e-01 -1.79738760e-01
-1.09055424e+00 7.35012472e-01 1.38094521e+00 -1.43012315e-01
-3.66742283e-01 -3.49891484e-02 -9.22359943e-01 7.68804610e-01
1.25312781e+00 -3.30234081e-01 -9.58560288e-01 -6.03862941e-01
3.92652869e-01 7.21697360e-02 -2.09611312e-01 -1.10812593e+00
3.35621089e-01 -6.41700327e-01 7.34188929e-02 -3.17267537e-01
1.11265235e-01 -8.11952829e-01 -5.75949430e-01 7.81337798e-01
-4.04759675e-01 1.68400928e-01 5.48160315e-01 4.96582031e-01
-1.43339023e-01 -2.77594566e-01 6.34881914e-01 -3.40917826e-01
-1.00812149e+00 2.39969596e-01 -8.77777755e-01 2.63448417e-01
1.11506438e+00 -1.00919820e-01 -6.77626673e-03 -4.50675637e-01
-3.31203759e-01 6.95226192e-01 5.21774553e-02 5.94454527e-01
6.20771945e-01 -9.58991468e-01 -5.22801757e-01 4.66245979e-01
6.91939592e-02 1.80057183e-01 1.03699595e-01 1.76451728e-01
-6.09654069e-01 3.65903407e-01 -3.15407366e-01 -1.43122733e-01
-1.07233059e+00 6.68800771e-01 3.69880497e-01 -7.70458817e-01
-5.37549675e-01 7.03360558e-01 3.97821754e-01 -8.18873286e-01
3.19706261e-01 -4.24453944e-01 -4.87728596e-01 -1.71058416e-01
8.20296586e-01 1.84023812e-01 -1.18470021e-01 5.44057228e-02
-3.37107420e-01 1.14364438e-01 -2.78011650e-01 -2.14580610e-01
1.39969730e+00 -4.12854962e-02 9.56252310e-03 3.18299443e-01
5.93295455e-01 -1.34957150e-01 -1.20641577e+00 -3.91663373e-01
2.68910378e-01 -1.97065607e-01 2.29030265e-03 -8.26384366e-01
-8.26395214e-01 1.02381349e+00 4.11869556e-01 2.15043887e-01
1.30874193e+00 -3.96491677e-01 6.98384345e-01 7.71194160e-01
5.95630229e-01 -1.40174043e+00 3.01021487e-01 1.01915228e+00
5.62287867e-01 -1.31391490e+00 -2.64469773e-01 -3.21523488e-01
-6.14659011e-01 9.69324589e-01 1.14472699e+00 -1.16413064e-01
3.01077902e-01 4.45212126e-01 1.41752511e-01 7.17022419e-02
-9.06057298e-01 -4.17980075e-01 -2.17963710e-01 6.93038046e-01
3.72030079e-01 3.52345221e-02 -4.53397453e-01 9.65325952e-01
-2.25331768e-01 7.31201470e-02 4.28192466e-01 1.11108696e+00
-6.06782973e-01 -1.30861831e+00 -1.67630658e-01 2.41596535e-01
-1.99191600e-01 -7.11794943e-02 -3.53875875e-01 6.92605436e-01
1.28159091e-01 1.15721452e+00 -7.30245560e-02 -4.75871086e-01
5.63821375e-01 7.28927180e-02 3.76046926e-01 -7.50467360e-01
-5.91641963e-01 -1.84095073e-02 2.66009104e-02 -7.57919550e-01
-1.37962252e-01 -3.53305727e-01 -1.54658508e+00 -1.68845505e-01
-1.68256283e-01 2.54983008e-01 1.06692657e-01 1.09707999e+00
3.60796154e-01 7.09607661e-01 4.51835304e-01 -3.89612764e-01
-7.63142824e-01 -8.16438973e-01 -2.32980564e-01 3.12526137e-01
6.89245239e-02 -5.37336707e-01 2.30430309e-02 -2.73245186e-01] | [3.8535170555114746, 1.5352736711502075] |
891d50e8-64cd-431f-b2e1-55f84ba9c25b | the-first-proven-performance-guarantees-for | 2305.13459 | null | https://arxiv.org/abs/2305.13459v2 | https://arxiv.org/pdf/2305.13459v2.pdf | The First Proven Performance Guarantees for the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) on a Combinatorial Optimization Problem | The Non-dominated Sorting Genetic Algorithm-II (NSGA-II) is one of the most prominent algorithms to solve multi-objective optimization problems. Recently, the first mathematical runtime guarantees have been obtained for this algorithm, however only for synthetic benchmark problems. In this work, we give the first proven performance guarantees for a classic optimization problem, the NP-complete bi-objective minimum spanning tree problem. More specifically, we show that the NSGA-II with population size $N \ge 4((n-1) w_{\max} + 1)$ computes all extremal points of the Pareto front in an expected number of $O(m^2 n w_{\max} \log(n w_{\max}))$ iterations, where $n$ is the number of vertices, $m$ the number of edges, and $w_{\max}$ is the maximum edge weight in the problem instance. This result confirms, via mathematical means, the good performance of the NSGA-II observed empirically. It also shows that mathematical analyses of this algorithm are not only possible for synthetic benchmark problems, but also for more complex combinatorial optimization problems. As a side result, we also obtain a new analysis of the performance of the global SEMO algorithm on the bi-objective minimum spanning tree problem, which improves the previous best result by a factor of $|F|$, the number of extremal points of the Pareto front, a set that can be as large as $n w_{\max}$. The main reason for this improvement is our observation that both multi-objective evolutionary algorithms find the different extremal points in parallel rather than sequentially, as assumed in the previous proofs. | ['Simon Wietheger', 'Yakob Kahane', 'Benjamin Hebras', 'Benjamin Doerr', 'Sacha Cerf'] | 2023-05-22 | null | null | null | null | ['combinatorial-optimization'] | ['methodology'] | [ 4.83320326e-01 7.76615217e-02 5.96661605e-02 1.16958961e-01
-5.21977961e-01 -6.60224915e-01 -3.99968386e-01 4.07356203e-01
-5.98030567e-01 1.02173674e+00 -6.50884926e-01 -3.62349033e-01
-1.05125868e+00 -1.13699162e+00 -7.95533717e-01 -1.07503939e+00
-6.54469967e-01 7.39888728e-01 1.54883951e-01 -3.87813479e-01
4.68153924e-01 4.37322199e-01 -2.06781793e+00 -3.48152190e-01
1.08497310e+00 1.15628469e+00 9.68174487e-02 7.52025664e-01
-1.37386695e-01 -1.80945098e-01 -6.87529624e-01 -3.91034126e-01
3.81122023e-01 -5.34978390e-01 -8.56142223e-01 -2.49204226e-03
-1.08854108e-01 2.69081712e-01 3.32074314e-01 1.29317403e+00
4.68955070e-01 2.64464468e-01 3.16270828e-01 -1.60349250e+00
2.46264925e-03 7.25090086e-01 -9.96934414e-01 -1.24494366e-01
1.23571791e-01 3.03435400e-02 1.10962808e+00 -9.74944755e-02
6.39311254e-01 9.44019377e-01 2.98694134e-01 4.01083052e-01
-1.07491708e+00 -7.03351617e-01 7.52924159e-02 1.27309129e-01
-1.43305838e+00 3.97459827e-02 3.65972161e-01 4.86038812e-03
9.13727403e-01 1.00007534e+00 6.42777503e-01 -7.34702591e-03
4.26447727e-02 4.79663968e-01 8.75007927e-01 -6.96754336e-01
4.01510686e-01 -2.82332450e-01 5.46734743e-02 7.90667057e-01
1.03504288e+00 1.18953750e-01 -2.72871464e-01 -2.07485124e-01
1.92221683e-02 -3.98000449e-01 -3.73282760e-01 -2.58465439e-01
-6.56011164e-01 8.60166013e-01 9.62762609e-02 5.76348484e-01
-1.59938589e-01 4.87269849e-01 4.04198617e-02 2.95027763e-01
4.22848277e-02 7.18307972e-01 -6.76598012e-01 -3.24689150e-01
-9.39751804e-01 1.63622811e-01 7.97786534e-01 7.01243341e-01
7.38498569e-01 -3.44915651e-02 3.31352562e-01 3.51996779e-01
5.03913462e-02 6.72066152e-01 3.96553911e-02 -1.04345167e+00
6.04825795e-01 8.98458958e-01 2.27996483e-01 -9.89722729e-01
-4.97750938e-01 -7.92744994e-01 -7.60402977e-01 4.61130679e-01
7.97579229e-01 -3.35831940e-01 -3.77908677e-01 1.87285423e+00
2.93088824e-01 -3.98543477e-01 -2.29457825e-01 5.17911077e-01
-9.31696221e-03 5.92019856e-01 -4.01802808e-01 -8.39515507e-01
1.26991308e+00 -6.59017920e-01 -4.04730946e-01 -7.25480169e-02
8.52634609e-01 -6.02736950e-01 6.27013028e-01 4.96710241e-01
-1.35599279e+00 3.71841714e-02 -1.06783926e+00 7.96224356e-01
-3.72543931e-01 -2.17267841e-01 7.79693544e-01 1.24409699e+00
-8.63119483e-01 5.93732893e-01 -5.76775849e-01 -1.17227070e-01
1.80355936e-01 8.43946874e-01 -9.33459029e-02 -1.94897026e-01
-5.71842074e-01 4.19670731e-01 4.86478388e-01 2.02247277e-01
-1.90078169e-01 -5.30166805e-01 -5.31713963e-01 4.42847580e-01
9.31800544e-01 -5.64127743e-01 4.51848894e-01 -8.77552569e-01
-1.10274827e+00 7.06897795e-01 -4.55168724e-01 -1.60461381e-01
4.03964460e-01 5.75386941e-01 -5.38424626e-02 3.34894471e-02
-2.06646279e-01 1.13303363e-01 1.54910251e-01 -1.26934898e+00
-9.47475016e-01 -8.71175706e-01 1.74298495e-01 -1.89447209e-01
-3.16274017e-01 1.47515789e-01 -1.95702121e-01 -1.18034147e-01
1.41504869e-01 -9.62538421e-01 -5.84068894e-01 -4.58302259e-01
-3.54934901e-01 -1.43525451e-01 1.74402937e-01 -1.68023303e-01
1.52734685e+00 -1.85050023e+00 3.98661107e-01 9.05113399e-01
4.28127907e-02 6.46940917e-02 -1.41904041e-01 5.18509150e-01
6.49268478e-02 3.84159029e-01 -5.44577181e-01 8.30600485e-02
2.43691832e-01 1.80655956e-01 3.08655232e-01 5.06940305e-01
-4.49209511e-01 5.95447123e-01 -7.16614723e-01 -1.99876949e-01
-2.19691068e-01 -1.00336090e-01 -5.62597454e-01 -4.04013067e-01
-2.59208471e-01 -2.34972268e-01 -3.14026326e-01 5.15579283e-01
1.05278742e+00 -2.46408775e-01 6.28476799e-01 2.05871686e-01
-3.88643980e-01 -4.15841907e-01 -1.79049313e+00 1.43622863e+00
-2.93708712e-01 2.18635812e-01 4.40193713e-01 -1.21462154e+00
5.58355808e-01 -8.82850438e-02 9.64947402e-01 -7.08199739e-01
4.07634109e-01 5.69416523e-01 2.34412089e-01 -2.38675162e-01
4.29188669e-01 -1.76098466e-01 -2.38746628e-01 7.39245117e-01
-3.91628265e-01 2.25380793e-01 9.79562044e-01 -1.93233192e-01
1.29330587e+00 -3.25480342e-01 -5.17145582e-02 -5.45303464e-01
5.96167505e-01 1.42652363e-01 8.26646388e-01 6.61915243e-01
1.12032041e-01 1.41406894e-01 9.04552281e-01 -3.95800732e-02
-6.83841765e-01 -6.84325635e-01 5.70992976e-02 1.03620529e+00
5.22714674e-01 -2.95016795e-01 -8.16764593e-01 -3.43039185e-01
1.16796074e-02 8.42600644e-01 -6.33139789e-01 1.51046840e-02
-7.91283607e-01 -1.39381349e+00 2.44459659e-01 1.76982433e-01
-3.11901607e-03 -6.81507289e-01 -1.04272020e+00 3.52775306e-01
-6.18133992e-02 -7.80358613e-01 -1.21267550e-01 4.92417127e-01
-9.33194757e-01 -1.28432488e+00 -4.18755174e-01 -6.03570402e-01
1.06091273e+00 -1.81761868e-02 8.58844697e-01 4.52703059e-01
-5.92883825e-01 8.37142020e-03 -1.45288959e-01 -3.06386322e-01
-1.57054588e-01 4.44278829e-02 -1.84824675e-01 -2.47617289e-01
-2.29066215e-03 -5.74363232e-01 -4.28056747e-01 5.44324279e-01
-1.07380557e+00 -2.75114834e-01 4.32761997e-01 5.89434445e-01
7.68122792e-01 8.24901104e-01 3.30830097e-01 -5.00597060e-01
3.69739622e-01 -2.66874786e-02 -1.34936714e+00 5.47285676e-01
-7.83154309e-01 4.13662672e-01 6.85060859e-01 -8.13939422e-02
-5.35077214e-01 7.99622610e-02 1.56789683e-02 2.35710442e-01
3.17310750e-01 4.40450341e-01 -3.13740492e-01 -3.49989712e-01
9.51826274e-02 2.39958674e-01 -1.51731133e-01 -3.86160731e-01
3.88354734e-02 2.06844434e-01 2.70036638e-01 -7.41536796e-01
8.38412941e-01 4.98664647e-01 9.18101311e-01 -5.34288287e-01
-3.52545321e-01 -1.82651684e-01 -7.72505924e-02 -1.57220915e-01
3.59021574e-01 7.91310072e-02 -1.76098359e+00 1.66638028e-02
-6.77522540e-01 -8.60194638e-02 -3.42582852e-01 2.64741510e-01
-4.95456219e-01 3.80181879e-01 4.92385067e-02 -1.34244609e+00
-2.79880047e-01 -1.08623779e+00 5.03865421e-01 3.13103318e-01
5.11662737e-02 -6.53101087e-01 -2.60273963e-01 4.07491267e-01
3.65405202e-01 7.01712370e-01 1.35492432e+00 -3.17570210e-01
-6.24734402e-01 -1.63841933e-01 -7.24656433e-02 -1.39772817e-01
-2.09310517e-01 1.40784070e-01 -1.01520918e-01 -4.87578750e-01
-7.42725581e-02 1.58744678e-01 6.03809237e-01 5.06440997e-01
1.10780752e+00 -3.28537166e-01 -5.71012914e-01 4.55619395e-01
2.05183864e+00 4.91743833e-01 4.87742960e-01 4.97106969e-01
-2.59572808e-02 5.62692583e-01 5.75868785e-01 7.30935097e-01
2.35222671e-02 6.79314613e-01 8.84266257e-01 1.40613735e-01
4.18641686e-01 4.33375210e-01 1.48919255e-01 4.56873745e-01
-3.17258656e-01 -8.36570859e-01 -7.57085860e-01 6.12134993e-01
-1.62572205e+00 -7.19487906e-01 -3.31932515e-01 2.71676469e+00
5.50403893e-01 1.80612713e-01 2.90146083e-01 7.04379499e-01
9.33156908e-01 -3.30446064e-01 -2.94172347e-01 -8.85394454e-01
-3.89521182e-01 8.40014994e-01 9.23104167e-01 5.20649910e-01
-3.85962367e-01 2.93357044e-01 4.78951406e+00 1.03333259e+00
-7.87233174e-01 -2.02922136e-01 4.03772205e-01 -8.19958389e-01
-3.76610428e-01 1.08990602e-01 -5.47632337e-01 7.62491822e-01
8.63095403e-01 -6.48040831e-01 7.19751239e-01 4.80530560e-01
1.12437606e-01 -6.27541840e-01 -8.34153295e-01 6.80378437e-01
-2.22846568e-01 -1.13850868e+00 -3.73069882e-01 7.32475102e-01
9.05907810e-01 -6.78150058e-01 5.35683893e-02 -2.47413620e-01
1.91271856e-01 -1.07595110e+00 4.39799637e-01 -1.40802100e-01
5.63547432e-01 -1.51049113e+00 8.05128753e-01 3.48070741e-01
-1.24747956e+00 -4.96116042e-01 -1.98232934e-01 1.90541476e-01
4.09220219e-01 8.52029681e-01 -4.56027389e-02 1.04236233e+00
7.30888665e-01 -6.34718060e-01 -4.93863747e-02 1.48640978e+00
1.65461637e-02 2.15647191e-01 -8.65480065e-01 -3.21563214e-01
3.65563512e-01 -3.30034196e-01 7.15794206e-01 6.56469464e-01
6.82770967e-01 2.37576276e-01 -2.10112363e-01 5.93656540e-01
-7.78221637e-02 3.24071258e-01 9.44679230e-02 -7.17838034e-02
3.30432236e-01 1.03288710e+00 -1.18932748e+00 2.62445152e-01
1.93262205e-01 4.65753347e-01 -2.86945794e-02 -1.44081470e-02
-1.02378404e+00 -9.96095717e-01 7.01338828e-01 -1.53820524e-02
4.93275851e-01 -4.02745277e-01 -5.71682274e-01 -4.69231874e-01
2.70828336e-01 -5.84970295e-01 5.73821664e-01 -1.39120728e-01
-5.40213764e-01 3.32366467e-01 -6.99345171e-02 -6.78707361e-01
-2.00411618e-01 -7.43683279e-01 -3.99898112e-01 5.47933936e-01
-1.27921605e+00 -4.52105224e-01 -1.14096086e-02 1.66370690e-01
-6.37958273e-02 1.53629988e-01 7.31900156e-01 3.07414681e-01
-7.30270147e-01 7.77986526e-01 6.21200383e-01 -4.04365510e-01
-1.51272625e-01 -9.10427690e-01 -2.81609714e-01 9.55539405e-01
-3.34548295e-01 3.68967205e-01 9.93587434e-01 -2.28679344e-01
-1.89437628e+00 -3.88021767e-01 9.14086163e-01 3.03655922e-01
4.53713059e-01 -1.25641331e-01 -2.76376307e-01 1.82497472e-01
6.20596781e-02 -4.47763920e-01 8.43939960e-01 -4.87974361e-02
3.52935195e-01 -3.74594063e-01 -1.46345305e+00 5.65778852e-01
1.24146044e+00 5.33628643e-01 1.52617976e-01 2.45754123e-01
4.10949796e-01 -2.41018146e-01 -9.06218112e-01 5.46278417e-01
5.37328482e-01 -1.04035044e+00 8.92949343e-01 -3.99713546e-01
2.76854873e-01 -4.14693981e-01 -3.07092845e-01 -9.99079347e-01
4.21709828e-02 -9.23821568e-01 -5.75299524e-02 1.02865112e+00
7.14015841e-01 -8.71545553e-01 1.01911819e+00 4.98758852e-01
9.77100879e-02 -1.07610881e+00 -1.49872971e+00 -1.24495709e+00
1.67986214e-01 -2.68842518e-01 9.28838551e-01 5.56821585e-01
-1.83885008e-01 -1.80403426e-01 -6.84182867e-02 1.18154421e-01
8.61077249e-01 6.57964766e-01 3.35228413e-01 -1.33917093e+00
-7.15711772e-01 -8.36009920e-01 -1.20698437e-01 -5.17441869e-01
-1.01203464e-01 -7.00091243e-01 -2.12900952e-01 -1.40541112e+00
1.30958185e-01 -7.74578691e-01 -3.08610857e-01 3.77976626e-01
4.58154231e-02 1.06956869e-01 3.21692675e-01 -4.79965329e-01
-4.74540979e-01 1.40087530e-01 1.19243169e+00 -8.33373144e-02
-2.49308690e-01 1.41220897e-01 -7.71957517e-01 5.17523766e-01
7.05537379e-01 -8.09308171e-01 -9.47459042e-02 -4.81037676e-01
8.34868670e-01 3.12207192e-01 -2.30962038e-01 -8.60900819e-01
1.24731630e-01 -5.82091749e-01 -8.74415115e-02 -4.95805144e-01
2.93056160e-01 -1.02708912e+00 5.72186172e-01 9.54796910e-01
2.10905373e-02 2.05465794e-01 2.80119121e-01 2.87630767e-01
1.17153004e-01 -8.28325391e-01 7.30335414e-01 -1.34768579e-02
-2.07855284e-01 -2.33936738e-02 -2.35879868e-01 -1.05139866e-01
1.45933259e+00 -6.32733107e-01 -5.17629206e-01 -1.05961718e-01
-4.97033387e-01 5.61567783e-01 4.92416203e-01 -1.68446705e-01
1.99482292e-01 -8.07793736e-01 -6.57328665e-01 -1.08948454e-01
-2.33573496e-01 -4.22197804e-02 3.66969109e-01 1.07145154e+00
-6.65568471e-01 3.76775891e-01 -1.94795340e-01 -2.56317973e-01
-1.46784246e+00 6.73563004e-01 2.34093681e-01 -5.95725596e-01
1.30460471e-01 9.75529134e-01 -3.56660545e-01 4.68993187e-02
2.72692684e-02 2.40041073e-02 3.68505597e-01 6.91783950e-02
2.02185139e-01 1.08957994e+00 1.70016706e-01 -2.85770655e-01
-6.30396843e-01 7.75882185e-01 4.37335581e-01 -2.02828333e-01
1.63069308e+00 -8.70089307e-02 -6.72332704e-01 -2.70780653e-01
1.15432096e+00 2.28695840e-01 -3.98801714e-01 4.04148012e-01
4.65511307e-02 -5.25779486e-01 -2.11122334e-01 -7.92297721e-01
-1.43339026e+00 6.28798723e-01 5.44651806e-01 4.82178509e-01
1.77228272e+00 -2.76348591e-01 7.66052961e-01 3.57068688e-01
6.74039006e-01 -1.32927680e+00 -3.81171077e-01 2.23876864e-01
4.46947485e-01 -3.98151398e-01 2.52144724e-01 -6.25687778e-01
4.60662022e-02 1.22268939e+00 3.65542978e-01 1.90951228e-01
2.01376811e-01 4.46629286e-01 -6.21680677e-01 9.86167602e-03
-5.31915605e-01 -3.76152247e-01 -1.49831042e-01 8.95961002e-02
-3.05205770e-03 9.45362896e-02 -1.03958619e+00 6.86323583e-01
-2.47871161e-01 -5.23116551e-02 4.33494329e-01 1.11851346e+00
-8.22117627e-01 -1.52756572e+00 -5.86402893e-01 3.05161029e-01
-6.02741480e-01 1.67330608e-01 -2.51839817e-01 7.33538687e-01
3.08626831e-01 1.20979714e+00 -1.21275231e-01 -6.37865886e-02
1.45225748e-01 -6.99202865e-02 9.85584319e-01 2.41966899e-02
-8.39003980e-01 -1.01806670e-01 1.10425673e-01 -5.62859356e-01
-2.19910353e-01 -3.48692358e-01 -1.63952112e+00 -7.92969525e-01
-5.77873409e-01 6.83404922e-01 9.80604768e-01 8.19778740e-01
3.08762580e-01 4.81273264e-01 6.88291907e-01 -3.32519889e-01
-2.81434238e-01 -3.15478325e-01 -7.62586236e-01 -1.20165601e-01
-3.86995643e-01 -5.25803328e-01 -5.23534358e-01 -5.14606774e-01] | [6.301247596740723, 4.466187953948975] |
1101a9b7-5830-4ffe-af75-ee78a3c570e3 | glt-t-global-local-transformer-for-3d-siamese | 2304.00242 | null | https://arxiv.org/abs/2304.00242v1 | https://arxiv.org/pdf/2304.00242v1.pdf | GLT-T++: Global-Local Transformer for 3D Siamese Tracking with Ranking Loss | Siamese trackers based on 3D region proposal network (RPN) have shown remarkable success with deep Hough voting. However, using a single seed point feature as the cue for voting fails to produce high-quality 3D proposals. Additionally, the equal treatment of seed points in the voting process, regardless of their significance, exacerbates this limitation. To address these challenges, we propose a novel transformer-based voting scheme to generate better proposals. Specifically, a global-local transformer (GLT) module is devised to integrate object- and patch-aware geometric priors into seed point features, resulting in robust and accurate cues for offset learning of seed points. To train the GLT module, we introduce an importance prediction branch that learns the potential importance weights of seed points as a training constraint. Incorporating this transformer-based voting scheme into 3D RPN, a novel Siamese method dubbed GLT-T is developed for 3D single object tracking on point clouds. Moreover, we identify that the highest-scored proposal in the Siamese paradigm may not be the most accurate proposal, which limits tracking performance. Towards this concern, we approach the binary score prediction task as a ranking problem, and design a target-aware ranking loss and a localization-aware ranking loss to produce accurate ranking of proposals. With the ranking losses, we further present GLT-T++, an enhanced version of GLT-T. Extensive experiments on multiple benchmarks demonstrate that our GLT-T and GLT-T++ outperform state-of-the-art methods in terms of tracking accuracy while maintaining a real-time inference speed. The source code will be made available at https://github.com/haooozi/GLT-T. | ['Jing Zhang', 'Mingyu Gao', 'Xudong Lv', 'Yuxiang Yang', 'Zhiwei He', 'Jiahao Nie'] | 2023-04-01 | null | null | null | null | ['3d-single-object-tracking'] | ['computer-vision'] | [-2.60981232e-01 -1.67448923e-01 -3.28884840e-01 -3.37548077e-01
-9.45643127e-01 -6.12284005e-01 6.46510303e-01 2.18179282e-02
-4.19879615e-01 3.75976533e-01 -7.04126209e-02 -2.37485711e-02
-1.17719807e-01 -6.37556136e-01 -8.86138499e-01 -5.81319332e-01
2.78495601e-04 5.71771383e-01 7.19866753e-01 -3.70251350e-02
3.01423639e-01 6.79998815e-01 -1.42515683e+00 -2.02875301e-01
8.78291607e-01 1.21570742e+00 1.95864469e-01 1.57468125e-01
7.22037479e-02 2.72068560e-01 -6.26891434e-01 -4.05544937e-01
8.04515183e-01 3.51229161e-02 -1.21100908e-02 -3.72980624e-01
1.10269237e+00 -3.18696916e-01 -1.87734485e-01 9.95635748e-01
5.60849190e-01 1.56839490e-02 6.65985107e-01 -1.43108273e+00
-2.97270596e-01 4.01087224e-01 -8.10521483e-01 -1.36597231e-02
3.00660711e-02 3.31747562e-01 1.18101001e+00 -1.26628816e+00
4.63299334e-01 1.19930875e+00 1.01042485e+00 4.16011781e-01
-1.21937978e+00 -1.04031384e+00 4.39797103e-01 -1.71863526e-01
-1.62978983e+00 -1.80365145e-01 8.81059587e-01 -3.74159008e-01
5.16078174e-01 1.99735671e-01 7.95499146e-01 7.78065562e-01
2.42617399e-01 8.06816995e-01 8.39958966e-01 -6.84746802e-02
1.25501439e-01 -1.28324926e-01 -2.08955407e-02 8.13629568e-01
6.47917569e-01 5.02624989e-01 -6.13285422e-01 -3.51306319e-01
9.30372417e-01 1.11379460e-01 -4.62596752e-02 -9.88207400e-01
-1.30125380e+00 7.43264139e-01 1.10950041e+00 3.36194336e-02
-3.67193043e-01 4.36560035e-01 3.34804654e-02 -7.55450353e-02
4.99490529e-01 4.35968786e-01 -2.69419044e-01 3.41264367e-01
-1.15307152e+00 5.98856807e-01 3.91373336e-01 9.45140958e-01
5.46179831e-01 -1.06549300e-01 -6.86601877e-01 5.80747902e-01
6.66923940e-01 8.50515306e-01 -1.93677455e-01 -1.07792985e+00
3.87037426e-01 7.02441752e-01 3.80153537e-01 -1.00083041e+00
-3.36384952e-01 -9.09143269e-01 -4.38065231e-01 5.47150493e-01
4.87984151e-01 1.80088550e-01 -9.98957813e-01 1.90806878e+00
7.92354524e-01 2.80284882e-01 -4.45803940e-01 1.39172220e+00
7.05768645e-01 3.74995172e-01 1.66362207e-02 1.70209914e-01
1.30149436e+00 -8.51005495e-01 -1.33280650e-01 -1.27408653e-01
3.42348784e-01 -7.34426916e-01 6.19042218e-01 2.10769668e-01
-1.05660045e+00 -5.81424296e-01 -9.00778055e-01 1.92022890e-01
-2.85253506e-02 4.26404029e-01 4.86519516e-01 4.25455719e-01
-9.01959062e-01 4.07907367e-01 -8.50141585e-01 -1.64086774e-01
6.69677198e-01 4.59929466e-01 4.46072780e-02 7.79117048e-02
-8.80708814e-01 1.00037837e+00 5.00023887e-02 9.78824794e-02
-9.51455355e-01 -1.01256371e+00 -6.75443232e-01 -9.92655531e-02
4.82998192e-01 -8.07456791e-01 1.29484785e+00 -3.47090781e-01
-1.26916504e+00 6.68184102e-01 -1.66116282e-01 -4.75309402e-01
6.89854920e-01 -1.95603684e-01 -5.48084043e-02 -7.17600510e-02
4.02715355e-01 9.85131204e-01 9.36715841e-01 -1.47552478e+00
-8.72091413e-01 -3.01822513e-01 -8.12885687e-02 1.47852406e-01
7.27349594e-02 -2.11672530e-01 -6.23469949e-01 -7.65021801e-01
3.32468122e-01 -9.79988396e-01 -2.80750275e-01 5.98980784e-01
-1.95345029e-01 -5.81204951e-01 6.62090659e-01 -1.47541255e-01
9.70600307e-01 -1.94524837e+00 -1.39976054e-01 4.03368086e-01
5.05557775e-01 2.60081917e-01 -1.66393220e-01 -3.77092138e-02
4.13591146e-01 -2.69052535e-01 -1.89593397e-02 -4.54241157e-01
3.03007513e-01 -4.58782315e-02 -3.20644736e-01 5.70052385e-01
3.81714463e-01 1.02150059e+00 -1.10666835e+00 -5.33226788e-01
4.05440986e-01 5.36465704e-01 -7.10256457e-01 -8.94485265e-02
-3.84983301e-01 3.64494354e-01 -7.12231994e-01 9.41856086e-01
8.84039521e-01 -3.58511180e-01 -2.07091257e-01 -4.50791985e-01
-4.10996854e-01 2.62982577e-01 -1.07910919e+00 1.58813119e+00
-1.13854960e-01 2.48408690e-01 5.44402599e-02 -3.98385137e-01
1.08761501e+00 -1.06601648e-01 6.72165096e-01 -6.03208899e-01
1.37367874e-01 4.50123250e-01 -7.04073459e-02 2.86766052e-01
7.65153587e-01 2.95670666e-02 -2.13403821e-01 1.89333215e-01
-1.02532499e-01 -1.21198252e-01 -5.89646865e-03 1.65942356e-01
1.09706247e+00 4.49225426e-01 8.61282125e-02 -7.86797404e-02
3.05197984e-01 2.94850290e-01 1.00049794e+00 1.01437569e+00
-5.23438454e-01 7.29075789e-01 2.58562416e-02 -3.27790171e-01
-9.40129995e-01 -1.19939911e+00 -3.37424427e-01 9.21091735e-01
7.08408713e-01 -1.99366659e-01 -3.26171279e-01 -8.66631150e-01
5.85561574e-01 4.37697053e-01 -3.91535610e-01 -3.07125393e-02
-6.59138203e-01 -5.08233011e-01 4.50558066e-01 5.17799258e-01
3.23875129e-01 -6.89777911e-01 -7.09852517e-01 2.51098752e-01
5.33088222e-02 -8.21649194e-01 -7.99108922e-01 9.71946418e-02
-8.47514033e-01 -9.89503145e-01 -8.75572741e-01 -4.85825837e-01
8.04739714e-01 5.89915395e-01 9.93179262e-01 5.51339835e-02
-2.97362246e-02 2.60478556e-01 -2.75049597e-01 -3.13630015e-01
5.63351698e-02 1.04509084e-03 1.75715432e-01 -1.36447847e-01
3.69576752e-01 -2.90916860e-01 -8.95014048e-01 7.08073139e-01
-3.87316763e-01 -1.70047194e-01 8.39173198e-01 7.01392949e-01
7.53114939e-01 -3.18055958e-01 1.91923872e-01 -2.20275894e-01
1.63297862e-01 -2.65166104e-01 -1.12377739e+00 8.25842768e-02
-5.29047132e-01 1.11259453e-01 2.96440065e-01 -5.91722786e-01
-6.87984884e-01 2.96272576e-01 -8.77197075e-04 -9.38369095e-01
2.92283297e-01 1.93469226e-01 8.04519653e-03 -6.72599494e-01
5.89603305e-01 1.12091303e-02 -4.58375625e-02 -4.75439250e-01
2.96535075e-01 1.40171632e-01 5.24234116e-01 -6.25275612e-01
1.37984216e+00 5.36386251e-01 1.08711198e-01 -2.07112253e-01
-1.05908370e+00 -4.80206221e-01 -3.93189847e-01 -4.79416013e-01
6.29862785e-01 -1.10334003e+00 -7.20599890e-01 2.93734848e-01
-1.07502747e+00 -1.64294377e-01 -4.13884342e-01 5.99263489e-01
-2.78522104e-01 2.05195189e-01 -1.97191164e-01 -8.02673519e-01
-3.02261233e-01 -1.10746610e+00 1.43577909e+00 4.43472415e-01
-5.14861522e-03 -5.54802239e-01 1.03485256e-01 2.85970062e-01
4.52000976e-01 2.62628913e-01 2.65424550e-01 -4.74053234e-01
-1.18160057e+00 -3.86303633e-01 -3.56739044e-01 -8.08650777e-02
-2.51869351e-01 -1.10360965e-01 -7.58590937e-01 -5.39691389e-01
-2.25373313e-01 -1.43147171e-01 9.99917746e-01 6.77210391e-01
9.42386389e-01 5.52287176e-02 -5.74819803e-01 6.43371224e-01
1.31493223e+00 -1.80471346e-01 1.01895958e-01 3.03201020e-01
5.74000359e-01 4.64253910e-02 1.10252690e+00 4.50004160e-01
6.04669452e-01 1.11844003e+00 7.42836773e-01 6.69727698e-02
-3.28806400e-01 -6.57194316e-01 3.73823524e-01 5.14016032e-01
-4.82753776e-02 -3.22071947e-02 -6.93761230e-01 4.45516974e-01
-1.91065478e+00 -8.82718801e-01 -9.14292932e-02 2.49836564e+00
6.01267219e-01 3.83107334e-01 1.80782288e-01 -3.24765652e-01
8.49264801e-01 2.60447621e-01 -7.19263911e-01 4.87698168e-01
8.98830220e-02 1.13373874e-02 7.97834694e-01 4.22157824e-01
-1.26398611e+00 9.49043751e-01 4.93114138e+00 1.01334417e+00
-1.20516479e+00 1.74982339e-01 6.25475645e-02 -2.50009090e-01
-2.36279398e-01 1.89628497e-01 -1.40847850e+00 6.28485501e-01
2.94395775e-01 2.75038425e-02 1.89473629e-02 8.52236986e-01
2.37012371e-01 2.19291728e-02 -7.76209056e-01 8.21529508e-01
-5.49822040e-02 -1.20759082e+00 -8.30662325e-02 1.64637834e-01
6.30996585e-01 4.03252870e-01 1.96414277e-01 3.86641741e-01
5.01766205e-01 -4.60918784e-01 1.21767211e+00 4.74752307e-01
4.23037231e-01 -4.87567991e-01 4.66069847e-01 1.91765070e-01
-1.62123871e+00 -1.03422878e-02 -5.26982307e-01 3.31246912e-01
2.58251369e-01 6.64946973e-01 -7.35672116e-01 6.80632591e-01
7.57461309e-01 7.45553255e-01 -6.23621762e-01 1.71576595e+00
-3.88183087e-01 3.96684140e-01 -8.10667574e-01 -3.27153094e-02
2.39766568e-01 1.14166006e-01 9.89358962e-01 7.30057836e-01
5.68246782e-01 -1.09068818e-01 4.31258202e-01 1.14876699e+00
-1.00542031e-01 -3.31038296e-01 -2.70719349e-01 4.58399892e-01
9.19227839e-01 1.41944826e+00 -7.27313757e-01 -1.52057186e-01
-1.69966549e-01 3.25019509e-01 3.19535494e-01 3.78602780e-02
-1.19025159e+00 -6.71532676e-02 7.00374424e-01 2.81718701e-01
6.46600366e-01 -1.93781704e-01 -2.38027215e-01 -1.09628737e+00
1.27965167e-01 -5.54853082e-01 2.24478930e-01 -6.56644166e-01
-1.48016071e+00 4.10525024e-01 8.55677277e-02 -1.93125105e+00
2.51303911e-01 -5.37810326e-01 -6.09206915e-01 6.54019773e-01
-1.70137501e+00 -1.24989510e+00 -5.19342482e-01 2.76638925e-01
2.53798127e-01 2.77007334e-02 5.93418740e-02 3.58251035e-01
-3.06083411e-01 7.17545867e-01 -2.07103997e-01 1.54928658e-02
8.90987217e-01 -1.09603620e+00 4.23753232e-01 8.03726315e-01
4.34676670e-02 6.84995353e-01 6.30793095e-01 -8.41705978e-01
-1.37031853e+00 -1.30982745e+00 7.91273534e-01 -8.38178515e-01
5.09105623e-01 -2.69430548e-01 -7.05507755e-01 4.30503786e-01
-4.26231146e-01 3.68496329e-01 1.64524049e-01 -1.05276369e-01
-3.79002661e-01 -3.37807328e-01 -1.12132013e+00 5.20232916e-01
1.21978164e+00 -6.10447526e-02 -6.61785245e-01 1.80214092e-01
6.48796916e-01 -6.79598331e-01 -9.41866994e-01 7.86712706e-01
7.06605911e-01 -7.76042700e-01 1.26055670e+00 1.82795465e-01
-4.94484678e-02 -1.01408625e+00 7.48348907e-02 -9.56296444e-01
-4.45113480e-01 -3.68543178e-01 -3.48523051e-01 1.01953042e+00
2.35492483e-01 -6.25681639e-01 1.02131772e+00 3.88278127e-01
-3.62126797e-01 -6.82742178e-01 -1.31940079e+00 -1.07813370e+00
-8.46589655e-02 -3.07264864e-01 6.65788651e-01 6.55099511e-01
-5.89375675e-01 -1.47167733e-02 -3.44974905e-01 4.56566989e-01
1.25644028e+00 3.91222447e-01 1.07794631e+00 -1.60021567e+00
-6.85458705e-02 -7.06539452e-01 -4.06158358e-01 -1.53269207e+00
-1.87221110e-01 -9.74375904e-01 3.07921976e-01 -1.39405310e+00
1.14994720e-01 -1.02435422e+00 -3.64660859e-01 5.87942123e-01
-2.33424470e-01 3.99078667e-01 4.99155670e-01 5.66936851e-01
-9.01754797e-01 8.98263216e-01 1.27507281e+00 -2.00020313e-01
-1.28906935e-01 3.10090691e-01 -4.97840345e-01 5.18009245e-01
5.03826022e-01 -8.04459274e-01 9.05820355e-02 -3.50513160e-01
-1.27606150e-02 -2.17252657e-01 8.60530198e-01 -1.19249511e+00
5.44509649e-01 -4.27766405e-02 5.47650754e-01 -1.19066072e+00
6.10654831e-01 -9.80086803e-01 9.46972743e-02 6.24685109e-01
-8.81639570e-02 -2.64060795e-01 1.78255644e-02 5.90368807e-01
-5.64098619e-02 -3.66347283e-02 7.73334086e-01 1.33190215e-01
-6.20133936e-01 6.79893374e-01 2.58270293e-01 -1.66970626e-01
9.74439085e-01 -4.04769331e-01 -4.46238160e-01 9.28901695e-03
-1.81169435e-01 7.06086576e-01 7.31569409e-01 5.01539230e-01
4.74388450e-01 -1.60412943e+00 -5.66479325e-01 1.36344180e-01
2.85611242e-01 2.64667124e-01 1.20624630e-02 1.15965497e+00
-1.89549550e-01 4.47741091e-01 4.30692956e-02 -9.68810439e-01
-1.14730251e+00 2.21881703e-01 3.70739549e-01 -3.01669687e-01
-6.01383686e-01 8.20600092e-01 2.68552899e-01 -7.05672920e-01
3.68047476e-01 -5.46202302e-01 5.49721383e-02 -2.06219882e-01
1.54237062e-01 1.45387858e-01 -1.46271139e-01 -6.09155595e-01
-5.67312300e-01 9.49554205e-01 4.94923219e-02 1.05109401e-01
1.24332619e+00 8.61318037e-02 3.22067559e-01 -6.67736381e-02
5.61250269e-01 7.69556463e-02 -1.84878707e+00 -4.01455194e-01
9.91338938e-02 -6.84246659e-01 1.37098357e-01 -7.45355606e-01
-1.20535016e+00 3.95559520e-01 6.43200696e-01 -6.93451837e-02
7.16901481e-01 8.55717137e-02 7.87167668e-01 1.77079663e-01
7.34088778e-01 -7.38510787e-01 6.55073896e-02 5.83439410e-01
6.98821008e-01 -1.37919807e+00 2.23074928e-01 -3.25372547e-01
-2.72747785e-01 7.54480720e-01 8.90132189e-01 -3.55855614e-01
4.34802264e-01 6.97183460e-02 -3.59820127e-02 -3.37188095e-01
-5.61425328e-01 -3.82764071e-01 6.26679599e-01 4.09230560e-01
1.06486700e-01 -1.48648247e-01 -3.28548223e-01 4.38087791e-01
-1.21847212e-01 1.30538139e-02 -1.49331406e-01 1.01992381e+00
-6.67348087e-01 -1.06544363e+00 -7.46269763e-01 3.99100840e-01
-4.65564467e-02 1.91252142e-01 -2.37551302e-01 7.35001922e-01
2.21499234e-01 5.21406293e-01 -1.59821808e-02 -4.29062784e-01
5.24662375e-01 -3.64503920e-01 5.48092902e-01 -3.67901534e-01
-7.23952889e-01 1.64367959e-01 -3.02816182e-01 -7.15986431e-01
-4.11957115e-01 -6.39325976e-01 -1.17899585e+00 -2.35100284e-01
-6.15751266e-01 2.14211509e-01 5.94383776e-01 6.17260516e-01
5.24534702e-01 2.93580651e-01 5.55851042e-01 -1.34395134e+00
-9.07303691e-01 -6.63694501e-01 -2.39692748e-01 1.25537172e-01
3.39754075e-01 -1.20965278e+00 -4.22589630e-01 -6.27453387e-01] | [6.514237880706787, -2.2397820949554443] |
890bbbc2-7beb-434f-811f-b65bd9566a61 | heart-sound-classification-considering | 2106.01865 | null | https://arxiv.org/abs/2106.01865v1 | https://arxiv.org/pdf/2106.01865v1.pdf | Heart Sound Classification Considering Additive Noise and Convolutional Distortion | Cardiac auscultation is an essential point-of-care method used for the early diagnosis of heart diseases. Automatic analysis of heart sounds for abnormality detection is faced with the challenges of additive noise and sensor-dependent degradation. This paper aims to develop methods to address the cardiac abnormality detection problem when both types of distortions are present in the cardiac auscultation sound. We first mathematically analyze the effect of additive and convolutional noise on short-term filterbank-based features and a Convolutional Neural Network (CNN) layer. Based on the analysis, we propose a combination of linear and logarithmic spectrogram-image features. These 2D features are provided as input to a residual CNN network (ResNet) for heart sound abnormality detection. Experimental validation is performed on an open-access heart sound abnormality detection dataset involving noisy recordings obtained from multiple stethoscope sensors. The proposed method achieves significantly improved results compared to the conventional approaches, with an area under the ROC (receiver operating characteristics) curve (AUC) of 91.36%, F-1 score of 84.09%, and Macc (mean of sensitivity and specificity) of 85.08%. We also show that the proposed method shows the best mean accuracy across different source domains including stethoscope and noise variability, demonstrating its effectiveness in different recording conditions. The proposed combination of linear and logarithmic features along with the ResNet classifier effectively minimizes the impact of background noise and sensor variability for classifying phonocardiogram (PCG) signals. The proposed method paves the way towards developing computer-aided cardiac auscultation systems in noisy environments using low-cost stethoscopes. | ['Taufiq Hasan', 'Ian Mclane', 'Md. Istiaq Ansari', 'Farhat Binte Azam'] | 2021-06-03 | null | null | null | null | ['sound-classification'] | ['audio'] | [ 4.05843467e-01 -3.54680657e-01 5.47082365e-01 -6.39029872e-03
-7.24424958e-01 -4.50265855e-01 -1.67599529e-01 2.65003979e-01
-3.74814332e-01 3.45939100e-01 7.93425832e-03 -5.19746900e-01
-2.40365118e-01 -3.89519483e-01 -1.86793238e-01 -6.51554465e-01
-4.47959363e-01 -2.79514611e-01 5.03837503e-02 1.37875333e-01
-6.40804172e-02 5.08760750e-01 -1.44402623e+00 -2.43241200e-03
6.60308361e-01 1.28834796e+00 5.34419939e-02 1.47822368e+00
5.54247260e-01 3.51831704e-01 -1.08643031e+00 1.40488595e-01
2.61533767e-01 -8.49339664e-01 -2.51051843e-01 -2.16767192e-01
2.32973412e-01 -2.66401321e-01 3.31952982e-02 8.38507533e-01
1.33933854e+00 -8.27678218e-02 4.16851372e-01 -5.05034626e-01
-1.37199700e-01 3.74776483e-01 6.01832522e-03 9.16662633e-01
2.28382409e-01 1.86620340e-01 6.29446089e-01 -7.03346193e-01
-6.22665994e-02 6.13300383e-01 1.31451654e+00 1.26391649e-01
-8.98926556e-01 -4.98651624e-01 -7.60569274e-01 4.40134332e-02
-1.29040301e+00 -1.18225016e-01 9.04916942e-01 -4.54390407e-01
9.34582353e-01 5.12797117e-01 7.73577750e-01 4.38026458e-01
4.86092240e-01 -5.85821718e-02 1.07315171e+00 -6.73331141e-01
-7.03022769e-03 3.37995179e-02 2.48089075e-01 6.48035526e-01
5.04940808e-01 2.12633282e-01 -1.42214298e-01 -3.77924681e-01
7.33540654e-01 1.23490497e-01 -5.33891141e-01 4.71122712e-01
-1.03386724e+00 3.60374093e-01 1.22478284e-01 7.83056140e-01
-6.12722278e-01 1.15612634e-01 5.25735259e-01 4.26370740e-01
2.51256227e-01 6.66815639e-01 -5.57577789e-01 -4.12187487e-01
-8.71313095e-01 5.48734628e-02 8.55015397e-01 6.17665574e-02
-4.61208783e-02 5.58532238e-01 -2.46331245e-01 1.05200553e+00
1.91503108e-01 8.65743399e-01 6.50822580e-01 -7.43631184e-01
2.06471100e-01 1.90291092e-01 1.28330246e-01 -1.29769266e+00
-7.50906765e-01 -1.05701089e+00 -1.04301500e+00 -1.72075242e-01
2.73183405e-01 -5.29386640e-01 -6.46745622e-01 1.27547383e+00
1.00586019e-01 5.17745852e-01 1.81365401e-01 9.54838276e-01
9.42948639e-01 4.04446304e-01 2.95010787e-02 -3.62365305e-01
1.46038365e+00 -3.50599587e-01 -9.31621611e-01 2.50332326e-01
4.14290220e-01 -7.20184803e-01 7.76654065e-01 4.49604243e-01
-9.36617017e-01 -9.91406381e-01 -1.19620979e+00 5.33473253e-01
-7.77790621e-02 3.57508153e-01 -2.49339446e-01 1.19346654e+00
-9.46509242e-01 8.79991829e-01 -8.21745455e-01 -2.51935065e-01
2.52657562e-01 3.20813179e-01 5.92496730e-02 3.67892623e-01
-1.35069942e+00 5.41450977e-01 -4.47567590e-02 4.16071296e-01
-5.01166880e-01 -7.23961473e-01 -7.18382418e-01 2.71159619e-01
-2.22823352e-01 -6.20561182e-01 1.02296054e+00 -5.99314213e-01
-1.45341265e+00 3.85124892e-01 2.05105469e-01 -6.19507432e-01
4.07680631e-01 -3.95104170e-01 -8.08463931e-01 3.22119862e-01
-3.48071724e-01 -2.92530775e-01 1.05836880e+00 -5.76396227e-01
-4.60965395e-01 -3.13174337e-01 -4.71104294e-01 -2.00365037e-02
-3.50176573e-01 1.32042304e-01 1.36225775e-01 -8.56038988e-01
4.47504997e-01 -6.96429491e-01 2.18318701e-02 -4.03271735e-01
-2.01670796e-01 2.14015365e-01 6.35889709e-01 -1.04855025e+00
1.50554490e+00 -2.16297102e+00 -5.35174787e-01 3.64727229e-01
3.00908536e-01 9.25147593e-01 2.09057704e-01 1.97133571e-01
-1.57968521e-01 2.00768292e-01 -3.47235084e-01 -3.34304012e-02
-5.17328501e-01 -1.79243997e-01 3.88656743e-02 6.64494216e-01
1.34852096e-01 7.03790545e-01 -6.17061913e-01 -1.22286148e-01
6.10205770e-01 7.64965534e-01 -1.28982142e-01 5.14887929e-01
4.10678506e-01 5.88636577e-01 -1.49879605e-01 3.95988435e-01
4.86667693e-01 -3.36058512e-02 -1.01358807e-02 -3.73437047e-01
2.96278670e-02 1.70633748e-01 -1.32754362e+00 1.23100805e+00
-4.76922095e-01 6.13107741e-01 -9.11849141e-02 -8.91086042e-01
1.23964405e+00 8.98591101e-01 4.97660607e-01 -3.81562620e-01
3.75608861e-01 5.35054922e-01 5.25151908e-01 -1.06149697e+00
-4.19325382e-01 -3.05648416e-01 2.22714037e-01 3.99802387e-01
-6.86786100e-02 -1.07444994e-01 -3.34849209e-01 -4.96882528e-01
1.24453461e+00 -4.25623715e-01 4.56556767e-01 -4.19288933e-01
8.13841462e-01 -5.46478510e-01 4.46471184e-01 1.10187781e+00
-4.98898208e-01 8.58236790e-01 1.02796271e-01 -7.35657752e-01
-7.37097859e-01 -7.21954346e-01 -1.84192300e-01 4.33593988e-01
-3.24988395e-01 -5.24583980e-02 -7.07176208e-01 -1.61193743e-01
-1.44591302e-01 5.30616231e-02 -3.30765694e-01 -9.69382823e-02
-9.05390441e-01 -7.31911540e-01 1.23406827e+00 5.68894804e-01
5.28809965e-01 -1.17498159e+00 -1.15730417e+00 4.79134947e-01
-1.36815220e-01 -8.75949144e-01 -1.15698859e-01 1.03308678e-01
-9.74721909e-01 -1.13007748e+00 -8.18388283e-01 -7.05412924e-01
8.16530958e-02 -1.58962816e-01 8.33408952e-01 2.16497302e-01
-7.53744602e-01 4.53602493e-01 -2.72576481e-01 -6.78234398e-01
-5.43508053e-01 -9.04373825e-02 6.41792044e-02 9.71724838e-02
1.05195373e-01 -7.27223039e-01 -1.02779746e+00 -2.68300194e-02
-6.74567759e-01 -6.11193538e-01 2.99875557e-01 9.18989003e-01
3.49842578e-01 1.31594479e-01 1.07581830e+00 -6.56049192e-01
1.17712474e+00 -4.46629137e-01 -5.02139688e-01 -2.05348626e-01
-7.76190817e-01 -6.39463663e-01 9.09028292e-01 -2.88568884e-01
-5.53721547e-01 -1.68374255e-01 -4.96616930e-01 -5.04714549e-01
-4.54764187e-01 3.81236285e-01 3.01167995e-01 1.50705487e-01
9.47153807e-01 2.16947690e-01 1.85264871e-01 -5.65174639e-01
-4.59153116e-01 9.26211655e-01 8.01027417e-01 -1.12936899e-01
5.35412610e-01 7.71371201e-02 1.96891382e-01 -1.09430361e+00
-4.46446866e-01 -6.90910757e-01 -3.85340303e-01 -3.46138626e-01
9.27544177e-01 -7.34744906e-01 -6.63989186e-01 8.19982708e-01
-1.12363207e+00 1.51611984e-01 -3.12716603e-01 7.72022963e-01
-1.41551942e-01 5.01152098e-01 -6.41928554e-01 -1.34907258e+00
-1.00210297e+00 -8.54822814e-01 6.40649736e-01 1.52687803e-01
-4.02776122e-01 -9.66639876e-01 1.13560438e-01 2.05196857e-01
8.66720855e-01 7.57074356e-01 6.24756634e-01 -8.40744436e-01
1.44302353e-01 -5.45855522e-01 1.67674437e-01 1.02329886e+00
4.38951582e-01 -2.78580725e-01 -1.25371909e+00 -3.57011676e-01
5.36618412e-01 1.58495769e-01 6.28437281e-01 8.00143301e-01
1.14859104e+00 -1.56077564e-01 2.55010754e-01 5.89513838e-01
1.52008641e+00 5.36168575e-01 5.22299767e-01 -1.03029199e-01
5.85006595e-01 3.17596108e-01 1.32616118e-01 5.51943123e-01
-2.18191564e-01 7.21490458e-02 1.98146135e-01 -5.31305432e-01
-3.05852562e-01 1.45078704e-01 4.13828436e-03 1.17439747e+00
-1.91036686e-01 -2.32979521e-01 -1.00961602e+00 6.24035537e-01
-1.30775321e+00 -7.24342585e-01 -4.55755144e-01 2.26164103e+00
4.91820961e-01 -3.05516310e-02 1.45097256e-01 9.75556612e-01
9.02315557e-01 -6.07700981e-02 -4.67283189e-01 -5.79838932e-01
-4.05444019e-02 8.07552814e-01 2.14588761e-01 3.90724897e-01
-1.34998739e+00 -7.41972476e-02 5.87283850e+00 1.95760742e-01
-1.49211502e+00 1.99240521e-01 6.41461253e-01 3.17302346e-01
3.31021219e-01 -6.77059233e-01 -9.75543112e-02 5.10778368e-01
1.27107966e+00 3.21285218e-01 8.03256929e-02 4.59373951e-01
5.12531042e-01 1.54410213e-01 -6.74434602e-01 1.18290687e+00
1.35942444e-01 -1.04481435e+00 -6.09114408e-01 -4.15909410e-01
5.00044405e-01 -6.05487078e-03 3.70856412e-02 -1.93672702e-01
-7.93463767e-01 -1.00308287e+00 3.07811022e-01 6.90891266e-01
1.12218618e+00 -5.02981544e-01 1.31071651e+00 1.54348314e-01
-1.14574063e+00 -3.39164466e-01 1.32621571e-01 -2.02403709e-01
-1.86501250e-01 7.89586246e-01 -1.01289749e+00 2.38786563e-01
6.83037400e-01 5.22495627e-01 -2.18982935e-01 1.20652556e+00
1.11663610e-01 1.29611003e+00 -4.51879561e-01 -2.73980517e-02
-2.76517898e-01 3.74651939e-01 8.56085658e-01 1.44321156e+00
5.42778313e-01 3.44619453e-01 -4.79577780e-02 8.04858267e-01
2.47816354e-01 2.10909948e-01 -5.98998547e-01 1.92307845e-01
5.60980856e-01 9.22553837e-01 -6.05638683e-01 -3.67129534e-01
-1.96730211e-01 2.57846177e-01 -6.70036435e-01 2.64071226e-01
-6.30451560e-01 -9.41708446e-01 2.15414420e-01 3.87030035e-01
1.52944401e-01 2.16836259e-01 -6.36871219e-01 -5.99412382e-01
2.31987044e-01 -1.03426158e+00 4.08586234e-01 -3.48839670e-01
-9.81725216e-01 8.22931409e-01 -3.40027601e-01 -1.38728619e+00
-3.37075949e-01 -4.11141038e-01 -9.00604129e-01 1.22156417e+00
-1.35315561e+00 -5.47470272e-01 -4.65252131e-01 4.07184511e-01
4.15747166e-01 -1.89315826e-01 1.10288107e+00 4.92808074e-01
-3.31101477e-01 5.81601024e-01 -1.98769018e-01 1.13411233e-01
3.45381171e-01 -1.22590852e+00 1.51886016e-01 1.11152184e+00
-3.32940340e-01 6.26797318e-01 7.20973432e-01 -5.11621952e-01
-9.96721625e-01 -1.16980660e+00 8.82603347e-01 9.53309312e-02
2.33964175e-01 1.74020469e-01 -8.84699762e-01 -1.29066169e-01
-7.98570141e-02 3.17104816e-01 7.03241169e-01 -4.34558123e-01
9.74529311e-02 -4.60199058e-01 -1.21717525e+00 -4.55414020e-02
2.86613226e-01 -5.21728992e-01 -5.89175344e-01 -3.91424960e-03
4.44932580e-01 -3.33743453e-01 -1.22632241e+00 7.85588622e-01
8.59353721e-01 -9.65131760e-01 8.10637772e-01 -1.73049793e-01
7.69367740e-02 -2.12689921e-01 6.82663023e-02 -1.16134417e+00
-2.06865147e-01 -1.06248391e+00 -8.89717266e-02 8.09008420e-01
2.93423891e-01 -9.62165296e-01 2.30045334e-01 9.17944387e-02
-2.49418348e-01 -9.53617096e-01 -6.66465342e-01 -5.29503286e-01
-3.62765819e-01 -6.53319597e-01 2.83259362e-01 5.88491857e-01
-2.23295107e-01 -1.07599460e-02 -3.71180803e-01 2.46037543e-01
3.93150687e-01 -3.22254509e-01 2.10841745e-01 -1.34455705e+00
-3.91914994e-01 -8.04900527e-02 -5.64453661e-01 -3.04221183e-01
-6.19577587e-01 -4.29675728e-01 1.69401631e-01 -1.11003637e+00
-2.68439233e-01 -3.68741006e-01 -9.96835709e-01 7.72614554e-02
-4.47239220e-01 4.69195724e-01 2.83451453e-02 1.25270545e-01
5.31034637e-03 -1.92920387e-01 1.06815612e+00 1.25255555e-01
-5.20979226e-01 5.37598789e-01 -2.79775500e-01 8.16316843e-01
1.10833633e+00 -5.19989133e-01 -4.50442344e-01 -1.08983349e-02
-1.60808161e-01 5.02567470e-01 5.30701160e-01 -1.51110375e+00
-6.83906972e-02 6.09218717e-01 4.28755462e-01 -4.10666287e-01
1.41313672e-01 -7.79968143e-01 3.47063571e-01 9.70168352e-01
-4.31856126e-01 2.31528610e-01 2.43912026e-01 4.40515488e-01
-2.75566846e-01 -4.53825444e-02 9.89119411e-01 -1.03164673e-01
1.84011564e-01 -8.15119445e-02 -6.72611177e-01 2.14888737e-01
5.22732913e-01 -3.46381754e-01 -1.11347564e-01 -3.39067906e-01
-9.33020175e-01 -5.60860574e-01 -4.65306461e-01 2.02154323e-01
9.05628145e-01 -9.31531131e-01 -8.56125593e-01 6.67222142e-01
-7.86232576e-02 -3.33550841e-01 3.98059964e-01 1.14984024e+00
-8.94710839e-01 4.40265119e-01 -1.12197109e-01 -8.46449256e-01
-1.40095007e+00 -2.14673430e-01 9.38634455e-01 -2.36637682e-01
-6.90713525e-01 7.92047322e-01 -4.74390805e-01 1.05185226e-01
4.28348184e-01 -8.30453098e-01 -4.53147441e-01 -3.52139205e-01
5.03627002e-01 7.35775292e-01 4.66090739e-01 -4.06949908e-01
-2.74449587e-01 7.03473389e-01 5.01347184e-01 8.56054500e-02
1.11782193e+00 -1.49599707e-03 1.02796787e-02 5.08538067e-01
1.10670173e+00 -4.36251387e-02 -5.20130336e-01 -5.41914739e-02
-2.27428734e-01 -2.29779303e-01 1.02918319e-01 -1.01259065e+00
-1.08677959e+00 1.20233953e+00 1.55214083e+00 7.07353354e-01
1.48209035e+00 -5.20381272e-01 1.03363264e+00 1.85478300e-01
-2.21925884e-01 -8.11707377e-01 5.62392063e-02 8.69434103e-02
6.67483032e-01 -8.75049472e-01 -2.67380357e-01 -3.91786695e-02
-1.87988684e-01 1.23413122e+00 1.61741853e-01 -3.47032070e-01
1.20211053e+00 2.58998334e-01 6.13500595e-01 -2.19874397e-01
-2.79935271e-01 -4.37761508e-02 3.19161922e-01 6.32694364e-01
6.30659640e-01 2.75877357e-01 -5.54258108e-01 8.39351177e-01
-1.42018631e-01 7.50322416e-02 3.56296331e-01 8.24927270e-01
-6.27479017e-01 -5.60205221e-01 -5.80307603e-01 7.74546027e-01
-1.40821564e+00 -6.24886043e-02 5.34479320e-02 3.27038139e-01
6.17864132e-01 1.45732117e+00 -5.70491701e-02 -4.05345112e-01
4.58729446e-01 3.28705847e-01 7.06341192e-02 -4.88590300e-01
-1.13180947e+00 4.21861321e-01 -5.34965694e-02 -1.30811244e-01
-3.09992194e-01 -4.07814413e-01 -9.44450557e-01 3.59415025e-01
-5.48292696e-01 1.02980308e-01 8.25365424e-01 7.21537948e-01
4.30575132e-01 1.06238270e+00 6.48531854e-01 -3.81564409e-01
-7.14437783e-01 -1.31157124e+00 -6.85552776e-01 4.16432053e-01
9.72669601e-01 -1.55753523e-01 -5.22180557e-01 3.50451916e-01] | [14.271682739257812, 3.2607057094573975] |
61e41678-b564-4bdc-a981-cedd67abc1f8 | new-frontiers-in-graph-autoencoders-joint | 2211.08972 | null | https://arxiv.org/abs/2211.08972v1 | https://arxiv.org/pdf/2211.08972v1.pdf | New Frontiers in Graph Autoencoders: Joint Community Detection and Link Prediction | Graph autoencoders (GAE) and variational graph autoencoders (VGAE) emerged as powerful methods for link prediction (LP). Their performances are less impressive on community detection (CD), where they are often outperformed by simpler alternatives such as the Louvain method. It is still unclear to what extent one can improve CD with GAE and VGAE, especially in the absence of node features. It is moreover uncertain whether one could do so while simultaneously preserving good performances on LP in a multi-task setting. In this workshop paper, summarizing results from our journal publication (Salha-Galvan et al. 2022), we show that jointly addressing these two tasks with high accuracy is possible. For this purpose, we introduce a community-preserving message passing scheme, doping our GAE and VGAE encoders by considering both the initial graph and Louvain-based prior communities when computing embedding spaces. Inspired by modularity-based clustering, we further propose novel training and optimization strategies specifically designed for joint LP and CD. We demonstrate the empirical effectiveness of our approach, referred to as Modularity-Aware GAE and VGAE, on various real-world graphs. | ['Michalis Vazirgiannis', 'Romain Hennequin', 'George Dasoulas', 'Johannes F. Lutzeyer', 'Guillaume Salha-Galvan'] | 2022-11-16 | null | null | null | null | ['community-detection'] | ['graphs'] | [-1.86144397e-01 3.65507096e-01 8.27059075e-02 3.66126060e-01
-2.47177690e-01 -4.78107929e-01 6.99967980e-01 5.59650421e-01
-2.61165679e-01 6.97341263e-01 2.84201473e-01 -2.52412647e-01
-3.88606191e-01 -1.06701279e+00 -8.60517263e-01 -6.41488969e-01
-4.87673759e-01 5.48699439e-01 8.37058350e-02 -1.64487839e-01
-9.85186249e-02 3.60693544e-01 -1.34283257e+00 4.21380736e-02
8.97746086e-01 3.45425487e-01 1.05510980e-01 8.83560300e-01
5.01629524e-03 8.14589083e-01 -4.31842655e-01 -9.05636609e-01
-1.42273251e-02 -3.02356243e-01 -8.72442901e-01 2.51741022e-01
1.53490290e-01 2.67759450e-02 -5.50973892e-01 8.98361862e-01
3.17937911e-01 -5.91637231e-02 8.73051643e-01 -1.39290845e+00
-8.51174533e-01 9.27881002e-01 -5.49650013e-01 4.67766300e-02
8.54557380e-02 -3.22706789e-01 1.69836068e+00 -6.81065142e-01
9.97812510e-01 1.06739223e+00 9.97749507e-01 3.03006202e-01
-1.59574378e+00 -2.32630923e-01 -1.22490926e-02 4.71041650e-01
-1.56617594e+00 -7.57474229e-02 1.09901392e+00 -7.14514554e-01
1.05338490e+00 2.29251478e-02 6.73805416e-01 1.05701208e+00
1.17659740e-01 7.87635803e-01 6.06710315e-01 -4.61208761e-01
1.75118312e-01 1.32322326e-01 -1.05573982e-01 8.55015159e-01
5.69343925e-01 -9.45748463e-02 -2.88417876e-01 -4.12978560e-01
6.86012387e-01 -1.45625204e-01 -3.61224085e-01 -9.24481332e-01
-1.21710896e+00 1.27577364e+00 7.30948627e-01 6.36346281e-01
-6.27269328e-01 2.23514944e-01 4.48679596e-01 4.41305786e-01
5.18850863e-01 3.45705867e-01 -1.10999934e-01 3.31904948e-01
-8.91820550e-01 5.81067055e-02 9.54318225e-01 7.51681566e-01
7.78706491e-01 -2.32348610e-02 -3.90765592e-02 6.28611982e-01
3.84277463e-01 -7.68966451e-02 -5.18446155e-02 -7.55123556e-01
1.30870044e-01 6.27404928e-01 -2.94192135e-01 -1.54857326e+00
-4.19467807e-01 -9.07231331e-01 -1.25666535e+00 -5.12956381e-02
1.24572717e-01 -5.30128218e-02 -3.07645738e-01 1.84281921e+00
3.18637311e-01 2.82667786e-01 7.69176707e-02 6.50319755e-01
7.76123762e-01 5.66039085e-01 1.00065574e-01 -1.43180907e-01
1.14444077e+00 -8.49541605e-01 -5.94631076e-01 1.32067591e-01
6.65411770e-01 -4.13862348e-01 4.48592871e-01 2.57858425e-01
-8.30596149e-01 -3.48634034e-01 -1.04776800e+00 2.83275563e-02
-3.39527458e-01 7.93654099e-02 8.49153519e-01 5.49838781e-01
-1.55331910e+00 9.01267946e-01 -8.63106430e-01 -6.38092339e-01
4.21262801e-01 3.60493243e-01 -5.75516820e-01 1.37813091e-01
-1.13661849e+00 7.77713776e-01 3.64414155e-01 -1.30314112e-01
-6.16547763e-01 -3.12953949e-01 -8.48470688e-01 4.54860300e-01
4.83939856e-01 -1.01241350e+00 4.82322514e-01 -6.74667358e-01
-1.15922594e+00 6.44223809e-01 2.22474396e-01 -8.91906977e-01
3.28607112e-01 3.09586912e-01 -1.69090495e-01 3.46002012e-01
-8.18182454e-02 6.29886568e-01 7.86042631e-01 -1.46336854e+00
-1.15141705e-01 -7.14914799e-02 4.24010195e-02 -1.42374299e-02
-7.93333650e-01 -1.97002649e-01 -3.46027762e-01 -6.50749862e-01
-1.81445941e-01 -7.44981527e-01 -1.55189589e-01 1.38344079e-01
-5.27059913e-01 -3.21113139e-01 5.85438550e-01 -9.53604460e-01
1.33629358e+00 -1.84400678e+00 8.31486523e-01 3.20761412e-01
9.27915573e-01 2.95784593e-01 -3.54397476e-01 9.46594596e-01
2.53630225e-02 2.81068444e-01 -3.79784763e-01 -4.84632879e-01
1.14823274e-01 3.73848379e-01 2.85341084e-01 5.80342352e-01
5.21203041e-01 1.06034255e+00 -8.84355783e-01 -7.05476999e-01
1.86486155e-01 8.18633378e-01 -1.05522799e+00 -1.54319284e-02
-1.89589456e-01 2.67175853e-01 -1.74242437e-01 3.79583865e-01
3.81241769e-01 -8.03710818e-01 7.09485173e-01 -2.08696365e-01
1.37127444e-01 -1.01653367e-01 -1.23494911e+00 1.22989035e+00
-1.77106351e-01 8.77109289e-01 3.90863270e-01 -1.46454322e+00
8.23033512e-01 4.42350954e-01 7.45479584e-01 -1.71058208e-01
1.14395902e-01 -7.80670196e-02 8.79852399e-02 -1.08994469e-01
3.48559052e-01 -1.62586793e-02 3.04526895e-01 2.98580647e-01
3.96644056e-01 2.97446460e-01 4.43755478e-01 6.60638034e-01
1.31679261e+00 -7.38456249e-02 5.31230152e-01 -3.39982718e-01
6.45167410e-01 -2.60048479e-01 2.71546692e-01 4.30283546e-01
-2.54829079e-01 4.31803584e-01 8.83360565e-01 -1.38588428e-01
-1.31094968e+00 -8.91408563e-01 1.60868138e-01 8.92839432e-01
-8.21843967e-02 -7.98324168e-01 -6.88533425e-01 -5.79514802e-01
1.26032919e-01 2.18017668e-01 -7.48514175e-01 -4.51561995e-02
-3.58530432e-01 -7.06779718e-01 3.30589563e-01 2.63246059e-01
1.12209350e-01 -7.59591520e-01 -9.72460508e-02 4.61564690e-01
-1.70713112e-01 -1.02621949e+00 -1.32972836e-01 4.27436642e-02
-8.02060843e-01 -1.16927826e+00 -8.56638193e-01 -8.11032534e-01
3.78315359e-01 2.00430483e-01 1.25841475e+00 4.81262594e-01
-9.32016820e-02 4.86351907e-01 -5.73497415e-01 8.63010958e-02
-5.74365377e-01 1.82218105e-01 -5.81419580e-02 2.26526260e-01
5.70455454e-02 -1.00162697e+00 -4.19673085e-01 -1.73881561e-01
-6.71678007e-01 8.38388409e-03 6.82492375e-01 1.12216163e+00
2.98765361e-01 1.14783019e-01 5.07670701e-01 -8.33852589e-01
7.32734084e-01 -7.00339496e-01 -2.73625940e-01 3.14041704e-01
-8.55273366e-01 1.99650735e-01 6.39351249e-01 -1.76398754e-01
-3.18171918e-01 -2.39523739e-01 -1.75938472e-01 -5.93584478e-01
2.43607491e-01 8.07242930e-01 1.12369694e-01 -2.56484151e-01
4.85356450e-01 1.75661072e-01 2.33272240e-01 -3.96778882e-01
4.38699037e-01 4.37481701e-01 2.07139403e-01 -3.56091619e-01
8.63861501e-01 4.18361306e-01 1.23806810e-02 -9.78236139e-01
-1.82101086e-01 -4.92240220e-01 -6.49215817e-01 -3.01859468e-01
9.49791312e-01 -9.80004609e-01 -6.98919535e-01 -4.43962924e-02
-1.21646070e+00 -9.08540115e-02 1.23109728e-01 2.87319243e-01
-3.67520332e-01 8.40274036e-01 -7.40480423e-01 -6.95100009e-01
-3.63276303e-01 -8.52831662e-01 8.95759761e-01 -1.40528426e-01
-1.16366126e-01 -1.52954292e+00 2.50739902e-01 2.05215603e-01
3.70185167e-01 4.22100008e-01 1.05899751e+00 -6.84764862e-01
-5.75904608e-01 2.71842675e-03 -3.88101250e-01 2.26906851e-01
-2.31463745e-01 3.54410678e-01 -6.38255298e-01 -5.99508047e-01
-7.18529344e-01 1.51064783e-01 9.75293577e-01 3.92805099e-01
7.77802885e-01 -4.87625450e-01 -4.11130399e-01 4.14124072e-01
1.74378383e+00 -4.66674119e-01 4.48937744e-01 1.77403033e-01
8.96134436e-01 5.79512656e-01 -2.25388423e-01 5.09009957e-01
6.52153850e-01 6.15593493e-01 5.89436829e-01 1.85472127e-02
-3.13674152e-01 -3.03871244e-01 1.44501254e-01 1.40769327e+00
-4.43051696e-01 -6.04157150e-01 -9.32059586e-01 8.74077499e-01
-1.90168631e+00 -9.82890427e-01 -3.88277173e-01 1.64877975e+00
4.82475579e-01 -5.85787185e-02 4.29599822e-01 2.96349913e-01
8.62505555e-01 2.88255155e-01 2.33741589e-02 -2.72633374e-01
-3.00800830e-01 1.41956851e-01 2.84585118e-01 6.27594709e-01
-1.19862187e+00 5.73210895e-01 5.37315369e+00 5.75931668e-01
-5.73172927e-01 3.61521453e-01 2.24903315e-01 3.57109725e-01
-6.18481994e-01 8.59221295e-02 -1.81781903e-01 2.28177488e-01
9.32144821e-01 -9.44075435e-02 5.63851058e-01 6.73902392e-01
-2.27232277e-01 4.21302468e-01 -9.44086432e-01 7.34723926e-01
8.90550166e-02 -1.54164231e+00 -1.02436528e-01 3.12572181e-01
8.07308018e-01 7.14161843e-02 -2.72086203e-01 4.86943334e-01
4.57569420e-01 -8.43375385e-01 3.01195323e-01 3.83778900e-01
2.14326665e-01 -6.66372240e-01 7.27024138e-01 2.41858140e-01
-1.28670633e+00 -2.90845279e-02 -4.55507100e-01 1.34540517e-02
-3.62349041e-02 7.27549493e-01 -7.51900971e-01 9.90817189e-01
4.51497585e-01 9.93097842e-01 -6.85816109e-01 1.02467597e+00
-1.36566237e-01 6.99389279e-01 -2.46007234e-01 -2.43080080e-01
2.02046677e-01 -2.98612297e-01 9.42965686e-01 1.14764094e+00
1.75261080e-01 -3.27616692e-01 4.71681692e-02 1.13661623e+00
-2.23021418e-01 3.02974910e-01 -6.46487832e-01 -3.94457400e-01
4.35066611e-01 1.27829993e+00 -6.77964509e-01 -1.37470722e-01
-5.42250276e-01 1.02573562e+00 7.72696912e-01 4.07173455e-01
-6.26156211e-01 -2.41695747e-01 5.30707538e-01 5.68945967e-02
7.31385052e-01 -3.26096028e-01 2.22422555e-02 -1.35593462e+00
-1.10150605e-01 -5.70063293e-01 5.81095695e-01 -3.95852119e-01
-1.43652570e+00 5.00509501e-01 -2.58533865e-01 -9.98383820e-01
-1.86846256e-01 -6.72443449e-01 -6.21298969e-01 4.57230240e-01
-1.41043103e+00 -1.32117522e+00 1.32695392e-01 3.88596863e-01
-3.98571510e-03 -1.62431747e-01 8.15626323e-01 5.16126096e-01
-6.31714523e-01 4.86914188e-01 5.77668071e-01 2.20807359e-01
2.32703269e-01 -1.42278981e+00 2.13601395e-01 8.30742419e-01
6.30407035e-01 4.70406413e-01 7.81805933e-01 -7.22425759e-01
-1.38523543e+00 -1.12134385e+00 1.15899491e+00 -2.56488979e-01
8.90266538e-01 -5.37864447e-01 -1.07606912e+00 7.71116436e-01
4.24234033e-01 -9.25248936e-02 5.56921780e-01 4.58417684e-01
-2.44376913e-01 2.89409459e-01 -7.71547854e-01 5.95883965e-01
1.02575231e+00 -5.73243260e-01 -3.63673717e-01 4.32646066e-01
7.24852681e-01 3.17443371e-01 -1.43133593e+00 2.15594992e-01
2.64880538e-01 -1.10508323e+00 1.18808925e+00 -4.78477478e-01
6.33592248e-01 -1.25028789e-01 -1.45430371e-01 -1.40976000e+00
-7.35413194e-01 -3.30038905e-01 -5.67255676e-01 1.13802361e+00
1.76110208e-01 -6.06499374e-01 8.38447690e-01 -3.14985275e-01
1.50612891e-01 -7.04903424e-01 -9.92147088e-01 -6.81311429e-01
2.05311239e-01 -1.11404411e-01 4.11218256e-01 1.21198273e+00
3.27979960e-02 4.45703894e-01 -5.18626630e-01 2.77408749e-01
9.55707431e-01 1.25465065e-01 6.99901044e-01 -1.64033663e+00
-5.90058684e-01 -7.19938278e-01 -7.05664456e-01 -4.83529866e-01
2.86725521e-01 -1.25899482e+00 -3.82088721e-01 -1.76992118e+00
3.85098070e-01 3.34429741e-02 -2.92024434e-01 2.83400476e-01
-2.58714080e-01 1.49893031e-01 2.75659233e-01 1.46349758e-01
-7.19746768e-01 8.11247528e-01 1.07028759e+00 -3.19375068e-01
1.01042829e-01 -3.03024352e-01 -5.32178700e-01 2.87411004e-01
5.81617713e-01 -3.33965838e-01 -1.48816332e-01 -2.85329551e-01
4.67636853e-01 5.76496758e-02 6.25207365e-01 -9.66004193e-01
4.01190370e-01 3.91984999e-01 1.41024739e-01 -3.75302702e-01
6.03040569e-02 -5.50603449e-01 3.84244829e-01 6.70961738e-01
-2.27420866e-01 -1.59537658e-01 -6.06892891e-02 1.12470448e+00
-2.36888424e-01 -9.16853026e-02 3.91669154e-01 2.62805186e-02
-6.50210202e-01 3.59422535e-01 -4.72208470e-01 -3.08844835e-01
8.60488534e-01 -8.46767128e-02 -2.04242751e-01 -7.08235919e-01
-1.00267649e+00 3.68388116e-01 4.31999147e-01 2.29233101e-01
4.76275444e-01 -1.41203761e+00 -1.17231333e+00 7.75059089e-02
1.29514024e-01 -4.11275744e-01 2.89988637e-01 1.13556850e+00
-5.05596101e-01 2.81943589e-01 -1.32137939e-01 -4.75680023e-01
-1.25058508e+00 9.01916683e-01 1.64082944e-01 -6.25486314e-01
-6.50231957e-01 7.58713484e-01 -1.25168622e-01 -6.34339154e-01
9.26697254e-02 1.22371867e-01 -3.72416466e-01 1.66532770e-01
3.68994251e-02 3.90235484e-01 -6.88927174e-02 -6.23478830e-01
-3.78292620e-01 3.16671252e-01 2.10657656e-01 2.55596727e-01
1.63715160e+00 -2.27105230e-01 -3.78156215e-01 1.05929680e-01
1.37691772e+00 2.19465699e-02 -9.00250971e-01 -2.83869803e-01
1.74712241e-01 4.04227190e-02 1.90414935e-01 -2.27098644e-01
-1.15960562e+00 1.02715898e+00 2.85989285e-01 6.64204538e-01
7.42571235e-01 1.40974775e-01 5.06582201e-01 1.74054697e-01
3.03265274e-01 -8.59522998e-01 3.23310941e-02 2.36548662e-01
8.22544992e-01 -1.17202318e+00 1.37457758e-01 -4.74597454e-01
-4.39763129e-01 1.05532575e+00 1.60388723e-01 -2.89758086e-01
7.11924434e-01 -1.17731318e-01 -5.93272865e-01 -3.84483874e-01
-9.42344785e-01 -5.24491549e-01 4.38228935e-01 7.32298672e-01
3.61467063e-01 3.08808267e-01 -4.07100856e-01 4.78080451e-01
5.77464737e-02 -4.57127631e-01 5.51254332e-01 6.21747315e-01
-3.71305287e-01 -1.18496001e+00 -2.38412380e-01 5.47562897e-01
-1.31876469e-01 -7.92024061e-02 -6.45883739e-01 9.28176045e-01
-2.96042543e-02 8.08799028e-01 2.99057104e-02 -6.89679861e-01
-1.43443584e-01 1.07036993e-01 3.80375981e-01 -4.72426027e-01
-5.54591477e-01 1.57578602e-01 3.75041693e-01 -2.62745976e-01
-6.24400198e-01 -7.02631533e-01 -6.87610984e-01 -7.23317027e-01
-4.90635812e-01 1.85970142e-01 4.17916745e-01 6.73623085e-01
6.04697883e-01 7.09796846e-01 4.93726104e-01 -7.72116065e-01
-3.60191405e-01 -9.12546158e-01 -6.90406263e-01 4.27727699e-01
2.68293083e-01 -7.11906433e-01 -3.66271853e-01 -2.19048753e-01] | [7.1987624168396, 6.099703788757324] |
cc0db26e-a445-4f2d-a2ad-9956d7d7dde4 | graph-based-aspect-representation-learning | null | null | https://aclanthology.org/2020.textgraphs-1.2 | https://aclanthology.org/2020.textgraphs-1.2.pdf | Graph-based Aspect Representation Learning for Entity Resolution | Entity Resolution (ER) identifies records that refer to the same real-world entity. Deep learning approaches improved the generalization ability of entity matching models, but hardly overcame the impact of noisy or incomplete data sources. In real scenes, an entity usually consists of multiple semantic facets, called aspects. In this paper, we focus on entity augmentation, namely retrieving the values of missing aspects. The relationship between aspects is naturally suitable to be represented by a knowledge graph, where entity augmentation can be modeled as a link prediction problem. Our paper proposes a novel graph-based approach to solve entity augmentation. Specifically, we apply a dedicated random walk algorithm, which uses node types to limit the traversal length, and encodes graph structure into low-dimensional embeddings. Thus, the missing aspects could be retrieved by a link prediction model. Furthermore, the augmented aspects with fixed orders are served as the input of a deep Siamese BiLSTM network for entity matching. We compared our method with state-of-the-art methods through extensive experiments on downstream ER tasks. According to the experiment results, our model outperforms other methods on evaluation metrics (accuracy, precision, recall, and f1-score) to a large extent, which demonstrates the effectiveness of our method. | ['Bin Gu', 'Xiangnan He', 'Yufan Huang', 'Dingxian Wang', 'Yuchen Guo', 'Zhenqi Zhao'] | null | null | null | null | coling-textgraphs-2020-12 | ['entity-resolution'] | ['natural-language-processing'] | [-2.55250931e-01 4.14825201e-01 -3.93146425e-01 -2.82044470e-01
-5.15212476e-01 -2.34057009e-01 5.61519504e-01 5.17066777e-01
-5.72213113e-01 6.56283617e-01 4.31505054e-01 -5.78308702e-02
-2.85326153e-01 -1.34994209e+00 -1.05915558e+00 -3.37015837e-01
-4.90682013e-02 7.17275620e-01 2.53790021e-01 -2.69442201e-01
-1.37002766e-01 2.05226809e-01 -1.14929295e+00 4.00665365e-02
1.05492532e+00 8.14335644e-01 9.53992158e-02 -1.19444415e-01
-6.78812504e-01 5.88279307e-01 -4.91805226e-01 -9.18792486e-01
-1.10270254e-01 -8.78952146e-02 -6.47055387e-01 -3.41430902e-01
1.53045535e-01 -2.20992312e-01 -8.46405029e-01 1.16267538e+00
4.57645863e-01 1.93367496e-01 3.57242733e-01 -1.35547674e+00
-1.19818139e+00 9.80759859e-01 -5.31351328e-01 9.06159058e-02
2.75252491e-01 -1.82937101e-01 1.31694019e+00 -9.08642054e-01
8.01195323e-01 1.29033935e+00 6.97802424e-01 2.97783703e-01
-1.06615067e+00 -8.33255470e-01 4.49363232e-01 5.14535546e-01
-1.48504150e+00 -1.16721787e-01 6.99414551e-01 -7.80442730e-02
8.60084116e-01 -5.74501697e-03 6.53623521e-01 9.25194979e-01
-2.75255859e-01 9.89158630e-01 4.08478707e-01 6.12139329e-03
-7.29190707e-02 1.59252867e-01 2.32948452e-01 7.33434439e-01
7.83001721e-01 -5.61364405e-02 -2.61731476e-01 -2.70424634e-01
6.50633931e-01 1.44361779e-01 -4.28288519e-01 -6.02301478e-01
-1.34140134e+00 6.44902706e-01 1.10412240e+00 3.14197809e-01
-4.64283645e-01 4.61328365e-02 3.91117066e-01 1.35508552e-01
3.59044164e-01 4.53692555e-01 -5.04351079e-01 4.44256723e-01
-4.27173823e-01 2.20977888e-01 6.65802360e-01 1.18857872e+00
9.55469966e-01 -3.62535954e-01 -4.39006716e-01 6.98322833e-01
5.41569412e-01 4.65212047e-01 3.31454396e-01 -5.02787411e-01
9.63160098e-01 1.20360279e+00 1.82834402e-01 -1.34180713e+00
-4.78436291e-01 -6.81271613e-01 -1.04587567e+00 -6.12128079e-01
1.52072636e-02 -1.06068701e-01 -8.92801046e-01 1.94876719e+00
5.75571060e-01 5.63680053e-01 2.16262013e-01 1.02681351e+00
1.21757782e+00 7.46211052e-01 4.95359242e-01 6.56832615e-03
1.45306957e+00 -1.00114179e+00 -9.54802811e-01 -1.28674328e-01
6.90919816e-01 -3.50584954e-01 8.78833115e-01 -2.19924584e-01
-7.31044054e-01 -4.44734871e-01 -8.90293479e-01 -2.27043316e-01
-5.80922723e-01 2.42252693e-01 8.42349648e-01 1.60935789e-01
-7.23149002e-01 6.96478128e-01 -6.57675028e-01 -2.10955203e-01
4.81261492e-01 2.70394206e-01 -6.81322753e-01 -2.51449615e-01
-1.80008554e+00 6.20088458e-01 6.60067499e-01 2.37784550e-01
-3.75269294e-01 -7.17139423e-01 -1.12864292e+00 5.46178222e-01
6.43231273e-01 -1.03743148e+00 8.04773867e-01 -2.49869063e-01
-6.14536643e-01 6.26783907e-01 -2.90397584e-01 -3.26633930e-01
2.05191776e-01 -3.30522031e-01 -7.73145080e-01 -1.56443473e-02
3.07600051e-01 4.24190313e-01 3.04466009e-01 -1.24924874e+00
-7.13368237e-01 -4.84294176e-01 3.76813591e-01 2.87590712e-01
-4.87390339e-01 -4.12755430e-01 -9.05541301e-01 -6.53381944e-01
1.37047082e-01 -7.43816793e-01 -2.38035202e-01 -7.37543106e-02
-4.12489861e-01 -4.74420279e-01 4.52388495e-01 -7.92744756e-01
1.46779776e+00 -2.04830480e+00 1.26682565e-01 1.50716648e-01
4.70931649e-01 3.80714774e-01 -2.77802467e-01 4.02111799e-01
5.38924001e-02 3.17485303e-01 -2.94371992e-01 -1.75385267e-01
1.87828705e-01 1.10119112e-01 -3.14085990e-01 8.72934908e-02
1.41683415e-01 1.32746041e+00 -9.72571909e-01 -5.06538570e-01
-2.41670370e-01 5.43567896e-01 -3.15057874e-01 3.01300734e-01
-2.94433236e-01 5.55551276e-02 -7.20024049e-01 5.53159297e-01
8.63909066e-01 -3.92174363e-01 1.82472631e-01 -5.08403718e-01
3.99923205e-01 4.37476903e-01 -1.22233438e+00 1.85991383e+00
-3.51478726e-01 2.53118455e-01 -2.86829621e-01 -9.07896280e-01
1.02600265e+00 1.62225693e-01 2.69157559e-01 -8.84083390e-01
-2.41045982e-01 2.30184302e-01 -2.51826823e-01 -5.18526137e-01
6.29529178e-01 3.60378504e-01 -8.67287442e-02 5.61425015e-02
2.54882853e-02 7.36456275e-01 2.75785148e-01 4.38980907e-01
9.76022601e-01 5.87613508e-02 1.67621955e-01 6.43529817e-02
5.92440426e-01 2.47523990e-02 7.58476138e-01 5.37634134e-01
2.23547727e-01 1.37624949e-01 5.96694827e-01 -2.85721570e-01
-9.25091028e-01 -1.01707625e+00 6.10498041e-02 8.71994615e-01
7.65463412e-01 -4.67582226e-01 -4.63178515e-01 -8.72941792e-01
2.78258681e-01 6.52103245e-01 -6.48361385e-01 -4.83496577e-01
-7.27749765e-01 -7.87482738e-01 4.59331721e-01 7.50644028e-01
6.83910668e-01 -1.13731539e+00 2.17007115e-01 3.33318412e-01
-3.80312204e-01 -1.18037570e+00 -5.42686164e-01 -2.16288835e-01
-7.47195542e-01 -1.20728970e+00 -7.22612560e-01 -8.75447631e-01
6.38457775e-01 2.22032130e-01 1.10305619e+00 2.69105077e-01
1.09945573e-01 4.14694138e-02 -2.27969959e-01 -9.35427845e-02
1.66199714e-01 4.63843197e-01 -1.22634277e-01 1.41134918e-01
7.41642177e-01 -5.72505414e-01 -8.24675441e-01 2.80055970e-01
-8.86268198e-01 -1.95190728e-01 9.64593530e-01 8.52900267e-01
8.91156375e-01 -1.08914629e-01 7.26168215e-01 -1.24036670e+00
6.73613906e-01 -8.89823020e-01 -4.44923639e-01 4.74693120e-01
-8.68457139e-01 3.01568359e-01 5.83810806e-01 -3.91590208e-01
-1.11140144e+00 -2.41589561e-01 -6.43096566e-02 -5.30910671e-01
1.21107027e-01 8.99450779e-01 -7.32320428e-01 3.63955885e-01
1.62561595e-01 1.92827761e-01 -5.03361285e-01 -7.05216110e-01
5.44432282e-01 4.62416083e-01 5.71406424e-01 -5.29061377e-01
8.51615965e-01 2.38579974e-01 3.39271352e-02 -2.29423910e-01
-9.98507321e-01 -4.61466551e-01 -3.10501039e-01 3.02057564e-01
6.04692400e-01 -1.18895245e+00 -6.50064945e-01 2.12136686e-01
-1.20734584e+00 2.23906904e-01 -1.97530806e-01 5.26285768e-01
3.59663852e-02 3.24916631e-01 -6.91018522e-01 -3.68068308e-01
-5.23636758e-01 -7.36169279e-01 1.03422821e+00 5.69709480e-01
2.27025479e-01 -8.20462584e-01 1.21990308e-01 2.60246992e-01
3.56401116e-01 1.15700789e-01 1.17919886e+00 -1.07368922e+00
-9.26452994e-01 -2.61039048e-01 -6.86417401e-01 -3.12589735e-01
-9.89203155e-02 -4.01085138e-01 -7.04768777e-01 -1.40126958e-01
-6.09535575e-01 1.61885515e-01 1.07649040e+00 -1.28070086e-01
1.11351490e+00 -5.04050970e-01 -8.05503309e-01 6.75694048e-01
1.36486697e+00 -6.49054348e-02 6.97134256e-01 5.70699632e-01
1.00931084e+00 4.31525946e-01 8.30482304e-01 1.40911922e-01
9.28188622e-01 6.41350448e-01 5.42663634e-01 -8.56393948e-02
-8.01207870e-02 -9.06359494e-01 -2.29014680e-01 7.46445417e-01
6.77797347e-02 -5.09884059e-01 -7.30778277e-01 8.38898540e-01
-2.12566090e+00 -8.72708678e-01 -3.09450567e-01 2.05437350e+00
5.77231884e-01 1.65248021e-01 -1.82243839e-01 -2.33731940e-01
1.00496209e+00 1.60044134e-01 -7.94275641e-01 2.76367038e-01
-1.16934009e-01 -1.33554503e-01 4.32814270e-01 1.11337975e-01
-1.18541574e+00 8.93198669e-01 4.27326822e+00 8.30902159e-01
-5.47574937e-01 -8.86945706e-03 2.34194890e-01 3.23811144e-01
-8.08763444e-01 7.16874525e-02 -9.64327753e-01 5.81865251e-01
5.36840737e-01 -6.30038679e-01 1.87240362e-01 8.68860483e-01
-3.02563399e-01 5.18498659e-01 -1.05698359e+00 7.06113398e-01
-1.12474076e-01 -1.29124343e+00 3.64555538e-01 4.21579294e-02
5.80189586e-01 3.28928307e-02 -1.54805049e-01 8.61691594e-01
1.89345583e-01 -7.75840104e-01 1.45642072e-01 7.27539837e-01
6.22763157e-01 -7.61790216e-01 1.03452718e+00 2.16405243e-01
-1.64749873e+00 -1.80631857e-02 -6.54044330e-01 5.54792702e-01
1.49111122e-01 6.16652846e-01 -4.54401672e-01 9.86821413e-01
7.10186899e-01 8.06918085e-01 -5.23901701e-01 1.32819211e+00
-5.04495203e-01 2.92786121e-01 -3.35312009e-01 2.78949738e-04
7.78223798e-02 -2.38247320e-01 4.97342080e-01 1.08772147e+00
4.06844378e-01 2.13772580e-01 9.18670818e-02 8.67487669e-01
-8.25380385e-01 3.11523199e-01 -6.12699091e-01 -3.08433473e-02
9.58688796e-01 1.31341112e+00 -2.47999176e-01 -3.98546070e-01
-5.52251220e-01 7.85875618e-01 8.21806252e-01 4.52132314e-01
-7.66009748e-01 -8.31753433e-01 5.51801860e-01 2.10259646e-01
5.69530368e-01 2.24377662e-01 -9.51648597e-03 -1.47504675e+00
4.14104015e-01 -4.61240292e-01 7.93198109e-01 -7.00507820e-01
-1.37277460e+00 6.55489504e-01 -1.73949763e-01 -1.12374485e+00
5.10456078e-02 -6.75345287e-02 -6.26616955e-01 7.82358408e-01
-1.86460674e+00 -1.18736696e+00 -6.32302940e-01 3.93714726e-01
-4.59801406e-02 -1.80937663e-01 7.01754689e-01 7.88233936e-01
-6.03154182e-01 8.28948259e-01 1.39398217e-01 4.54399228e-01
6.86535180e-01 -1.16669726e+00 7.31309652e-01 7.15134144e-01
2.01581106e-01 7.97380030e-01 3.16065997e-01 -8.55396509e-01
-1.33199584e+00 -1.42314577e+00 1.20591748e+00 -1.68024957e-01
7.63701439e-01 -8.29799026e-02 -1.44223988e+00 9.40570772e-01
-5.75545281e-02 1.12715594e-01 4.54770714e-01 3.12804818e-01
-4.57001597e-01 -2.18759477e-01 -9.71008956e-01 5.80866218e-01
1.38820708e+00 -4.00389940e-01 -8.14325154e-01 5.07894531e-02
1.30135989e+00 -3.48863631e-01 -1.08948493e+00 7.55266011e-01
2.99222380e-01 -3.83383334e-01 1.01493859e+00 -9.47298229e-01
4.22508717e-01 -4.20510024e-01 5.88557264e-03 -1.38139057e+00
-4.47233289e-01 -4.06094268e-02 -7.62787342e-01 1.62032712e+00
6.11509621e-01 -6.23846531e-01 7.71918178e-01 6.50255144e-01
6.88510537e-02 -9.49163139e-01 -6.89330399e-01 -5.56450963e-01
-2.79503584e-01 1.12569094e-01 1.24687099e+00 1.21765041e+00
-1.80462405e-01 6.22460067e-01 -1.25124991e-01 5.68522334e-01
6.56395257e-01 5.24991333e-01 5.84631145e-01 -1.42229521e+00
-1.05558790e-01 -3.18511784e-01 -4.66117144e-01 -1.19345307e+00
3.52914274e-01 -1.04553246e+00 -4.61078972e-01 -1.98646951e+00
3.19564730e-01 -6.92020357e-01 -5.70332408e-01 5.39283037e-01
-7.22297013e-01 -1.37467235e-01 3.10333185e-02 2.70640969e-01
-7.63718128e-01 9.78420556e-01 1.16804218e+00 -3.88192147e-01
-1.47798538e-01 -4.31493148e-02 -9.15037215e-01 4.80507702e-01
6.00971162e-01 -5.84183693e-01 -3.24544668e-01 -7.90933967e-01
5.17850339e-01 1.20133713e-01 1.71426758e-01 -6.05488062e-01
5.24129212e-01 1.04950011e-01 3.10124755e-01 -5.84033430e-01
2.62502760e-01 -1.00237036e+00 1.93560332e-01 2.05941409e-01
-4.56883788e-01 -4.95842062e-02 -5.94322011e-03 9.87050235e-01
-4.10880119e-01 -1.22853838e-01 1.22934155e-01 2.46577840e-02
-9.03881371e-01 8.22193742e-01 4.37702715e-01 2.68323243e-01
8.22513163e-01 3.46452266e-01 -6.92435741e-01 -1.38799891e-01
-7.87708163e-01 8.51960659e-01 2.57955939e-01 6.08268142e-01
5.24946570e-01 -1.78505003e+00 -7.42704213e-01 2.00098641e-02
4.58297908e-01 3.28817189e-01 3.54788691e-01 6.12100124e-01
-1.33686408e-01 3.08495343e-01 1.50741369e-01 -2.28282407e-01
-9.04614210e-01 9.13254559e-01 2.92054176e-01 -6.69769287e-01
-6.42771363e-01 5.95367730e-01 3.00487071e-01 -7.53818154e-01
2.86044776e-01 1.23938002e-01 -5.79608381e-01 1.89310446e-01
3.75745356e-01 2.46850237e-01 4.02318165e-02 -5.07377028e-01
-3.59463394e-01 4.38968778e-01 -3.36246938e-01 3.93981665e-01
1.36339045e+00 -2.48026341e-01 -6.32777289e-02 1.07021764e-01
1.21543849e+00 -4.42854911e-02 -6.07775390e-01 -7.99011588e-01
1.94382608e-01 -3.68996859e-01 -1.23938292e-01 -6.30381048e-01
-1.31016743e+00 7.12625802e-01 2.55672723e-01 1.27902806e-01
8.99069071e-01 2.06220210e-01 1.21846092e+00 7.70976424e-01
2.58372366e-01 -7.83331990e-01 -4.84795481e-01 1.89330235e-01
5.55663705e-01 -1.22461331e+00 -1.49024248e-01 -5.22714615e-01
-6.42001688e-01 6.85282707e-01 1.01441967e+00 -5.74464388e-02
5.49853683e-01 -1.93230316e-01 -2.62668818e-01 -3.25108320e-01
-6.64852321e-01 -4.41925645e-01 4.49623615e-01 4.29480582e-01
1.79545715e-01 5.60680851e-02 -4.25562501e-01 1.08695483e+00
3.43357474e-02 -4.25369330e-02 1.63827911e-01 4.23583925e-01
-4.27778035e-01 -8.50774765e-01 1.41068488e-01 6.06791914e-01
-4.11493540e-01 -2.05920637e-01 -1.17811784e-01 8.35510314e-01
1.52478730e-02 5.02645195e-01 -8.64089876e-02 -3.81098032e-01
7.56153464e-01 -3.73100080e-02 -1.12872824e-01 -3.50987762e-01
-2.78354436e-01 -3.93821478e-01 1.50527820e-01 -5.53395629e-01
-1.85445592e-01 -2.91788757e-01 -1.44085646e+00 -2.29710415e-01
-4.14543957e-01 4.69033569e-01 4.57903862e-01 8.55291247e-01
7.21823990e-01 6.79931283e-01 5.54682016e-01 -2.26098567e-01
-3.78579885e-01 -9.10380244e-01 -5.51023006e-01 7.11795747e-01
1.27829984e-01 -6.62520528e-01 -1.91968173e-01 -4.44156617e-01] | [8.947144508361816, 8.177873611450195] |
b7c75cd1-4e0d-4567-b59d-8ddba8de40a9 | using-machine-learning-methods-for-automation | 2306.09775 | null | https://arxiv.org/abs/2306.09775v1 | https://arxiv.org/pdf/2306.09775v1.pdf | Using Machine Learning Methods for Automation of Size Grid Building and Management | Fashion apparel companies require planning for the next season, a year in advance for supply chain management. This study focuses on size selection decision making for Levi Strauss. Currently, the region and planning group level size grids are built and managed manually. The company suffers from the workload it creates for sizing, merchant and planning teams. This research is aiming to answer two research questions: "Which sizes should be available to the planners under each size grid name for the next season(s)?" and "Which sizes should be adopted for each planning group for the next season(s)?". We approach to the problem with a classification model, which is one of the popular models used in machine learning. With this research, a more automated process was created by using machine learning techniques. A decrease in workload of the teams in the company is expected after it is put into practice. Unlike many studies in the state of art for fashion and apparel industry, this study focuses on sizes where the stock keeping unit represents a product with a certain size. | ['Filipa Peleja', 'Dries Benoit', 'Salim Yunus'] | 2023-06-16 | null | null | null | null | ['management'] | ['miscellaneous'] | [-1.61082953e-01 2.90456321e-02 -3.78352404e-01 -3.82979155e-01
1.38331190e-01 -4.52796698e-01 4.23442051e-02 5.22351444e-01
-2.90093511e-01 7.19697833e-01 -1.49158090e-02 -3.00979912e-01
-4.66223210e-01 -1.12825704e+00 -4.42123920e-01 -5.71050882e-01
7.88328201e-02 1.22384477e+00 -1.95854008e-01 -3.65016550e-01
7.57079184e-01 6.52591884e-01 -1.25534117e+00 3.36889625e-01
4.41961944e-01 1.14189577e+00 7.22080588e-01 2.18065798e-01
-2.72091240e-01 9.83513534e-01 -4.45066988e-01 -1.54527605e-01
7.96773672e-01 -5.41663349e-01 -8.05099726e-01 7.85994709e-01
2.64286324e-02 -1.29681835e-02 5.21465421e-01 7.91879356e-01
-3.84921534e-03 1.32583618e-01 7.67353058e-01 -1.43786108e+00
-1.94268018e-01 1.22633147e+00 -4.46203947e-01 1.34953693e-01
2.69460857e-01 1.12654455e-02 9.81249869e-01 -4.83711541e-01
7.51082242e-01 9.54887092e-01 2.98503339e-01 -2.82230973e-01
-1.14099216e+00 -8.51012766e-01 1.97960824e-01 3.17473322e-01
-1.48760474e+00 2.00556844e-01 6.55011177e-01 -7.87054718e-01
9.54830587e-01 1.45577863e-01 1.09257305e+00 1.41309768e-01
7.59455502e-01 1.46142244e-01 1.34732330e+00 -5.97834587e-01
3.61798227e-01 6.94183230e-01 -1.37375176e-01 -9.30967629e-02
4.32083100e-01 -2.13072836e-01 -7.94670582e-02 4.92513508e-01
1.06238139e+00 -2.50000674e-02 5.04820883e-01 5.09917997e-02
-8.77558649e-01 1.29836345e+00 4.73091155e-02 6.95971966e-01
-8.00034046e-01 -2.38708183e-01 3.29528302e-01 6.87873960e-01
2.87640631e-01 1.03070283e+00 -6.71811581e-01 -1.39641732e-01
-1.12957263e+00 3.34507287e-01 1.31949663e+00 1.27082670e+00
5.82338750e-01 -2.33055428e-01 3.17614734e-01 2.80508995e-01
3.71725708e-01 -4.42656130e-02 -5.02186269e-02 -7.49074340e-01
4.01158601e-01 9.15596068e-01 1.65976092e-01 -9.12068844e-01
-7.54501283e-01 -5.29977500e-01 -5.12960613e-01 2.85031021e-01
6.82636321e-01 -2.28039891e-01 -6.27367735e-01 6.71830118e-01
1.34534419e-01 -5.02712846e-01 -2.61260867e-01 7.02176690e-01
2.70618707e-01 6.94316864e-01 -1.16505604e-02 -7.69293904e-01
1.49740183e+00 -9.60373819e-01 -9.43717480e-01 -1.40613923e-02
4.92832392e-01 -1.33117747e+00 5.75140476e-01 1.12907267e+00
-1.07269347e+00 -6.75837934e-01 -7.26387143e-01 4.81707990e-01
-3.90455067e-01 3.54449540e-01 7.61860251e-01 5.51090419e-01
-6.00510120e-01 7.08170891e-01 -4.08617944e-01 -5.27330220e-01
-4.13690120e-01 4.10288602e-01 -1.51786268e-01 2.97385380e-02
-7.57457852e-01 1.32138216e+00 4.55184251e-01 1.98261321e-01
-5.67811072e-01 -4.24839139e-01 -5.05670071e-01 -4.21031006e-03
6.44308686e-01 -3.10883999e-01 1.20049798e+00 -1.01098573e+00
-1.22234941e+00 5.02762496e-01 4.87408668e-01 -2.43552104e-01
4.97311831e-01 1.50111586e-01 -3.81101459e-01 -3.74357074e-01
1.74761921e-01 2.05373570e-01 3.16800922e-01 -1.30210447e+00
-1.36797380e+00 -3.15315872e-01 2.28597686e-01 1.00097872e-01
2.03531772e-01 4.08874571e-01 1.77673563e-01 -5.94159424e-01
3.57690006e-01 -1.09267974e+00 -5.80498159e-01 -9.85464811e-01
-1.63375035e-01 -3.90665740e-01 6.80267587e-02 -7.43356407e-01
1.61814964e+00 -1.54483783e+00 2.25423232e-01 4.56719339e-01
-2.17783973e-01 -5.11291742e-01 4.44498420e-01 1.03065825e+00
-4.40464541e-02 4.00341265e-02 5.00260830e-01 3.27532172e-01
2.73006260e-02 1.84895039e-01 3.66744131e-01 3.08191210e-01
-1.53592914e-01 -1.10829927e-01 -3.10415268e-01 -4.77052629e-01
1.08129047e-01 -4.76200551e-01 -3.48325014e-01 1.18906982e-01
-2.15334579e-01 2.66305625e-01 -3.88298541e-01 8.32345366e-01
1.02871382e+00 -3.21689295e-03 4.84561712e-01 -3.84382516e-01
-8.06427479e-01 -4.90777101e-03 -1.63917887e+00 1.06306148e+00
-7.28792012e-01 -8.96945521e-02 2.43085936e-01 -9.59095240e-01
1.39863741e+00 2.16637045e-01 6.17874742e-01 -4.47200626e-01
5.40946066e-01 2.96420455e-01 4.84302044e-01 -6.52370572e-01
6.28441870e-01 -1.98199421e-01 -2.65840054e-01 3.70109051e-01
-2.40915775e-01 -3.14447433e-01 9.79252934e-01 -4.25379485e-01
7.25145042e-01 1.07197583e-01 5.56503952e-01 -4.87878770e-01
2.96707451e-01 4.49376017e-01 8.18351984e-01 2.46789828e-01
5.57438619e-02 3.22015882e-02 6.70997441e-01 -7.18112707e-01
-1.42847812e+00 -3.90264690e-01 -4.30403128e-02 1.05265951e+00
9.14611295e-02 6.19837120e-02 -5.51600635e-01 -4.33948219e-01
-4.43127640e-02 9.07606661e-01 -4.31576759e-01 5.90712726e-01
-5.96545458e-01 -1.04946062e-01 -4.93056536e-01 4.73052084e-01
1.10515378e-01 -9.64561880e-01 -6.66171491e-01 8.14214170e-01
2.83536494e-01 -9.12965119e-01 -3.07163358e-01 4.30234760e-01
-7.94851184e-01 -1.01807511e+00 -4.81896400e-01 -8.76376212e-01
7.82423556e-01 -1.83252096e-01 1.11672509e+00 -1.88260704e-01
3.69589552e-02 -4.10512179e-01 -8.62379968e-01 -9.87917364e-01
-5.58254063e-01 6.01542473e-01 -1.10528193e-01 -1.89732000e-01
4.35518026e-01 -3.36270779e-01 -3.92256439e-01 8.26389730e-01
-3.75032991e-01 1.82832018e-01 9.91644919e-01 2.96630651e-01
5.54298103e-01 9.27558839e-01 7.55283773e-01 -9.81666863e-01
6.20739877e-01 -4.85645682e-01 -9.33448553e-01 2.72434026e-01
-8.96657526e-01 -2.83032179e-01 7.92595327e-01 -2.76450694e-01
-8.37193310e-01 5.17046452e-02 -6.48880377e-02 1.77266970e-01
-4.24372047e-01 6.96273923e-01 -9.69299823e-02 1.85766816e-01
1.78571343e-01 -9.66923684e-02 7.18093663e-02 -5.84413230e-01
-2.54120827e-01 5.41167021e-01 -3.46314162e-01 -2.46647269e-01
6.01052701e-01 -2.24045902e-01 9.45519432e-02 -3.08392406e-01
-4.60195988e-01 -5.29932261e-01 -8.51619303e-01 -5.60902774e-01
6.73798323e-01 -5.56500614e-01 -8.25334132e-01 1.15368679e-01
-8.44129860e-01 5.36459312e-03 -2.19206899e-01 9.40165401e-01
-6.10913932e-01 -5.82864046e-01 -4.10926282e-01 -9.81486917e-01
2.66389456e-02 -1.27269030e+00 3.13258618e-01 3.91338259e-01
-6.23160362e-01 -8.41859758e-01 -1.06339425e-01 6.69435322e-01
3.43839347e-01 3.84678036e-01 9.00832474e-01 -9.05769110e-01
-5.62489450e-01 -2.93712169e-01 7.34727606e-02 2.89757967e-01
2.04755574e-01 1.68490499e-01 8.46867263e-02 -1.96387202e-01
1.82849318e-01 2.23535657e-01 6.42433763e-02 6.74999177e-01
4.66199994e-01 -4.17155772e-01 -2.57909864e-01 -2.94738673e-02
1.75130630e+00 1.10963643e+00 4.41238195e-01 9.09494996e-01
9.56967697e-02 1.13710332e+00 1.70070064e+00 9.20656264e-01
4.20505047e-01 6.12370193e-01 2.53160328e-01 6.12448528e-02
3.69948536e-01 1.25040701e-02 6.94993511e-02 8.59283149e-01
-5.55133462e-01 -7.69779980e-02 -7.71542013e-01 4.84826505e-01
-1.69343352e+00 -7.87124991e-01 3.43786110e-03 1.87646317e+00
4.58062589e-01 3.51840347e-01 4.89450574e-01 3.83328706e-01
5.44175506e-01 -2.92224109e-01 8.93350225e-03 -9.59636927e-01
3.84956688e-01 -8.52029473e-02 1.04586291e+00 9.22052935e-02
-8.67919803e-01 3.73094469e-01 5.74053764e+00 7.88180590e-01
-1.00142968e+00 -2.06230372e-01 6.84132516e-01 -1.06251799e-02
-1.44422084e-01 5.38223028e-01 -1.26817000e+00 6.24479830e-01
6.42904282e-01 -1.09894775e-01 4.70672220e-01 9.69232798e-01
7.80746520e-01 -5.57694316e-01 -1.24137414e+00 6.24091566e-01
-8.62450004e-02 -1.53410256e+00 -4.90785420e-01 3.65702420e-01
8.76762569e-01 -6.88293874e-01 -3.04742932e-01 1.45363241e-01
2.50614196e-01 -1.04136479e+00 1.13088310e+00 5.62346518e-01
1.59847841e-01 -1.20438218e+00 1.40660739e+00 4.87351447e-01
-1.34529018e+00 -5.97099185e-01 -4.74691629e-01 -5.78444839e-01
5.59917510e-01 4.09072012e-01 -1.06337249e+00 9.03553963e-01
4.89491731e-01 6.53484538e-02 2.29543317e-02 1.02318263e+00
1.34687424e-01 2.36913294e-01 -1.71693459e-01 -4.56886977e-01
3.70412916e-01 -8.83361697e-01 -4.15894799e-02 8.97757947e-01
7.10138023e-01 -1.52732044e-01 6.87166035e-01 6.60495758e-01
5.14565349e-01 6.74638331e-01 -4.63222712e-01 -8.48057792e-02
6.77576065e-01 1.05015314e+00 -1.29490030e+00 1.13300458e-01
-2.50753433e-01 1.69976950e-01 -3.05142313e-01 -1.56900242e-01
-7.05784142e-01 -5.63161492e-01 -3.86825278e-02 9.93537486e-01
4.36302990e-01 -2.27652788e-01 -6.07178152e-01 -9.73872989e-02
-2.48466283e-01 -9.90433097e-01 1.62506208e-01 -2.34313220e-01
-1.08284342e+00 3.44688535e-01 4.16856587e-01 -1.35616457e+00
-7.42421299e-02 -6.91846669e-01 -5.43126583e-01 6.27674758e-01
-1.02298021e+00 -1.34628344e+00 7.38640204e-02 -9.43321213e-02
9.32165682e-01 -5.79272330e-01 3.16474795e-01 3.15074950e-01
-5.12512565e-01 1.85236245e-01 2.36488916e-02 -1.46517098e-01
3.57507437e-01 -1.04442525e+00 -3.44494849e-01 4.06600326e-01
-5.02582133e-01 3.79601330e-01 1.12058568e+00 -1.00538099e+00
-1.35694981e+00 -5.28507411e-01 1.38020480e+00 1.62918523e-01
7.68506348e-01 -6.04194514e-02 -1.87908068e-01 5.74322641e-01
5.77718377e-01 -8.63335729e-01 8.10839653e-01 4.31764007e-01
6.28174126e-01 -5.91092467e-01 -1.32291293e+00 -8.05694759e-02
4.44213659e-01 4.33410496e-01 -3.53087187e-01 5.31679928e-01
2.74329036e-01 -1.75859988e-01 -1.67566597e+00 5.57203963e-02
5.59201479e-01 -6.66755140e-01 2.39518732e-01 -2.69014150e-01
3.48670363e-01 -1.30572632e-01 5.44755459e-02 -1.30376744e+00
-6.35199606e-01 -4.94263887e-01 7.18072891e-01 1.40144134e+00
9.60265338e-01 -4.58020121e-01 8.70306373e-01 6.82017326e-01
-4.60894518e-02 -9.88988757e-01 -7.72294462e-01 -7.72447288e-01
1.18917916e-02 4.02795374e-02 7.82261252e-01 7.95478880e-01
-1.46168796e-02 2.43325755e-01 -2.79643476e-01 -2.76312102e-02
4.21811104e-01 5.21067321e-01 8.73812914e-01 -1.12374234e+00
-2.17017204e-01 -3.31707329e-01 -3.92908424e-01 -3.87509942e-01
-4.42550868e-01 -5.09985268e-01 -2.78595954e-01 -1.96223950e+00
-1.04163378e-01 -6.74092710e-01 2.95808744e-02 1.40835240e-01
7.51324296e-01 -3.78186315e-01 6.28112555e-01 8.89911056e-02
3.41757163e-02 -3.73639435e-01 1.63630950e+00 5.12552559e-02
-5.66226542e-01 5.84901154e-01 -8.29070032e-01 7.40904748e-01
8.90422523e-01 -4.18586731e-01 -4.78710204e-01 1.12865604e-01
8.83536518e-01 2.42673740e-01 -5.84088922e-01 -6.18490338e-01
3.17573786e-01 -9.17136371e-01 9.98507664e-02 -9.66959059e-01
-1.67825624e-01 -1.40308464e+00 1.03521669e+00 5.72746158e-01
-8.34390596e-02 6.10680997e-01 -2.22155005e-01 -6.98312372e-02
-2.93411642e-01 -7.61940777e-01 5.67066133e-01 -5.05367756e-01
-5.53332746e-01 -8.61622300e-03 -5.98588943e-01 -6.75412476e-01
1.72573972e+00 -6.39605165e-01 1.39880076e-01 -4.20736074e-02
-1.08431673e+00 3.95055085e-01 2.45613083e-01 9.96168330e-02
8.26064423e-02 -1.10582638e+00 -9.22992647e-01 -2.19881937e-01
-1.14934593e-01 -1.19957045e-01 2.57475942e-01 1.18062603e+00
-1.22366953e+00 4.53819633e-01 -7.82093406e-01 -6.02553859e-02
-1.00811493e+00 7.63721287e-01 -1.90344915e-01 -8.50468814e-01
-2.50687212e-01 7.68295228e-01 -2.91543067e-01 7.19584301e-02
-8.21633339e-02 -7.98690200e-01 -8.13280523e-01 8.41169357e-01
1.13913126e-01 7.42181659e-01 8.88167098e-02 -6.06445253e-01
-1.38860539e-01 6.48293853e-01 -1.76038936e-01 1.21646971e-01
1.55944657e+00 -3.23361486e-01 -2.60096103e-01 4.95531827e-01
5.54934740e-01 5.07098176e-02 -8.25145960e-01 2.09004253e-01
2.58989513e-01 -7.57231832e-01 -1.83778971e-01 -6.18379235e-01
-1.14792609e+00 6.37887940e-02 3.59628856e-01 6.77680314e-01
1.31663597e+00 -1.45592213e-01 4.57814246e-01 -1.41987577e-01
6.86231017e-01 -1.84942627e+00 -4.35697198e-01 2.60929018e-01
1.17395473e+00 -1.00657856e+00 3.01386803e-01 -6.12646699e-01
-9.69445646e-01 1.21236420e+00 3.68928909e-01 -3.86945099e-01
8.45476151e-01 6.65287614e-01 -1.38013616e-01 -2.44053438e-01
-6.69860482e-01 3.60803157e-02 -1.97268605e-01 4.24053401e-01
6.06487215e-01 5.14328957e-01 -1.28530586e+00 9.37080562e-01
-6.51078343e-01 2.41733655e-01 8.48010838e-01 1.09323728e+00
-7.21122563e-01 -1.41733432e+00 -6.25189126e-01 8.36693645e-01
-3.23878527e-01 2.53927767e-01 2.44040340e-02 1.19610620e+00
1.05342817e+00 1.10212135e+00 2.71943927e-01 -4.28616524e-01
7.01350451e-01 -3.49779636e-01 5.95964074e-01 -7.96354592e-01
-1.15973651e+00 4.50745791e-01 6.28605306e-01 -5.88570535e-02
-3.32283556e-01 -1.02410018e+00 -1.07259429e+00 -5.78051448e-01
-7.44465828e-01 3.77605677e-01 1.08270729e+00 8.38773668e-01
-3.06041211e-01 7.32046306e-01 1.08030248e+00 -6.79002941e-01
-6.32741809e-01 -1.17982149e+00 -1.24298525e+00 1.77501842e-01
-6.27789378e-01 -9.01073098e-01 -1.32353492e-02 1.48664787e-01] | [9.111751556396484, 5.92186164855957] |
cbafa84d-a8fa-4f82-aba6-f34f81c0488a | adaptive-action-supervision-in-reinforcement | 2305.13030 | null | https://arxiv.org/abs/2305.13030v2 | https://arxiv.org/pdf/2305.13030v2.pdf | Adaptive action supervision in reinforcement learning from real-world multi-agent demonstrations | Modeling of real-world biological multi-agents is a fundamental problem in various scientific and engineering fields. Reinforcement learning (RL) is a powerful framework to generate flexible and diverse behaviors in cyberspace; however, when modeling real-world biological multi-agents, there is a domain gap between behaviors in the source (i.e., real-world data) and the target (i.e., cyberspace for RL), and the source environment parameters are usually unknown. In this paper, we propose a method for adaptive action supervision in RL from real-world demonstrations in multi-agent scenarios. We adopt an approach that combines RL and supervised learning by selecting actions of demonstrations in RL based on the minimum distance of dynamic time warping for utilizing the information of the unknown source dynamics. This approach can be easily applied to many existing neural network architectures and provide us with an RL model balanced between reproducibility as imitation and generalization ability to obtain rewards in cyberspace. In the experiments, using chase-and-escape and football tasks with the different dynamics between the unknown source and target environments, we show that our approach achieved a balance between the reproducibility and the generalization ability compared with the baselines. In particular, we used the tracking data of professional football players as expert demonstrations in football and show successful performances despite the larger gap between behaviors in the source and target environments than the chase-and-escape task. | ['Yoshinobu Kawahara', 'Naoya Takeishi', 'Hiroshi Nakahara', 'Atom Scott', 'Kazushi Tsutsui', 'Keisuke Fujii'] | 2023-05-22 | null | null | null | null | ['dynamic-time-warping'] | ['time-series'] | [-1.38344780e-01 -2.61844575e-01 7.60667911e-03 2.60608107e-01
-2.13473141e-01 -7.13100314e-01 6.11101747e-01 -3.36169332e-01
-8.42478871e-01 1.14864147e+00 -4.78918672e-01 5.82147799e-02
-3.40281785e-01 -4.68513429e-01 -9.03286636e-01 -1.15338218e+00
-3.88565898e-01 4.59091246e-01 4.84631598e-01 -6.53644145e-01
1.18352018e-01 5.78569055e-01 -1.45920193e+00 -1.29702166e-01
9.30185854e-01 4.10504788e-01 5.89223623e-01 8.22469175e-01
4.35385734e-01 6.68993473e-01 -9.42251980e-01 3.60525638e-01
4.32773262e-01 -6.71623170e-01 -2.69050539e-01 -2.27193326e-01
-2.93487132e-01 -9.71299335e-02 -4.55560297e-01 8.41557384e-01
5.12452900e-01 5.15083551e-01 5.22676766e-01 -1.73125863e+00
-5.09832263e-01 4.40951854e-01 -3.31711650e-01 2.03283370e-01
1.94355458e-01 8.38297606e-01 4.17714387e-01 -6.86902478e-02
7.98334897e-01 1.32214510e+00 4.10472304e-01 8.73227954e-01
-1.14750588e+00 -8.64348650e-01 2.78032064e-01 3.13529491e-01
-1.06534088e+00 -1.11238785e-01 5.42392373e-01 -5.41469812e-01
7.64910758e-01 -1.62252203e-01 1.03236926e+00 1.78295982e+00
7.79481888e-01 6.22288048e-01 1.22798586e+00 1.28567708e-03
6.56028926e-01 -7.56550133e-02 -2.09491745e-01 6.38129532e-01
8.86295810e-02 1.01040220e+00 -3.93684030e-01 -1.69595405e-01
1.06023753e+00 -2.54140683e-02 -3.50671589e-01 -4.79753047e-01
-1.69235861e+00 6.95773542e-01 4.20232892e-01 2.99207479e-01
-4.73165840e-01 2.72110552e-01 2.43162110e-01 6.81462228e-01
-2.33177736e-01 8.67515087e-01 -5.57057798e-01 -3.07001203e-01
-2.09936202e-01 6.39193952e-01 8.31337631e-01 8.65114331e-01
2.84712672e-01 5.29724360e-01 7.56938010e-02 4.66620773e-01
1.21051408e-01 6.75967038e-01 7.78517544e-01 -9.69104290e-01
3.61952603e-01 3.96406174e-01 5.57958961e-01 -7.47060537e-01
-5.95260859e-01 -4.63947326e-01 -6.40022457e-01 8.43809128e-01
8.07824790e-01 -5.92885256e-01 -7.68865466e-01 2.24118495e+00
5.47087491e-01 4.22784328e-01 5.32083988e-01 1.02930343e+00
2.01584488e-01 6.78285062e-01 -5.93380034e-02 -2.94590116e-01
9.46083426e-01 -1.01991916e+00 -6.01011276e-01 -1.73079997e-01
3.52349162e-01 -2.34962311e-02 1.16834342e+00 3.56542408e-01
-7.22929657e-01 -6.54114366e-01 -1.29605436e+00 7.57096052e-01
-4.27127123e-01 -1.45788044e-01 2.86460161e-01 1.34094968e-01
-7.88638771e-01 1.03528440e+00 -1.21918023e+00 -3.98393035e-01
-1.43378034e-01 3.69262815e-01 -3.35542083e-01 5.31468987e-01
-1.31707919e+00 1.09852004e+00 5.52136898e-01 -1.04165219e-01
-1.71788526e+00 -5.18550813e-01 -5.29692769e-01 -2.37366930e-01
6.45946801e-01 -5.31516552e-01 1.19258070e+00 -8.72082829e-01
-1.99637055e+00 8.00559595e-02 6.67607844e-01 -3.33212972e-01
6.54166162e-01 2.04583406e-02 -2.22281039e-01 -1.88669842e-02
-2.91141141e-02 4.79367584e-01 8.45400393e-01 -1.09761167e+00
-5.95983803e-01 -2.18642086e-01 1.58159018e-01 4.32963371e-01
6.69605657e-03 -4.15338486e-01 2.91593492e-01 -5.64131141e-01
-1.60490230e-01 -1.40598071e+00 -2.26743579e-01 7.98432380e-02
-1.33947030e-01 -1.63673133e-01 8.15375507e-01 -3.59791040e-01
4.99263644e-01 -2.10911489e+00 7.69772768e-01 -1.13122240e-01
-4.93411766e-03 2.43555471e-01 -4.21208143e-01 6.48618758e-01
1.68142349e-01 -1.83508292e-01 7.21180290e-02 1.76786721e-01
-6.89241663e-02 3.35098535e-01 -2.66725361e-01 4.42202330e-01
-1.18083164e-01 7.23908186e-01 -1.41593564e+00 -2.93042392e-01
-1.09286226e-01 3.48353118e-01 -1.72574490e-01 4.19299960e-01
-4.05335844e-01 1.12314069e+00 -7.27089822e-01 4.63357180e-01
-1.48902042e-02 -4.28115465e-02 2.17794880e-01 2.28609860e-01
-2.37656295e-01 -3.88173372e-01 -1.15268612e+00 1.49111724e+00
-4.13464874e-01 5.01005948e-01 2.10644662e-01 -8.68052542e-01
9.72752035e-01 3.59836340e-01 6.63077414e-01 -6.31699264e-01
2.21869737e-01 2.60795265e-01 7.84834921e-01 -5.50043404e-01
1.08468749e-01 -1.89055562e-01 -5.10941818e-02 5.40193677e-01
1.84611917e-01 -4.78521615e-01 3.02465916e-01 -2.22407475e-01
1.21266961e+00 6.22080326e-01 2.66020089e-01 -1.96114302e-01
3.64968181e-01 1.89002499e-01 8.14276338e-01 8.62984359e-01
-5.47819257e-01 -6.75194839e-04 3.85081917e-01 -3.70978624e-01
-1.01418030e+00 -1.12561893e+00 3.50040615e-01 9.35927629e-01
4.09270048e-01 -8.35127756e-03 -6.02282822e-01 -4.90313292e-01
6.57986430e-03 5.97757697e-01 -7.01270819e-01 -5.99298060e-01
-8.77242684e-01 -6.13321424e-01 5.90381622e-01 4.33584064e-01
4.11898375e-01 -1.60203230e+00 -1.16726196e+00 4.10956681e-01
2.43391586e-03 -9.89031792e-01 -3.80355090e-01 2.80592263e-01
-6.87910914e-01 -1.31686425e+00 -6.84933901e-01 -6.17073715e-01
3.75917077e-01 4.72445525e-02 4.84306753e-01 1.58328954e-02
-2.39843771e-01 4.29936022e-01 -2.34377742e-01 -4.81057942e-01
-7.60826051e-01 -2.48487934e-01 6.11429691e-01 -2.77653545e-01
-3.08495343e-01 -7.12172031e-01 -3.89679372e-01 7.24839866e-01
-9.01423573e-01 7.46887997e-02 5.17225862e-01 1.20473135e+00
3.99653465e-01 8.87841880e-02 7.32519209e-01 3.32351727e-03
8.06827724e-01 -4.65490252e-01 -9.93759036e-01 2.85423219e-01
-2.75990605e-01 2.35843927e-01 1.02137923e+00 -1.37676299e+00
-6.71584845e-01 7.06174970e-03 1.82487473e-01 -3.50555837e-01
-1.12869397e-01 2.24560961e-01 1.22282319e-02 -1.67225927e-01
7.40501225e-01 5.36061883e-01 4.13869083e-01 -1.20518230e-01
1.62612841e-01 3.54317188e-01 1.47680163e-01 -7.25045860e-01
8.61852288e-01 2.03213349e-01 2.22407237e-01 -7.36251831e-01
-1.54302895e-01 1.85690433e-01 -5.57783127e-01 -5.38290739e-01
6.95952952e-01 -6.27173901e-01 -1.28104496e+00 9.12523985e-01
-8.94227147e-01 -8.70874047e-01 -3.06577325e-01 8.31835151e-01
-1.12584627e+00 9.65186283e-02 -6.69507444e-01 -8.50873411e-01
2.62001157e-02 -1.44638014e+00 6.66374445e-01 5.63576877e-01
6.57647252e-02 -8.23380709e-01 4.42991704e-01 -1.40329644e-01
4.67014879e-01 5.40314555e-01 8.83360147e-01 -7.95617282e-01
-5.94330430e-01 1.60172746e-01 4.83628750e-01 3.29469927e-02
1.02597468e-01 -1.27184782e-02 -3.73825073e-01 -7.42830396e-01
1.70119852e-01 -6.56176746e-01 1.92693025e-01 1.73956856e-01
5.46964526e-01 -1.55818656e-01 -4.66952562e-01 1.94398552e-01
1.14932477e+00 7.94515967e-01 2.28759870e-01 5.70087731e-01
2.69540280e-01 5.23157775e-01 9.03505445e-01 3.92601192e-01
2.61968166e-01 8.60848546e-01 5.33892930e-01 3.18102121e-01
1.21691078e-01 -4.19255525e-01 7.93869436e-01 6.03304863e-01
-1.31096214e-01 -2.00157613e-01 -7.28307962e-01 3.44878227e-01
-2.19612527e+00 -1.16910279e+00 4.15723890e-01 2.16678667e+00
8.31718206e-01 6.02340028e-02 5.19769192e-01 -1.40451103e-01
7.65321553e-01 -1.75632223e-01 -1.28724444e+00 6.59631044e-02
-9.33917612e-02 -3.04665029e-01 2.96833009e-01 3.88366431e-01
-7.40366876e-01 8.30567539e-01 6.03408241e+00 5.84748566e-01
-1.20777941e+00 -1.77305162e-01 1.48041725e-01 -3.19908410e-01
4.45250839e-01 -2.79530883e-01 -7.63053596e-01 6.51210189e-01
1.00818884e+00 -3.06811094e-01 8.26047182e-01 7.67431855e-01
3.79120022e-01 4.51835915e-02 -1.34487855e+00 7.12428331e-01
-2.70230263e-01 -9.35974538e-01 -5.28856456e-01 2.19388492e-02
5.68355560e-01 8.76592174e-02 -4.22811508e-02 7.09737420e-01
8.10337126e-01 -7.50511408e-01 7.78212428e-01 5.69172919e-01
2.43734524e-01 -3.78570676e-01 4.41840708e-01 1.08721340e+00
-9.21217740e-01 -4.61937845e-01 -1.87452018e-01 -9.75425243e-02
1.38205543e-01 -2.74709374e-01 -6.59931600e-01 3.84846330e-01
4.77198750e-01 5.94883621e-01 -1.97324425e-01 1.06083691e+00
-1.76242664e-01 2.47316346e-01 -2.89372087e-01 -7.43120968e-01
2.73176908e-01 -6.27582133e-01 9.35266316e-01 4.57612723e-01
3.91634822e-01 -6.26962632e-02 4.15076137e-01 9.16988313e-01
5.36867917e-01 -2.46620864e-01 -8.15489888e-01 -1.56623006e-01
3.81974488e-01 1.06938434e+00 -7.01397717e-01 -2.37877369e-01
1.37390688e-01 5.26037455e-01 4.53369915e-01 5.61022997e-01
-1.34198105e+00 -4.41549271e-01 7.42750943e-01 -4.05898243e-01
1.84652895e-01 -5.81942081e-01 4.82380360e-01 -1.15175438e+00
-1.18728377e-01 -1.21725595e+00 1.08944528e-01 -5.89857697e-01
-1.18159556e+00 7.78233469e-01 2.47566015e-01 -1.54702401e+00
-6.32878363e-01 -5.84396064e-01 -4.91959929e-01 5.09783566e-01
-1.16844547e+00 -7.34681249e-01 -2.56550431e-01 6.95707560e-01
5.38091719e-01 -4.79325593e-01 5.33173084e-01 -1.82315081e-01
-4.82365668e-01 2.33865544e-01 4.55800712e-01 -1.00205638e-01
4.96684849e-01 -1.06405663e+00 1.00285642e-01 4.23257977e-01
9.95393917e-02 4.48624432e-01 9.77425635e-01 -6.45069063e-01
-1.63393462e+00 -7.19986558e-01 -1.65622830e-01 -2.11142734e-01
9.59743738e-01 -4.02603209e-01 -7.99304426e-01 5.44662476e-01
7.78741017e-02 -2.28434592e-01 1.87443674e-01 -4.19520468e-01
1.88609641e-02 1.84147600e-02 -1.18909669e+00 1.17680371e+00
1.12350798e+00 -3.16032767e-02 -5.99270821e-01 4.14537668e-01
8.11798513e-01 -5.40412307e-01 -6.80141568e-01 3.30384761e-01
7.80961514e-01 -6.33414567e-01 9.55677509e-01 -1.09998798e+00
1.33748233e-01 -5.83191574e-01 6.15403391e-02 -2.01338148e+00
-2.20349312e-01 -6.82690203e-01 -4.32614610e-02 6.51398957e-01
2.00542465e-01 -7.95373023e-01 5.36977947e-01 1.01603627e-01
-1.39796492e-02 -6.23510182e-01 -1.11670327e+00 -1.32959485e+00
3.69166940e-01 2.32341260e-01 4.85111207e-01 5.82252800e-01
2.59415269e-01 1.71483591e-01 -5.58964193e-01 1.45316184e-01
6.51260138e-01 2.09965989e-01 9.52138782e-01 -1.05504131e+00
-6.57852232e-01 -3.49508345e-01 -3.29603851e-01 -9.42266941e-01
5.57432711e-01 -5.82256377e-01 4.49390799e-01 -1.12553036e+00
-1.30636051e-01 -3.96630615e-01 -3.70924443e-01 3.27050537e-01
2.49529649e-02 -4.53014076e-01 4.87673730e-01 3.19753230e-01
-4.28830653e-01 8.91632736e-01 1.66055071e+00 -1.59168139e-01
-5.68194807e-01 1.58715814e-01 -1.25999793e-01 5.80581546e-01
1.04609621e+00 -5.95524669e-01 -6.85126126e-01 -2.65292406e-01
-1.98297277e-01 5.87811232e-01 4.05192524e-01 -1.32769692e+00
2.84916312e-01 -6.82714939e-01 9.59415734e-02 -1.54720783e-01
7.50398934e-01 -8.24530780e-01 3.27037513e-01 1.12962079e+00
-4.40209061e-01 3.19823533e-01 2.42261156e-01 1.02195942e+00
1.23836257e-01 -2.25955453e-02 8.47858489e-01 -2.27831393e-01
-6.02057874e-01 1.83557659e-01 -6.13585591e-01 -1.25416303e-02
1.48387027e+00 -9.55948830e-02 -6.63657665e-01 -3.36247891e-01
-7.97765195e-01 5.25502205e-01 4.28280115e-01 6.47161782e-01
6.34497881e-01 -1.16575491e+00 -5.31218827e-01 2.23882720e-01
-9.75824818e-02 -5.15349627e-01 1.81296140e-01 8.25030327e-01
-2.46550217e-01 1.66698560e-01 -9.47463155e-01 -5.34430206e-01
-1.10261989e+00 6.40856266e-01 6.20824158e-01 -3.08646530e-01
-5.93927443e-01 2.38802046e-01 1.06681965e-01 -6.06529415e-01
2.92341739e-01 -3.68413657e-01 -2.38329977e-01 -5.06541312e-01
3.70005608e-01 3.90583634e-01 -5.08844316e-01 -3.60541403e-01
-1.47960290e-01 4.37180161e-01 1.66923031e-01 -4.05745059e-01
1.37769020e+00 1.95140600e-01 3.60461444e-01 7.68009245e-01
5.25910258e-01 -4.67730403e-01 -1.74299049e+00 1.34947464e-01
8.90363473e-03 -2.82349110e-01 -4.73746777e-01 -8.84166300e-01
-8.35631788e-01 6.01797938e-01 6.34914875e-01 1.61151469e-01
6.92877114e-01 -2.27134228e-01 3.70245367e-01 7.27485180e-01
9.29937959e-01 -1.21877599e+00 6.85062408e-01 5.44140875e-01
1.20629358e+00 -1.09233487e+00 -3.14661950e-01 1.75523356e-01
-8.71978760e-01 1.03247833e+00 1.09076881e+00 -4.50433880e-01
4.71410930e-01 3.73436421e-01 1.12463474e-01 7.94832259e-02
-1.10189021e+00 1.04552269e-01 -1.76359966e-01 1.03922248e+00
-2.37475827e-01 -4.34677042e-02 -1.55288085e-01 3.94953430e-01
-3.84706451e-04 -1.22827083e-01 8.17801058e-01 1.06541574e+00
-4.62276280e-01 -1.05894947e+00 -4.41846400e-01 -1.00959644e-01
1.46685496e-01 4.85970914e-01 -2.47810438e-01 1.32073534e+00
-8.25675279e-02 9.45486844e-01 -3.20363224e-01 -5.21675766e-01
4.27007675e-01 -1.09968737e-01 5.54659307e-01 -3.61204594e-01
-6.91143692e-01 -1.26333773e-01 -2.50709265e-01 -6.28447831e-01
-2.86408663e-01 -6.22649014e-01 -1.41658771e+00 -8.59998688e-02
-3.10708612e-01 2.02478796e-01 6.76505744e-01 7.90080428e-01
3.79633397e-01 6.10030413e-01 7.08603024e-01 -1.00167572e+00
-1.21875226e+00 -1.03829134e+00 -7.16796696e-01 4.12774354e-01
5.41170299e-01 -1.20454180e+00 -6.07620716e-01 -1.88104168e-01] | [4.2978835105896, 1.598680853843689] |
2ba23d17-20d0-4011-802a-7ad1120daf38 | portmanteauing-features-for-scene-text | 2211.05036 | null | https://arxiv.org/abs/2211.05036v1 | https://arxiv.org/pdf/2211.05036v1.pdf | Portmanteauing Features for Scene Text Recognition | Scene text images have different shapes and are subjected to various distortions, e.g. perspective distortions. To handle these challenges, the state-of-the-art methods rely on a rectification network, which is connected to the text recognition network. They form a linear pipeline which uses text rectification on all input images, even for images that can be recognized without it. Undoubtedly, the rectification network improves the overall text recognition performance. However, in some cases, the rectification network generates unnecessary distortions on images, resulting in incorrect predictions in images that would have otherwise been correct without it. In order to alleviate the unnecessary distortions, the portmanteauing of features is proposed. The portmanteau feature, inspired by the portmanteau word, is a feature containing information from both the original text image and the rectified image. To generate the portmanteau feature, a non-linear input pipeline with a block matrix initialization is presented. In this work, the transformer is chosen as the recognition network due to its utilization of attention and inherent parallelism, which can effectively handle the portmanteau feature. The proposed method is examined on 6 benchmarks and compared with 13 state-of-the-art methods. The experimental results show that the proposed method outperforms the state-of-the-art methods on various of the benchmarks. | ['Joo Hwee Lim', 'Jung-jae Kim', 'Adams Wai-Kin Kong', 'Ernest Yu Kai Chew', 'Yew Lee Tan'] | 2022-11-09 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 6.31160438e-01 -4.02528763e-01 1.54470280e-01 -2.92116404e-01
-1.17913492e-01 -1.81761086e-01 7.80225217e-01 -2.47029915e-01
-3.70867908e-01 2.51900285e-01 1.36603564e-01 -2.69584246e-02
1.36028096e-01 -7.15768516e-01 -6.73042238e-01 -7.78416693e-01
8.82630169e-01 3.05367678e-01 2.83741623e-01 -2.99495161e-01
7.34693527e-01 3.40231478e-01 -1.59042180e+00 7.79211164e-01
7.42407799e-01 1.07412469e+00 3.29837322e-01 4.54504311e-01
-4.92766231e-01 6.48286462e-01 -5.77919781e-01 -4.62891310e-01
3.77099693e-01 -4.17641550e-01 -4.32219476e-01 2.32950181e-01
5.83330810e-01 -4.77300614e-01 -6.29845381e-01 1.18038857e+00
1.99690968e-01 1.75699979e-01 7.27046788e-01 -9.64872360e-01
-8.83944750e-01 5.45356989e-01 -8.41650426e-01 1.78853169e-01
3.14791888e-01 3.54886726e-02 8.31110060e-01 -1.40536368e+00
4.62890297e-01 1.31331122e+00 4.15848672e-01 1.36353016e-01
-8.83210063e-01 -7.38370359e-01 1.73095256e-01 5.55829167e-01
-1.53772867e+00 -4.21731561e-01 6.89007878e-01 -3.05940211e-01
1.13211381e+00 3.61537695e-01 4.10109550e-01 8.08718145e-01
6.04623139e-01 9.12050247e-01 8.09619367e-01 -4.71757650e-01
-2.43019551e-01 1.45244077e-01 6.11326694e-02 4.56239909e-01
1.49788111e-01 -1.62834704e-01 -4.80406493e-01 4.56423372e-01
6.55776560e-01 5.74892342e-01 -4.09923047e-01 -5.17795235e-02
-1.18998981e+00 4.68424559e-01 5.08741140e-01 4.07856464e-01
-2.67911822e-01 -1.79895326e-01 3.49870712e-01 2.45914817e-01
4.65891063e-02 -9.21641961e-02 -1.19353607e-01 -9.06723831e-03
-9.99788582e-01 -2.30885103e-01 4.82105017e-01 8.78965676e-01
6.68140888e-01 2.27283821e-01 -1.06379315e-01 7.67139971e-01
3.74282509e-01 6.58412397e-01 8.50501716e-01 1.97365254e-01
9.46997404e-01 1.08840024e+00 -2.01837495e-01 -1.33465219e+00
-3.91310930e-01 -3.67542088e-01 -1.41332448e+00 7.86330178e-02
6.75232187e-02 3.03147376e-01 -1.18840587e+00 9.36919510e-01
-7.18274191e-02 -3.41839902e-02 2.29631722e-01 1.11961818e+00
9.16602194e-01 1.05794680e+00 -5.31043291e-01 1.58994924e-02
1.17746305e+00 -1.30546451e+00 -7.87421107e-01 -3.69577885e-01
3.95123839e-01 -1.30523980e+00 1.02179623e+00 5.33979654e-01
-8.87326837e-01 -8.38738561e-01 -1.24051642e+00 -1.41040266e-01
-4.84105140e-01 5.93164921e-01 -1.00763425e-01 5.21657765e-01
-7.84741461e-01 3.22264254e-01 -6.20335162e-01 -4.90693152e-01
1.32984906e-01 3.71254951e-01 -4.57267940e-01 -2.55879641e-01
-9.43492353e-01 1.04335046e+00 4.86074746e-01 5.19670486e-01
-4.30807918e-01 -1.21556163e-01 -7.10333824e-01 4.53714520e-01
2.63918996e-01 -2.05237508e-01 7.93983221e-01 -1.40185332e+00
-1.52345550e+00 4.55885112e-01 -1.24504782e-01 -1.59704968e-01
7.28756189e-01 -5.85785061e-02 -5.42259932e-01 -3.52814682e-02
-3.33981186e-01 4.85079020e-01 1.31080878e+00 -8.43518913e-01
-7.72229612e-01 -3.27461004e-01 -3.33122283e-01 5.02287269e-01
-4.47827935e-01 -4.04604711e-02 -4.76486772e-01 -9.79225457e-01
3.73189330e-01 -7.23633587e-01 2.96773195e-01 -2.45228205e-02
-4.63739812e-01 4.17252295e-02 1.42727554e+00 -5.40485442e-01
1.07653296e+00 -2.35537863e+00 1.42794028e-01 1.23711482e-01
8.96757916e-02 5.27004838e-01 -1.52259499e-01 5.44756711e-01
-2.01695651e-01 2.44100969e-02 -6.65346831e-02 -2.48581290e-01
-1.40609711e-01 -2.68405471e-02 -6.37408495e-01 5.90520084e-01
2.32803240e-01 6.41841829e-01 -3.15300107e-01 -1.38864830e-01
5.70648611e-01 6.24662399e-01 -3.22980464e-01 2.62329221e-01
4.32955287e-02 2.13234454e-01 -2.76708663e-01 2.89398819e-01
1.10542214e+00 -1.00582927e-01 -5.28991036e-02 -4.65100557e-01
-2.82950938e-01 1.81087047e-01 -1.39893222e+00 1.23684359e+00
-2.31672317e-01 6.83580577e-01 -2.90706068e-01 -1.11216831e+00
1.28035998e+00 3.59383114e-02 5.28767221e-02 -7.48105586e-01
5.23625195e-01 2.71576554e-01 1.15906753e-01 -3.96948665e-01
7.35252321e-01 1.60129517e-01 2.56630838e-01 5.74288011e-01
-3.63265038e-01 -1.28393350e-02 1.42143503e-01 -4.44288515e-02
8.09676290e-01 9.25962105e-02 1.69490010e-01 -5.87240271e-02
1.24748588e+00 -3.36247049e-02 3.32139283e-01 4.63622153e-01
1.75869331e-01 7.84524262e-01 2.69417256e-01 -7.62480676e-01
-1.13212168e+00 -6.71747923e-01 -1.19979583e-01 7.84914851e-01
4.56804454e-01 -4.00088727e-01 -6.13404751e-01 -4.57135051e-01
-2.53178328e-01 3.59862030e-01 -4.89721745e-01 -2.23294526e-01
-5.59043586e-01 -6.55883670e-01 5.95268965e-01 5.40554285e-01
1.16139388e+00 -8.38516533e-01 -6.05494916e-01 -1.40309159e-03
-2.85832733e-02 -1.13364494e+00 -7.39187777e-01 -9.03367475e-02
-6.98195934e-01 -8.38708103e-01 -7.02258170e-01 -9.68684614e-01
1.03294206e+00 6.70443892e-01 3.76211405e-01 2.16617808e-01
-1.23780482e-01 -4.11396511e-02 -5.51992834e-01 -1.77656468e-02
-3.38780880e-01 3.64982970e-02 -6.38164133e-02 4.03785259e-01
3.53049219e-01 -2.96941698e-01 -5.49839437e-01 6.89396441e-01
-1.23955846e+00 5.84191203e-01 8.75495493e-01 1.30950809e+00
5.96573114e-01 4.78381038e-01 1.78001136e-01 -7.24637985e-01
3.55183721e-01 -5.41628599e-02 -6.85697913e-01 2.61133283e-01
-6.09297276e-01 5.51509038e-02 1.34152424e+00 -6.11312091e-01
-1.12895799e+00 1.87446982e-01 5.14944196e-02 -4.64058638e-01
5.85956573e-02 3.54511231e-01 -2.64577329e-01 -1.24331437e-01
3.78407955e-01 7.39070714e-01 -1.68071508e-01 -3.85681778e-01
1.61426216e-01 1.05884635e+00 3.89484346e-01 2.36786492e-02
9.17709470e-01 4.09205616e-01 -3.85655276e-02 -9.07855153e-01
-4.23900127e-01 -1.59279317e-01 -6.74552500e-01 1.33065938e-03
5.66577911e-01 -7.46177554e-01 -5.37254632e-01 9.90256667e-01
-1.25226080e+00 1.62353903e-01 2.55224526e-01 4.19299394e-01
-2.33389229e-01 5.42234719e-01 -5.14062703e-01 -5.50877273e-01
-5.90555429e-01 -1.53097749e+00 9.83046174e-01 5.70580304e-01
1.87194124e-01 -6.50445223e-01 -4.74565715e-01 2.36208841e-01
6.07200265e-01 -4.40113038e-01 9.77890372e-01 -7.77247131e-01
-7.67465949e-01 -2.46156424e-01 -6.72024667e-01 3.04539382e-01
2.77512491e-01 1.20782340e-02 -7.60367870e-01 -4.08367187e-01
1.75271198e-01 6.99431822e-02 8.09211075e-01 -1.23396851e-01
9.28588688e-01 -4.68479246e-01 -8.22204873e-02 6.80522919e-01
1.36000586e+00 3.65909338e-01 9.95417714e-01 4.28810239e-01
8.54726315e-01 2.99089462e-01 5.65853953e-01 4.48342264e-01
3.50078940e-01 5.70753992e-01 3.71108502e-01 -1.88188121e-01
-1.20716125e-01 -2.29911819e-01 5.66011846e-01 9.89040375e-01
3.07577997e-01 -5.45689762e-01 -8.18147957e-01 1.10592581e-01
-1.82233012e+00 -7.26568758e-01 -3.86632115e-01 2.02403545e+00
5.14671922e-01 2.15587065e-01 -4.90491837e-01 4.77356076e-01
1.13151276e+00 4.00147140e-01 -5.63189507e-01 -5.63558877e-01
-3.40380311e-01 4.03759591e-02 4.59241211e-01 2.08444685e-01
-8.62729311e-01 1.12318540e+00 5.54253197e+00 8.15511584e-01
-1.65318716e+00 -2.68250197e-01 4.12846297e-01 2.40041494e-01
5.11710905e-02 -5.45868613e-02 -8.48497212e-01 4.30458397e-01
5.07533193e-01 -1.72962248e-01 6.63736582e-01 5.87427616e-01
1.64236084e-01 -9.49717388e-02 -1.04696429e+00 1.22564340e+00
6.58782780e-01 -1.07542217e+00 4.03935015e-01 -3.61778319e-01
7.52584815e-01 -1.51295379e-01 2.76395500e-01 1.41156077e-01
-3.63768518e-01 -1.00804007e+00 5.84085822e-01 4.73964453e-01
9.90117192e-01 -6.88979566e-01 9.36456919e-01 5.07599533e-01
-1.31261301e+00 -1.55321346e-03 -7.23991454e-01 -7.83184394e-02
-1.51845947e-01 5.53292811e-01 -8.39596093e-01 4.93299335e-01
5.00273526e-01 1.05974054e+00 -5.98853588e-01 8.21328163e-01
-2.78031558e-01 2.00354546e-01 -2.31569707e-01 -5.04990295e-02
2.94301659e-01 -3.56422037e-01 3.91470790e-01 1.19947243e+00
4.56530213e-01 1.09372839e-01 6.95575774e-03 7.54770517e-01
-2.08873391e-01 4.77831930e-01 -5.07405460e-01 -6.03164621e-02
1.26497909e-01 1.17522049e+00 -6.82791054e-01 -3.95379335e-01
-5.49957335e-01 1.28321970e+00 -1.11419886e-01 9.06197652e-02
-7.26801634e-01 -8.64015222e-01 2.06093863e-01 1.04747834e-02
4.97987568e-01 -3.55195701e-02 -3.68821472e-01 -1.32790160e+00
2.89037466e-01 -1.08618617e+00 1.76317871e-01 -1.09163916e+00
-1.01037085e+00 7.98469245e-01 -4.12512600e-01 -1.44834423e+00
9.16236788e-02 -7.73925126e-01 -6.34741485e-01 9.82277513e-01
-1.59498477e+00 -1.15028548e+00 -6.61507845e-01 6.37026966e-01
9.97705638e-01 -3.26042980e-01 4.89725202e-01 3.56877089e-01
-9.48280513e-01 7.67032146e-01 3.52896631e-01 1.48527890e-01
9.55340087e-01 -6.62513316e-01 5.28998375e-01 1.08281004e+00
-6.30056113e-02 3.72295707e-01 3.81874353e-01 -7.27099538e-01
-1.90935326e+00 -1.18228412e+00 6.40658319e-01 4.32423577e-02
6.03408754e-01 -2.12175474e-01 -1.08258772e+00 7.21735716e-01
1.82911232e-01 -9.17791128e-02 1.15666509e-01 -6.92731559e-01
-5.06434500e-01 -3.29676032e-01 -1.10390389e+00 6.82890296e-01
7.28396118e-01 -3.47088099e-01 -7.67369151e-01 5.78767806e-03
2.29736850e-01 -7.63510168e-01 -5.38511634e-01 5.37107363e-02
6.70064867e-01 -1.00421309e+00 4.80094075e-01 -9.88939852e-02
6.80967033e-01 -5.79076767e-01 -1.68852627e-01 -1.10839128e+00
-1.89856887e-01 -3.17555338e-01 3.58171731e-01 1.26277578e+00
2.54391015e-01 -9.70104814e-01 5.08044302e-01 3.05523545e-01
1.32708484e-03 -4.27187800e-01 -9.01308894e-01 -5.62907636e-01
-9.47916582e-02 -8.92322287e-02 6.88484490e-01 8.29786658e-01
-1.41312825e-02 4.89044905e-01 -5.49576044e-01 6.77591236e-03
1.84167936e-01 7.60044754e-02 7.08999217e-01 -9.35589433e-01
5.90286814e-02 -4.77243513e-01 -7.38457561e-01 -1.43000150e+00
9.82147530e-02 -8.56258988e-01 1.37535408e-01 -1.35017121e+00
2.53582746e-01 -1.59411386e-01 -2.70016994e-02 4.77421999e-01
-9.36316550e-02 2.79761255e-01 4.56634730e-01 3.78220975e-01
-2.78220534e-01 6.78343296e-01 1.35469139e+00 -6.21004879e-01
-5.31396233e-02 -2.45616779e-01 -3.76549214e-01 6.81344271e-01
7.17916906e-01 -4.49701071e-01 -3.65248919e-01 -7.76956141e-01
1.80923149e-01 -1.15813263e-01 -5.11983875e-04 -1.04284489e+00
5.96872389e-01 -7.40432888e-02 6.74788177e-01 -1.12085474e+00
3.44941854e-01 -1.14430892e+00 2.41506919e-01 4.06243473e-01
-1.82372689e-01 3.96384865e-01 1.09368831e-01 2.71804363e-01
-3.47148716e-01 -3.32922906e-01 9.95059967e-01 1.87878519e-01
-4.36008692e-01 1.28579577e-02 -5.41541874e-01 -1.99459553e-01
8.14885497e-01 -5.32392740e-01 -7.38091707e-01 -2.06151158e-01
-7.33149946e-02 1.72670186e-02 5.96624017e-01 6.00290298e-01
1.02842939e+00 -1.20201230e+00 -7.06358373e-01 7.22927034e-01
-2.18576621e-02 6.51668608e-02 3.19088906e-01 9.23189819e-01
-6.08645797e-01 5.85380793e-01 -4.48842615e-01 -7.38071680e-01
-1.19887137e+00 5.27132452e-01 3.54323834e-01 -2.84081340e-01
-8.55800629e-01 1.64035261e-01 3.22886050e-01 -1.06937893e-01
1.41521990e-01 -3.73371035e-01 -3.18161219e-01 -1.81149915e-01
7.15585291e-01 3.53727341e-01 2.36171663e-01 -9.74680901e-01
-2.54733831e-01 1.05730176e+00 -5.05421460e-01 1.32608414e-01
1.10548520e+00 -1.73822641e-01 -2.33051330e-01 1.65696308e-01
1.01625884e+00 -1.14011578e-01 -9.04267311e-01 -4.04084980e-01
-4.13450927e-01 -8.55866075e-01 1.42393857e-01 -7.28852332e-01
-1.31259775e+00 1.09530079e+00 6.08256876e-01 -1.87518739e-03
1.44107771e+00 -9.25300598e-01 8.72253597e-01 6.70183361e-01
1.92764983e-01 -9.85980868e-01 6.46994710e-02 8.65881443e-01
1.05801344e+00 -1.11557829e+00 2.09664881e-01 -4.23706293e-01
-7.60311186e-01 1.57413411e+00 8.14106345e-01 -1.27497703e-01
3.46367389e-01 2.37235278e-01 5.51086962e-02 2.16267765e-01
-6.75308406e-01 1.25209525e-01 1.56005368e-01 2.55993962e-01
1.60751984e-01 -2.99655408e-01 -4.28178906e-01 6.56856954e-01
-1.14162356e-01 -9.81788263e-02 8.60250890e-01 8.05127561e-01
-3.18315268e-01 -1.11922503e+00 -8.42305958e-01 5.09352326e-01
-2.72655636e-01 -1.87782228e-01 -4.38495040e-01 6.07554853e-01
-1.29604423e-02 9.21475053e-01 2.60839224e-01 -6.12939894e-01
6.03598058e-01 -1.47225156e-01 1.96322650e-01 -2.97198087e-01
-7.61980534e-01 1.49396107e-01 -3.62804830e-01 -1.39277831e-01
-1.25393823e-01 -3.09991539e-01 -1.45581222e+00 -4.57931817e-01
-5.62971234e-01 -1.85827985e-01 7.85096526e-01 9.09095705e-01
3.14518452e-01 5.07605314e-01 9.35653567e-01 -8.50602567e-01
-4.69179183e-01 -9.95381355e-01 -3.81222576e-01 3.33808720e-01
1.75914824e-01 -5.09807885e-01 -4.06497866e-01 1.97628334e-01] | [11.87619400024414, 2.162376642227173] |
c1c04ff7-46b5-4d66-a334-60843c1c8a6c | pre-trained-models-or-feature-engineering-the | null | null | https://aclanthology.org/2022.osact-1.5 | https://aclanthology.org/2022.osact-1.5.pdf | Pre-trained Models or Feature Engineering: The Case of Dialectal Arabic | The usage of social media platforms has resulted in the proliferation of work on Arabic Natural Language Processing (ANLP), including the development of resources. There is also an increased interest in processing Arabic dialects and a number of models and algorithms have been utilised for the purpose of Dialectal Arabic Natural Language Processing (DANLP). In this paper, we conduct a comparison study between some of the most well-known and most commonly used methods in NLP in order to test their performance on different corpora and two NLP tasks: Dialect Identification and Sentiment Analysis. In particular, we compare three general classes of models: a) traditional Machine Learning models with features, b) classic Deep Learning architectures (LSTMs) with pre-trained word embeddings and lastly c) different Bidirectional Encoder Representations from Transformers (BERT) models such as (Multilingual-BERT, Ara-BERT, and Twitter-Arabic-BERT). The results of the comparison show that using feature-based classification can still compete with BERT models in these dialectal Arabic contexts. The use of transformer models have the ability to outperform traditional Machine Learning approaches, depending on the type of text they have been trained on, in contrast to classic Deep Learning models like LSTMs which do not perform well on the tasks | ['Simon Dobnik', 'Stergios Chatzikyriakidis', 'Kathrein Abu Kwaik'] | null | null | null | null | osact-lrec-2022-6 | ['dialect-identification'] | ['natural-language-processing'] | [-4.11867917e-01 -1.19419225e-01 2.10652515e-01 -4.66377467e-01
-3.06725532e-01 -6.89335525e-01 1.07968676e+00 7.28403449e-01
-8.28166842e-01 3.95896584e-01 4.36047375e-01 -4.36278135e-01
7.00727701e-02 -1.11899686e+00 -3.48383993e-01 -6.07600212e-01
-2.42645890e-01 7.94313312e-01 -2.13662572e-02 -1.17387569e+00
2.21210152e-01 5.20335138e-01 -1.33251381e+00 5.95071197e-01
7.19288409e-01 8.96330237e-01 -1.17149517e-01 4.79075968e-01
-5.56127787e-01 8.36315870e-01 -4.63638812e-01 -7.06999719e-01
-6.04377761e-02 -1.37875810e-01 -1.17236245e+00 -2.79744297e-01
-5.22074066e-02 -1.93119466e-01 -3.47382436e-03 6.49286747e-01
5.29061794e-01 7.72999600e-02 6.51900887e-01 -8.52469027e-01
-8.01594675e-01 8.96968842e-01 -4.52234596e-01 3.92277464e-02
5.80448151e-01 -3.91099811e-01 9.95502770e-01 -1.17150140e+00
4.97458875e-01 1.63387215e+00 9.15590525e-01 1.61081344e-01
-8.44534218e-01 -2.92440891e-01 2.61158720e-02 5.12261271e-01
-1.00092769e+00 -2.30431557e-01 3.21968347e-01 -3.11954170e-01
1.29795277e+00 -1.73746511e-01 3.68964136e-01 9.57100630e-01
2.41031393e-01 9.91899312e-01 1.25496924e+00 -1.12045944e+00
-1.16643578e-01 3.42068583e-01 3.63902450e-01 8.11569035e-01
-1.43593147e-01 -3.62729847e-01 -4.84830767e-01 1.03505217e-01
1.33255854e-01 -2.83963054e-01 1.16464905e-02 1.53649762e-01
-1.17662191e+00 1.27048743e+00 5.92567027e-01 9.51850712e-01
-3.25643539e-01 -3.27673376e-01 9.20629323e-01 7.59780526e-01
7.68607080e-01 4.06886309e-01 -6.32880330e-01 -8.68071616e-02
-5.42990208e-01 1.71992421e-01 9.25109863e-01 3.88569951e-01
6.55684650e-01 2.97147036e-01 1.90890521e-01 1.21650374e+00
5.64328611e-01 2.04406098e-01 1.10240340e+00 2.63739191e-02
4.51597780e-01 8.16354930e-01 -2.00592205e-01 -1.03342390e+00
-6.41044319e-01 2.39235103e-01 -6.87358201e-01 1.40021503e-01
8.10559869e-01 -2.61849403e-01 -7.74479091e-01 1.30734992e+00
-4.57793176e-02 -6.56358004e-01 3.14611554e-01 5.80037057e-01
7.13084757e-01 1.12316978e+00 -1.26285881e-01 3.53331685e-01
1.30761969e+00 -7.57827044e-01 -6.68988347e-01 -2.59916127e-01
8.02312791e-01 -1.07212305e+00 1.27508605e+00 5.47647059e-01
-8.00706089e-01 -3.27298313e-01 -1.11208296e+00 -2.66905785e-01
-1.46808457e+00 1.40520200e-01 4.26694721e-01 1.00530291e+00
-1.30469179e+00 2.91284829e-01 -5.63837945e-01 -7.84042001e-01
6.10978901e-02 5.71640670e-01 -4.75970328e-01 -2.61536181e-01
-1.58086169e+00 1.53499174e+00 4.75546628e-01 4.84831274e-01
-4.83368754e-01 -8.54276270e-02 -1.07572103e+00 -1.86659858e-01
-2.40629882e-01 2.53688484e-01 1.00823331e+00 -1.58320105e+00
-1.71385014e+00 1.16803098e+00 8.61383323e-03 -7.21119642e-01
1.54940158e-01 -4.81354713e-01 -5.45854926e-01 -2.34368682e-01
-7.84623548e-02 5.65622807e-01 5.77402055e-01 -1.03437817e+00
-6.80743217e-01 -4.63584423e-01 3.49363059e-01 2.28010610e-01
-6.81504309e-01 5.02043724e-01 1.62424520e-01 -6.23102725e-01
-1.41363561e-01 -7.96217263e-01 3.78852934e-02 -5.70446610e-01
-1.84473917e-01 -5.40933788e-01 7.01938033e-01 -9.86264884e-01
1.16704702e+00 -1.91824007e+00 1.57890737e-01 2.93292373e-01
-1.90861076e-01 5.66819727e-01 -2.17552409e-01 9.97196972e-01
3.76572385e-02 3.01641282e-02 -2.11937472e-01 -7.52113089e-02
9.38086808e-02 3.91902030e-01 -2.99854487e-01 4.26294595e-01
4.73913401e-01 6.82734191e-01 -7.69433081e-01 -9.11357701e-02
2.30262518e-01 6.23569250e-01 -1.45792961e-01 1.08722141e-02
-1.77614644e-01 2.08076704e-02 1.16051853e-01 5.91975868e-01
4.94377881e-01 3.39780003e-01 3.41441363e-01 7.83662591e-03
-4.69454616e-01 4.90331441e-01 -7.73750365e-01 1.24714661e+00
-7.85966277e-01 1.03780234e+00 1.18902558e-02 -1.15075696e+00
1.02476299e+00 5.14791787e-01 2.01623693e-01 -7.25044549e-01
3.63896936e-01 5.94707489e-01 2.70836443e-01 -3.42500210e-01
6.19304717e-01 -2.31875569e-01 -1.10570595e-01 7.10535586e-01
5.22140503e-01 2.25114986e-01 4.51183349e-01 -1.18546583e-01
4.60183084e-01 9.61929858e-02 2.41388708e-01 -4.69150096e-01
9.83762980e-01 8.66035521e-02 -1.22335792e-01 2.23627254e-01
-4.81388271e-02 3.95848304e-01 5.47447205e-01 -8.38125288e-01
-7.04729259e-01 -6.96665645e-01 -1.25185013e-01 1.69659305e+00
-4.29521561e-01 -2.67657846e-01 -9.04367149e-01 -5.86526573e-01
-2.67408282e-01 6.93696022e-01 -7.26990163e-01 1.56684324e-01
-9.41167176e-01 -1.35474050e+00 9.35912728e-01 2.03543380e-01
4.49951470e-01 -1.51521289e+00 -3.54005039e-01 4.01531756e-01
-1.18657380e-01 -8.02623749e-01 1.88622773e-01 4.59897459e-01
-5.48959911e-01 -9.42528903e-01 -7.42688715e-01 -1.25386846e+00
3.33164632e-01 -1.71213701e-01 1.19828546e+00 -2.43822336e-01
2.16384202e-01 3.16891998e-01 -9.32626784e-01 -8.60283732e-01
-6.19945526e-01 5.00047565e-01 1.67094741e-03 3.44551772e-01
1.00562012e+00 2.42914408e-02 1.81728289e-01 -1.17241882e-01
-1.11359596e+00 -5.09779334e-01 4.06408310e-01 8.21919978e-01
-1.86581135e-01 -2.33128667e-01 8.37709367e-01 -9.74978685e-01
1.12851846e+00 -6.89060390e-01 -2.36147910e-01 2.98939198e-01
-4.11645055e-01 7.74962828e-02 7.25913227e-01 -7.55118281e-02
-9.79041755e-01 -1.76646009e-01 -7.44886816e-01 4.46933419e-01
-2.23102704e-01 1.16517782e+00 1.01559959e-01 1.32346097e-02
8.53627264e-01 2.59805948e-01 3.17343295e-01 -3.99325699e-01
3.26023996e-01 1.03791535e+00 -1.83468834e-01 -3.24269086e-01
2.59842634e-01 2.36409068e-01 -5.26148736e-01 -1.11513042e+00
-5.17207384e-01 -3.01076829e-01 -1.03096664e+00 -2.33396918e-01
9.08832073e-01 -7.41175592e-01 -3.01742494e-01 1.02141261e+00
-1.03046274e+00 -4.49498117e-01 2.80961953e-02 2.43261442e-01
-2.92617470e-01 1.22401580e-01 -1.07181549e+00 -6.34926200e-01
-3.28506112e-01 -1.18517745e+00 6.27906382e-01 -6.69640973e-02
-2.35819116e-01 -1.49712491e+00 2.07309783e-01 -9.40944925e-02
6.75681412e-01 2.19100296e-01 1.41745365e+00 -8.80784094e-01
2.15884745e-01 -3.39383096e-01 -1.66246697e-01 4.97923076e-01
2.19648480e-01 1.12719938e-01 -1.11595166e+00 -2.23570362e-01
-8.23796615e-02 -5.83187938e-01 7.54292667e-01 2.55971491e-01
3.61502588e-01 -2.52135068e-01 2.01152533e-01 -1.16098024e-01
1.46504343e+00 1.90658376e-01 5.99975824e-01 8.82993698e-01
5.23764431e-01 1.08602095e+00 4.82445389e-01 3.36174443e-02
5.86869597e-01 3.24273527e-01 4.08743352e-01 -6.11621402e-02
2.79898047e-01 9.55479443e-02 1.00019133e+00 1.14995444e+00
1.43846581e-02 -3.26113909e-01 -1.35974002e+00 7.48173475e-01
-1.68144464e+00 -5.58627844e-01 -2.77748704e-01 2.03866553e+00
8.49329710e-01 -3.52844521e-02 3.62807214e-01 6.02878511e-01
4.73282933e-01 3.46536994e-01 2.68436760e-01 -1.48908842e+00
-2.71316528e-01 5.12292087e-01 2.96568185e-01 6.25435054e-01
-1.18995678e+00 1.21727431e+00 5.93107986e+00 7.12979555e-01
-1.52245402e+00 2.03315854e-01 6.13705158e-01 4.04678404e-01
-2.07663467e-03 -5.28361917e-01 -6.97422147e-01 2.47395784e-01
1.40903163e+00 1.81160539e-01 3.28258306e-01 3.50607932e-01
4.72441092e-02 -7.81889632e-02 -9.44601715e-01 5.10119081e-01
4.19755042e-01 -9.79575098e-01 3.61365795e-01 -3.64496410e-01
3.32593679e-01 4.61345643e-01 7.16397613e-02 5.39068162e-01
4.03156787e-01 -1.25202513e+00 7.84410715e-01 3.04343160e-02
5.02938807e-01 -9.23600912e-01 1.17079079e+00 1.18831247e-01
-7.65632451e-01 -2.84733146e-01 -3.53270173e-01 -2.17869341e-01
8.41798708e-02 3.24616104e-01 -8.54980886e-01 5.18930256e-01
8.69592845e-01 8.93453598e-01 -6.37093008e-01 6.10691786e-01
-1.70519829e-01 7.34670639e-01 -2.79823780e-01 -4.92947608e-01
1.01049185e+00 -3.37373763e-01 1.17888324e-01 1.61074305e+00
9.37800631e-02 -6.37778521e-01 -1.76869184e-01 1.61276758e-01
2.10771099e-01 6.70728564e-01 -6.48008525e-01 -1.90729558e-01
-1.04474783e-01 1.02213752e+00 -7.81447828e-01 -1.07379153e-01
-7.33523130e-01 8.11921895e-01 3.45187098e-01 1.44646749e-01
-4.54754919e-01 -5.27334511e-01 4.17372465e-01 9.85653773e-02
7.29857609e-02 -4.59596276e-01 -1.03702232e-01 -9.36128497e-01
-2.41785899e-01 -1.18238330e+00 5.61214149e-01 -4.89827007e-01
-1.61642230e+00 1.11353827e+00 -2.42404878e-01 -9.64689851e-01
-3.71319383e-01 -1.32406497e+00 -4.31751996e-01 8.80566776e-01
-1.63507831e+00 -1.54022431e+00 2.07154334e-01 6.24735653e-01
6.92222416e-01 -6.62231684e-01 1.32945991e+00 4.73191828e-01
-3.33321691e-01 2.17031360e-01 4.22091514e-01 5.98029673e-01
9.24487114e-01 -1.50224936e+00 1.85667768e-01 5.60782611e-01
3.20057780e-01 3.48648489e-01 3.06213915e-01 -6.60715327e-02
-1.03613317e+00 -7.66956389e-01 1.51343870e+00 -3.20189238e-01
8.69462907e-01 -5.72248459e-01 -7.32303321e-01 8.54147732e-01
9.00433004e-01 -6.05485201e-01 8.89790654e-01 3.34223717e-01
-4.08869058e-01 -2.13756830e-01 -1.07713640e+00 5.19341171e-01
7.67162582e-03 -5.04117489e-01 -6.41445398e-01 4.26108539e-01
1.76469281e-01 -6.04728721e-02 -6.46581292e-01 4.82866168e-02
5.55182755e-01 -9.88764107e-01 8.40184152e-01 -6.18648410e-01
5.90114534e-01 1.53275833e-01 -9.01587531e-02 -1.68906045e+00
2.07519848e-02 -1.75073549e-01 4.62340474e-01 1.18863654e+00
6.64621234e-01 -9.11007047e-01 2.96230406e-01 1.32792413e-01
5.37474593e-03 -6.15343451e-01 -6.93604648e-01 -2.35630423e-01
7.81022489e-01 -3.71087641e-01 4.41147566e-01 1.20578742e+00
2.16168419e-01 4.70299244e-01 -3.42458040e-02 -2.70327747e-01
-2.10079089e-01 -7.64453635e-02 2.98735082e-01 -1.24193561e+00
3.08183730e-01 -7.82454848e-01 -4.29184377e-01 -5.74528337e-01
2.57345676e-01 -1.01080239e+00 -5.70529439e-02 -1.68771291e+00
-7.78220117e-01 -4.38082010e-01 -1.77101940e-01 6.53448880e-01
3.14933836e-01 3.61932784e-01 4.49422672e-02 -9.20672715e-02
-1.00312389e-01 3.27633351e-01 5.37703276e-01 -2.49405175e-01
-2.63201714e-01 -9.53129753e-02 -4.35081929e-01 7.50035584e-01
1.01445282e+00 -3.15399587e-01 -2.07373984e-02 -8.71164262e-01
7.50346005e-01 -6.42796338e-01 -2.41681620e-01 -7.62034416e-01
-5.77038899e-02 4.56244946e-01 3.02811712e-01 -2.00954571e-01
9.00659710e-02 -6.18036151e-01 -8.85356486e-01 5.42696238e-01
-3.05760741e-01 6.45449281e-01 4.38794672e-01 -7.03586414e-02
-8.56454551e-01 -5.78197181e-01 7.40037739e-01 -2.66339064e-01
-9.37185049e-01 -1.02504663e-01 -1.34375167e+00 -1.96378529e-01
6.92856789e-01 -1.14692822e-01 -3.56930271e-02 -4.56558138e-01
-7.87835777e-01 4.91333380e-02 -6.69791847e-02 8.04591238e-01
3.80176187e-01 -1.01360905e+00 -1.03330076e+00 2.71503657e-01
1.14044674e-01 -2.86449909e-01 -3.18202317e-01 7.85649717e-01
-1.25400579e+00 5.11994362e-01 -6.36032581e-01 -2.71857023e-01
-9.81464803e-01 1.62665755e-01 3.72499764e-01 -5.49054384e-01
-9.13293660e-02 8.42834473e-01 -3.17026228e-01 -1.12653315e+00
4.96438183e-02 -9.26533639e-02 -9.60443556e-01 9.28443909e-01
3.81548971e-01 3.09367567e-01 6.26027048e-01 -1.21815538e+00
-2.19420969e-01 1.71604037e-01 -2.42776349e-01 -2.47034535e-01
1.55177259e+00 5.25721163e-02 -7.78028131e-01 9.13771749e-01
1.10841560e+00 7.16707110e-02 -1.84841543e-01 -1.70538485e-01
5.72014391e-01 2.39499081e-02 -7.43210763e-02 -9.01087224e-01
-8.89787078e-01 1.36204565e+00 7.98069417e-01 6.55871987e-01
9.46874917e-01 -5.97457767e-01 6.84522510e-01 5.13157248e-01
2.95751274e-01 -1.24952424e+00 -1.76804230e-01 1.30396450e+00
9.86669362e-01 -1.16357279e+00 -4.91697431e-01 1.31439250e-02
-5.89929402e-01 1.62143517e+00 3.17942649e-01 -4.30611968e-01
1.11875284e+00 1.74033582e-01 6.21034026e-01 -1.80940121e-01
-2.27759346e-01 -4.31805760e-01 8.40809494e-02 7.55695105e-01
1.10071445e+00 -1.16462283e-01 -4.84775990e-01 4.77668554e-01
-3.88731062e-01 -2.92569071e-01 7.02269673e-01 9.35912788e-01
-2.78182507e-01 -1.50388396e+00 -4.15609330e-01 4.94866908e-01
-7.77618349e-01 -2.68284410e-01 -7.39382744e-01 9.81797218e-01
4.30644900e-02 1.05798507e+00 1.20530732e-01 -2.53174365e-01
1.63763747e-01 4.67925727e-01 3.26296777e-01 -7.13719606e-01
-1.37001967e+00 -3.48277539e-01 2.57884443e-01 -2.73274407e-02
-6.54666007e-01 -6.43515348e-01 -1.10749185e+00 -4.53889042e-01
-3.52645069e-01 6.61802664e-02 7.90881336e-01 1.10884058e+00
-8.65585580e-02 1.02704369e-01 4.10751760e-01 -1.00280547e+00
-1.38218060e-01 -1.37082672e+00 -5.08203804e-01 2.34545305e-01
2.18155801e-01 -3.41082245e-01 -2.38113441e-02 -2.54898556e-02] | [11.148672103881836, 7.1966071128845215] |
1658c0ed-7d43-4c8c-8652-732114104f66 | learning-multi-scale-deep-features-for-high | 1611.03591 | null | http://arxiv.org/abs/1611.03591v1 | http://arxiv.org/pdf/1611.03591v1.pdf | Learning Multi-Scale Deep Features for High-Resolution Satellite Image Classification | In this paper, we propose a multi-scale deep feature learning method for
high-resolution satellite image classification. Specifically, we firstly warp
the original satellite image into multiple different scales. The images in each
scale are employed to train a deep convolutional neural network (DCNN).
However, simultaneously training multiple DCNNs is time-consuming. To address
this issue, we explore DCNN with spatial pyramid pooling (SPP-net). Since
different SPP-nets have the same number of parameters, which share the
identical initial values, and only fine-tuning the parameters in
fully-connected layers ensures the effectiveness of each network, thereby
greatly accelerating the training process. Then, the multi-scale satellite
images are fed into their corresponding SPP-nets respectively to extract
multi-scale deep features. Finally, a multiple kernel learning method is
developed to automatically learn the optimal combination of such features.
Experiments on two difficult datasets show that the proposed method achieves
favorable performance compared to other state-of-the-art methods. | ['Zhi Li', 'Renlong Hang', 'Qingshan Liu', 'Huihui Song'] | 2016-11-11 | null | null | null | null | ['satellite-image-classification'] | ['computer-vision'] | [ 2.72955024e-03 -6.01735234e-01 5.69465756e-02 -4.10297155e-01
-6.67883992e-01 -3.34283978e-01 2.34810919e-01 -1.80756509e-01
-7.84400165e-01 4.94867802e-01 -1.29028887e-01 -2.77431915e-03
-2.06837267e-01 -1.20603228e+00 -6.47838056e-01 -9.39669907e-01
-3.35390359e-01 -2.23922685e-01 5.74833393e-01 -1.00276047e-05
5.75738288e-02 8.70263636e-01 -1.33449686e+00 1.39038906e-01
8.44417930e-01 1.12559378e+00 4.77759421e-01 3.40633512e-01
-2.90425378e-03 5.91630399e-01 -4.20048058e-01 1.18456803e-01
3.22026789e-01 3.90432663e-02 -5.64954042e-01 1.96216017e-01
2.64049053e-01 -6.53166831e-01 -5.97242534e-01 1.29715657e+00
4.35496092e-01 3.35507661e-01 2.95073479e-01 -8.56739879e-01
-7.49328136e-01 4.15964693e-01 -8.43815029e-01 5.22413969e-01
-4.37026739e-01 2.77493745e-02 8.95064712e-01 -6.79247260e-01
8.18778053e-02 1.21411026e+00 7.06973374e-01 -3.11876144e-02
-1.17227805e+00 -9.29554939e-01 2.55209748e-02 4.19537425e-02
-1.77409530e+00 -4.09841649e-02 6.71254873e-01 -2.39941105e-01
5.60036957e-01 -9.02244002e-02 5.94715834e-01 3.79538208e-01
1.63766935e-01 6.14208996e-01 1.18318892e+00 1.55725449e-01
-2.42068414e-02 -1.38490796e-01 5.69051765e-02 7.80019939e-01
8.62795264e-02 -1.35025844e-01 -7.20450431e-02 -3.43669914e-02
1.28849578e+00 4.49510992e-01 -3.75907123e-01 1.26513049e-01
-1.17913210e+00 9.13568139e-01 1.08808208e+00 5.52711248e-01
-4.88322020e-01 -3.15556899e-02 3.23334754e-01 2.78684795e-01
3.32714677e-01 1.15768164e-01 -5.24363935e-01 5.02151966e-01
-9.06252682e-01 9.58758220e-02 1.53765485e-01 3.05775404e-01
1.36353791e+00 -1.19686671e-01 -3.52470130e-02 9.64812875e-01
1.61681905e-01 2.19922349e-01 6.59021795e-01 -3.57198298e-01
5.08078694e-01 9.14894402e-01 3.50522762e-03 -1.34841847e+00
-6.98504269e-01 -4.86528933e-01 -1.23467207e+00 2.85283864e-01
9.47626680e-02 -2.21345752e-01 -9.22307789e-01 1.28618848e+00
1.89732283e-01 5.49284220e-01 3.11363250e-01 1.17895353e+00
7.19528794e-01 1.11805761e+00 2.51599196e-02 1.35225952e-01
1.36571193e+00 -8.35797310e-01 -2.86987156e-01 -3.94521087e-01
2.48122156e-01 -5.87017834e-01 1.07871664e+00 4.23959792e-02
-5.20290732e-01 -7.76004434e-01 -1.27233648e+00 8.57668668e-02
-3.00456375e-01 6.34907544e-01 6.26773119e-01 -5.61701180e-03
-9.19908285e-01 8.30212772e-01 -8.51108909e-01 -6.02300167e-02
5.82772374e-01 4.32481170e-01 -3.99930090e-01 -9.84809846e-02
-1.39765704e+00 5.41759551e-01 7.19637275e-01 4.28485483e-01
-8.04656208e-01 -3.88771951e-01 -7.70483732e-01 2.27355093e-01
4.96295048e-03 -6.75879344e-02 9.83612359e-01 -1.03101242e+00
-1.46248817e+00 5.76877773e-01 1.95055544e-01 -8.87153298e-02
1.65976077e-01 -8.60831514e-02 -4.92884248e-01 2.42065713e-01
1.43895417e-01 5.70091248e-01 9.04229581e-01 -8.58922720e-01
-9.40471113e-01 -2.75100291e-01 2.41977856e-01 1.88132569e-01
-6.87117755e-01 2.32546076e-01 -4.83172148e-01 -4.91959065e-01
3.54775369e-01 -5.60861051e-01 -3.96376282e-01 8.44092444e-02
-7.47245848e-02 -3.55288982e-01 8.98099720e-01 -6.54364169e-01
1.02216578e+00 -2.49763107e+00 2.01786727e-01 1.23885415e-01
6.42001778e-02 4.02535707e-01 -3.38851094e-01 -4.58749793e-02
-2.67639812e-02 1.01603866e-01 -2.58472562e-01 3.86209227e-02
-3.45992893e-01 3.21055472e-01 -1.70650840e-01 5.98350585e-01
5.35673976e-01 7.29991078e-01 -6.10722125e-01 -5.72585940e-01
2.57042617e-01 5.90107024e-01 -1.41046956e-01 3.70515406e-01
1.40728340e-01 2.84562916e-01 -8.14199805e-01 5.31450510e-01
1.06165588e+00 -3.48898500e-01 -1.04869075e-01 -5.57734132e-01
-4.52925742e-01 -2.63243556e-01 -1.25254512e+00 1.38266909e+00
-4.64669734e-01 3.56754005e-01 1.41925393e-02 -1.16035807e+00
1.16891289e+00 1.73108547e-03 2.29927063e-01 -6.19835317e-01
1.92379117e-01 2.47892946e-01 -2.14108557e-01 -5.08691847e-01
3.68596762e-01 -2.93380648e-01 -6.68241084e-02 2.09279954e-01
-1.75969079e-02 1.67530596e-01 2.47449353e-02 -4.23562169e-01
5.87160230e-01 -1.25963584e-01 2.22857684e-01 -2.70475149e-01
8.20164263e-01 -2.78062969e-01 8.02845657e-01 4.48777676e-01
-2.10267276e-01 4.24657702e-01 3.36647779e-01 -9.27653909e-01
-9.57295656e-01 -7.37993896e-01 -3.40276599e-01 1.00989401e+00
3.07710618e-01 1.20763339e-01 -5.69201767e-01 -5.95732391e-01
6.41801208e-02 -7.37514570e-02 -7.39098668e-01 1.15221933e-01
-6.40863776e-01 -1.15019643e+00 7.61565328e-01 7.15046167e-01
1.12615728e+00 -1.11472929e+00 -5.93677580e-01 4.22745436e-01
3.92697044e-02 -1.02551925e+00 -2.54918307e-01 9.50256586e-02
-1.03432858e+00 -9.03483689e-01 -7.97638893e-01 -1.04534388e+00
6.32887304e-01 4.94741410e-01 4.36002046e-01 1.29173651e-01
-4.44169752e-02 -5.97832799e-01 -2.67613530e-01 2.99638778e-01
1.45359308e-01 2.54064143e-01 -1.16145894e-01 3.07039380e-01
3.60051095e-01 -7.86623478e-01 -6.89818442e-01 1.97256967e-01
-1.35161042e+00 -6.85673058e-02 1.09089828e+00 8.13470960e-01
6.11819565e-01 5.78075767e-01 3.66943210e-01 -1.76638991e-01
5.25064170e-01 -2.79788047e-01 -8.47795665e-01 1.97760388e-01
-5.68616353e-02 6.55638054e-02 1.01705003e+00 -5.72773218e-01
-7.46946871e-01 1.43171638e-01 -8.57576579e-02 -3.20517212e-01
-2.52919704e-01 9.13379431e-01 -1.10451870e-01 -3.39995772e-01
3.75355512e-01 7.06335723e-01 -1.87316850e-01 -6.34575248e-01
1.44889876e-01 8.24898601e-01 4.66208279e-01 -2.05750898e-01
1.00939953e+00 3.89010102e-01 -2.53172815e-01 -7.42283821e-01
-9.62836444e-01 -1.30784407e-01 -7.85547614e-01 8.52901116e-02
1.05113351e+00 -1.19876909e+00 -5.16693711e-01 9.42144752e-01
-9.76768315e-01 -2.14555442e-01 3.45358163e-01 5.94052136e-01
-4.68140766e-02 2.57784814e-01 -7.52352715e-01 -3.48788321e-01
-3.68441254e-01 -1.29663301e+00 1.04139316e+00 9.69871044e-01
6.97788477e-01 -7.09879398e-01 -1.17136322e-01 -1.14758171e-01
5.34300983e-01 4.60168384e-02 7.69931853e-01 -4.79051262e-01
-6.14644885e-01 -2.49615595e-01 -8.31266344e-01 5.17432094e-01
2.72626102e-01 1.59981623e-02 -6.84402168e-01 -5.32538593e-01
-2.18661398e-01 -4.57656711e-01 1.08110595e+00 3.82040918e-01
1.45636475e+00 -2.14510962e-01 -9.24957395e-02 1.04893231e+00
1.79298770e+00 3.35276760e-02 6.04285181e-01 8.97490382e-01
6.32088780e-01 1.93819135e-01 4.34556693e-01 4.41814423e-01
3.72180670e-01 4.08724815e-01 3.89341623e-01 -3.80891681e-01
4.65235621e-01 5.47293350e-02 2.46638536e-01 5.85024536e-01
-1.05529934e-01 4.14804637e-01 -7.12203681e-01 3.99273694e-01
-1.69649982e+00 -8.85771513e-01 1.82838455e-01 1.86386681e+00
8.73391271e-01 4.42359298e-02 -2.32000127e-01 -1.31639093e-01
8.26360643e-01 6.29555821e-01 -6.09370887e-01 1.88199654e-01
-2.41371647e-01 2.23614648e-01 7.02593029e-01 2.19053686e-01
-1.64259863e+00 1.11828399e+00 5.38997412e+00 8.80106151e-01
-1.66089046e+00 -8.30026250e-03 4.41074222e-01 2.64455855e-01
8.88013765e-02 -1.54225722e-01 -8.45048189e-01 3.42124343e-01
4.71194357e-01 -1.09504364e-01 5.61960638e-01 9.13510263e-01
2.65846342e-01 6.09483421e-02 -2.99920559e-01 7.80833364e-01
-1.44915059e-01 -1.06230819e+00 1.24691620e-01 -2.12000981e-01
7.14282751e-01 3.41221750e-01 -1.08250953e-01 3.65561396e-01
1.09049052e-01 -8.42433095e-01 3.16504836e-01 4.74077076e-01
5.84995627e-01 -1.12498653e+00 8.89604032e-01 2.97431409e-01
-1.67653310e+00 -3.51462483e-01 -1.02940023e+00 -3.27274501e-02
-1.87428966e-01 6.36800706e-01 6.32477850e-02 6.59265697e-01
9.68010306e-01 7.48415709e-01 -6.69650257e-01 9.87043917e-01
-2.91673690e-01 2.85914809e-01 -3.12412888e-01 2.38071755e-01
6.34890616e-01 -3.34032357e-01 2.45666597e-02 1.15585315e+00
4.78841782e-01 3.13901991e-01 5.44462502e-01 6.21499240e-01
-1.56052545e-01 1.33010253e-01 -2.13658720e-01 -1.44155368e-01
4.39548463e-01 1.82016575e+00 -6.94740295e-01 -3.72683764e-01
-6.78240597e-01 1.05803525e+00 6.80053711e-01 3.38638753e-01
-7.46690512e-01 -8.48318279e-01 8.52985442e-01 -3.78944367e-01
6.52476609e-01 -3.72277617e-01 1.25778869e-01 -1.36975563e+00
-1.49532661e-01 -7.05562592e-01 3.38957012e-01 -6.44833803e-01
-1.24329293e+00 8.59479249e-01 -2.95016259e-01 -1.39355528e+00
3.66387933e-01 -6.86053157e-01 -8.30136538e-01 1.08940876e+00
-2.17226338e+00 -1.34353602e+00 -5.36582410e-01 7.12552071e-01
3.00224483e-01 -4.45144027e-02 6.22272432e-01 3.64109874e-01
-8.34783792e-01 3.70283276e-01 3.67001325e-01 5.80143750e-01
4.99068141e-01 -9.14007604e-01 2.58508950e-01 9.65104997e-01
-3.62598330e-01 4.64554280e-01 5.63320443e-02 -3.46593678e-01
-1.16691601e+00 -1.40855122e+00 5.10796130e-01 6.41539931e-01
7.81161070e-01 7.71420375e-02 -1.29549575e+00 5.49391448e-01
-7.78570995e-02 4.51562732e-01 5.99900782e-01 -2.93893725e-01
-2.82731503e-01 -4.93708134e-01 -9.70841646e-01 3.07231516e-01
4.18789476e-01 -5.59190989e-01 -6.79236591e-01 1.18819319e-01
7.62709916e-01 -1.79452315e-01 -1.05049598e+00 5.23504198e-01
4.55758214e-01 -8.80907416e-01 8.99836838e-01 -5.10697961e-01
5.64829886e-01 -7.64607549e-01 -1.71711862e-01 -1.49177158e+00
-7.08337247e-01 5.72193190e-02 4.39877033e-01 9.08794463e-01
1.23136267e-01 -7.59416938e-01 4.43780601e-01 1.62444845e-01
1.89784225e-02 -7.39707351e-01 -8.05777431e-01 -5.85286140e-01
3.86575721e-02 1.16140254e-01 9.03387010e-01 1.16899061e+00
-5.48642933e-01 1.53286517e-01 -3.60493153e-01 9.42525804e-01
6.71262264e-01 5.36328137e-01 5.11839330e-01 -1.30667150e+00
-1.46802187e-01 -3.89667481e-01 -4.96750414e-01 -1.02994931e+00
9.71054286e-02 -6.73648596e-01 -2.52124425e-02 -1.44765866e+00
3.59243989e-01 -3.57961178e-01 -6.36833191e-01 8.60122859e-01
-5.44672549e-01 3.74353647e-01 -4.04224545e-03 4.36212569e-01
-1.94121644e-01 6.61167979e-01 1.31321847e+00 -5.80225885e-02
-3.16863626e-01 -5.54072633e-02 -5.72118282e-01 7.12949097e-01
9.96172428e-01 -4.30596590e-01 6.10206723e-02 -6.56469166e-01
-2.22102299e-01 -3.01551241e-02 4.70056117e-01 -1.22582853e+00
2.82348812e-01 -2.92665958e-01 7.82142103e-01 -5.62631130e-01
-4.20101024e-02 -7.37681150e-01 -5.49248792e-02 5.23351073e-01
-1.40312031e-01 1.26546901e-02 3.21412563e-01 2.93525994e-01
-4.21794176e-01 -1.85716629e-01 1.00181341e+00 -2.12481961e-01
-1.01370430e+00 7.10265636e-01 -2.48375505e-01 -5.01815856e-01
8.29969645e-01 6.25385717e-02 -2.90268838e-01 1.95803031e-01
-5.56494713e-01 3.12480122e-01 2.04673171e-01 3.83697361e-01
5.92966378e-01 -1.44198358e+00 -7.31804848e-01 4.37026888e-01
-1.08947586e-02 2.84752160e-01 5.70225954e-01 5.81247509e-01
-6.32884085e-01 -1.89848691e-02 -5.64442337e-01 -5.29809773e-01
-1.01763427e+00 2.94062287e-01 6.76859677e-01 -4.21246916e-01
-5.87111175e-01 6.35626674e-01 9.97730494e-02 -7.13228762e-01
-5.54190986e-02 -3.03723067e-01 -4.41678673e-01 2.18165308e-01
8.08235943e-01 -4.83617224e-02 -1.91623881e-01 -6.41596854e-01
-3.76950651e-01 9.08896744e-01 -2.32391551e-01 1.18612908e-01
1.79938865e+00 3.08340304e-02 -4.61586565e-01 1.34424299e-01
1.52121401e+00 -4.57386553e-01 -1.55788910e+00 -6.35586619e-01
-1.85516670e-01 -5.63602149e-01 4.83961403e-01 -3.31093967e-01
-1.39976537e+00 9.54464018e-01 7.96593904e-01 1.50170065e-02
1.39653623e+00 -2.40230218e-01 7.87930787e-01 7.32941270e-01
1.24580093e-01 -9.46192622e-01 -4.00821567e-02 5.95785916e-01
6.97808504e-01 -1.12918067e+00 1.13122061e-01 1.37300476e-01
-4.56945598e-01 1.47918940e+00 8.03958833e-01 -4.59782988e-01
6.47780895e-01 1.46991417e-01 2.24946603e-01 -8.29332396e-02
-1.09148644e-01 -3.74654800e-01 7.04000518e-02 1.58675656e-01
1.56737059e-01 1.30202780e-02 -2.45041788e-01 5.43903708e-01
1.37483075e-01 2.18353629e-01 3.04763317e-01 8.88365626e-01
-8.71222973e-01 -9.09776747e-01 -4.07069474e-01 4.28169340e-01
-4.04541552e-01 -8.93436894e-02 1.16198070e-01 6.46032751e-01
9.35669392e-02 4.70454276e-01 1.47377700e-01 -6.73592806e-01
3.76061231e-01 -1.69668645e-01 4.44573872e-02 -2.15739444e-01
-4.86081153e-01 -2.97337119e-02 -5.33789933e-01 -3.59884024e-01
-5.73069036e-01 -3.28652143e-01 -1.32146466e+00 -2.43521556e-01
-2.55827636e-01 1.37011036e-01 3.87339681e-01 1.00848663e+00
1.42758250e-01 4.74458575e-01 1.12395012e+00 -1.14952445e+00
-8.32189083e-01 -1.07453346e+00 -8.01947773e-01 1.30880490e-01
4.48598593e-01 -4.91409749e-01 -3.01998258e-01 -1.93781808e-01] | [9.886358261108398, -1.4813718795776367] |
a230dc40-c232-4540-a5d7-570fd2402d3b | seastar-vertex-centric-programming-for-graph | null | null | https://dl.acm.org/doi/10.1145/3447786.3456247 | https://dl.acm.org/doi/pdf/10.1145/3447786.3456247 | Seastar: vertex-centric programming for graph neural networks | Graph neural networks (GNNs) have achieved breakthrough performance in graph analytics such as node classification, link prediction and graph clustering. Many GNN training frameworks have been developed, but they are usually designed as a set of manually written, GNN-specific operators plugged into existing deep learning systems, which incurs high memory consumption, poor data locality, and large semantic gap between algorithm design and implementation. This paper proposes the Seastar system, which presents a vertex-centric programming model for GNN training on GPU and provides idiomatic python constructs to enable easy development of novel homogeneous and heterogeneous GNN models. We also propose novel optimizations to produce highly efficient fused GPU kernels for forward and backward passes in GNN training. Compared with the state-of-the art GNN systems, DGL and PyG, Seastar achieves better usability, up to 2 and 8 times less memory consumption, and 14 and 3 times faster execution, respectively. | ['Fan Yu', 'James Cheng', 'Chenguang Zheng', 'Boyang Li', 'Tatiana Jin', 'Zhenkun Cai', 'Kaihao Ma', 'Yidi Wu'] | 2021-04-21 | null | null | null | proceedings-of-the-sixteenth-european | ['graph-clustering'] | ['graphs'] | [-5.45588791e-01 -1.54297084e-01 -3.02658647e-01 -2.00812921e-01
3.73940431e-02 -3.83403689e-01 4.15190488e-01 3.93546134e-01
-1.89488590e-01 2.06604078e-01 -4.20556515e-01 -9.26401675e-01
3.83554921e-02 -1.49897075e+00 -5.24306953e-01 -4.36414778e-01
-2.16705844e-01 6.42540812e-01 5.44302642e-01 -2.47807845e-01
-2.39501894e-01 6.05296969e-01 -1.36334419e+00 3.18359844e-02
9.09051776e-01 6.60351992e-01 -1.96198188e-02 1.00578296e+00
-4.27935928e-01 9.30014074e-01 -2.77842641e-01 -7.52019405e-01
3.45126480e-01 -9.28121805e-02 -5.99862516e-01 -5.28526127e-01
3.89771223e-01 -1.57188922e-01 -6.74835503e-01 9.64927971e-01
5.44095814e-01 -6.62263185e-02 7.83343613e-02 -1.69528604e+00
-4.48935866e-01 9.56851602e-01 -6.31091654e-01 -3.45359817e-02
-2.36283652e-02 1.82084441e-01 8.86288762e-01 -3.81244987e-01
5.66940308e-01 1.01744294e+00 1.18906379e+00 3.41602772e-01
-1.13045394e+00 -7.38475382e-01 -1.19045615e-01 1.71928659e-01
-1.35958815e+00 -3.93140949e-02 6.39065206e-01 -1.85176477e-01
1.42520189e+00 3.58067930e-01 1.10722077e+00 1.00129473e+00
2.44197890e-01 6.04860425e-01 2.70711035e-01 -1.87790483e-01
1.02924824e-01 -2.83849776e-01 2.61079490e-01 1.18741143e+00
4.46572423e-01 -5.70285618e-02 -2.49358729e-01 -3.31148446e-01
1.11790621e+00 -1.94208790e-02 1.53677374e-01 -4.36780721e-01
-8.96177113e-01 8.66510332e-01 1.11111927e+00 3.30888154e-03
-2.66063422e-01 6.25963569e-01 1.08601344e+00 3.62841368e-01
4.06166524e-01 -7.30874985e-02 -4.56116199e-01 -1.07306175e-01
-7.21180916e-01 1.64999634e-01 1.19152188e+00 1.12577486e+00
8.73962641e-01 1.96910232e-01 -1.97987091e-02 4.31447268e-01
2.42910117e-01 2.01221943e-01 1.04547270e-01 -5.11869788e-01
3.74258786e-01 1.16340005e+00 -9.25778866e-01 -1.39755857e+00
-7.74193883e-01 -7.75048673e-01 -1.07134664e+00 -1.08378768e-01
2.78903805e-02 -2.08769456e-01 -9.42062140e-01 1.39236403e+00
5.12945592e-01 4.77931440e-01 -9.20133442e-02 6.58883631e-01
1.30545151e+00 6.41435444e-01 2.98773766e-01 5.73978305e-01
1.06117260e+00 -1.56467938e+00 -1.37432873e-01 -3.22188467e-01
1.37419260e+00 -4.91480321e-01 1.05271041e+00 7.75917666e-03
-8.76910150e-01 -3.99387389e-01 -8.07590485e-01 -4.43797737e-01
-5.40272951e-01 1.39665976e-03 1.45451510e+00 8.58306646e-01
-1.57226694e+00 7.58064091e-01 -1.14117849e+00 -5.63472807e-01
4.20451581e-01 5.59684813e-01 -3.92601967e-01 -1.69657059e-02
-6.89067125e-01 3.53141695e-01 8.20523858e-01 6.08186573e-02
-7.63058484e-01 -8.39310408e-01 -8.71023774e-01 3.62146974e-01
3.25963259e-01 -1.05987132e+00 1.03866649e+00 -8.05629075e-01
-1.46461093e+00 8.04625571e-01 3.91859055e-01 -4.80538964e-01
3.38364244e-01 1.26672536e-01 -5.07673085e-01 -1.61808789e-01
-1.60773963e-01 4.73755926e-01 2.83260405e-01 -4.96889353e-01
-3.73649031e-01 -2.19611287e-01 1.94985509e-01 9.10517797e-02
-3.35612506e-01 8.20415765e-02 -1.00646973e+00 -5.80951810e-01
1.01381801e-01 -8.80791306e-01 -4.42721337e-01 3.91174294e-02
-4.69078153e-01 -2.34945640e-01 9.63011980e-01 -3.82976353e-01
1.21604443e+00 -1.82492471e+00 -1.18391946e-01 5.89804590e-01
1.00679445e+00 6.32139683e-01 -1.61942124e-01 4.33923960e-01
2.18896866e-02 -1.78837597e-01 3.22172225e-01 -7.71636367e-02
1.35493493e-02 4.14327174e-01 1.68293059e-01 2.47612521e-01
-2.02939093e-01 1.27371693e+00 -1.08531225e+00 -5.13166964e-01
1.56825215e-01 3.89618009e-01 -8.17813277e-01 2.16898452e-02
-3.91864836e-01 -1.63888559e-02 -3.69132072e-01 6.09828591e-01
8.49054933e-01 -8.16106021e-01 5.96196771e-01 -7.58550391e-02
1.24684446e-01 3.28914911e-01 -8.54055762e-01 1.59745967e+00
-4.00276303e-01 5.90882778e-01 1.07515283e-01 -8.92294586e-01
9.22250450e-01 2.97048106e-03 1.43107951e-01 -5.83925426e-01
3.93203825e-01 1.17691576e-01 1.26320660e-01 -2.31608570e-01
4.13848162e-01 6.36705935e-01 1.44532263e-01 3.43167245e-01
2.55240113e-01 3.03643137e-01 4.52332169e-01 5.04869401e-01
1.62733352e+00 -2.21272223e-02 8.80692899e-02 -3.32765549e-01
3.51706326e-01 2.01373830e-01 3.43228966e-01 7.86140978e-01
1.25776350e-01 1.21252201e-01 8.32588315e-01 -9.29762423e-01
-1.10169482e+00 -8.38513911e-01 3.38208377e-01 1.58694553e+00
-1.75391324e-02 -1.24192798e+00 -8.93960357e-01 -6.68347418e-01
-6.62071677e-03 2.65903473e-01 -2.87887901e-01 -1.40399218e-01
-6.34105504e-01 -9.25505161e-01 9.62415874e-01 7.47244775e-01
8.04365575e-01 -1.04433250e+00 -2.05412120e-01 3.26473534e-01
5.28666854e-01 -1.12570560e+00 -1.23595893e-01 8.39689746e-02
-1.07281518e+00 -1.08211207e+00 -1.15337521e-01 -1.10039878e+00
7.01465726e-01 4.25760090e-01 1.59020388e+00 7.14257061e-01
-2.24342272e-01 -1.48079515e-01 -1.60385266e-01 -1.15392603e-01
-3.52778196e-01 6.51874423e-01 -2.22529590e-01 -4.60108280e-01
3.65059972e-01 -8.74310255e-01 -4.78977025e-01 7.59150237e-02
-5.95824063e-01 5.89543223e-01 4.75488633e-01 7.98918545e-01
2.98440546e-01 -3.72475423e-02 -8.42950121e-02 -1.45538783e+00
6.67596340e-01 -3.44418436e-01 -9.77489591e-01 2.29162499e-01
-8.85753989e-01 -7.72603452e-02 1.07276392e+00 -1.58327758e-01
-5.74264050e-01 -1.34527490e-01 -4.36188787e-01 -4.05432731e-01
7.83131868e-02 8.51958752e-01 -8.68322551e-02 -7.25931108e-01
9.27135229e-01 -1.91037115e-02 1.73822850e-01 -4.26420510e-01
2.76392460e-01 1.10021770e-01 3.23207974e-01 -6.95908308e-01
6.84391499e-01 2.41322577e-01 1.29298300e-01 -5.97553849e-01
-3.10256183e-01 -2.72270411e-01 -3.19370300e-01 -1.70744389e-01
4.54013556e-01 -9.09630358e-01 -1.07834673e+00 7.42328107e-01
-1.10737193e+00 -8.21893513e-01 2.88983077e-01 8.73520300e-02
-1.02777302e-01 2.37104163e-01 -1.12817836e+00 -6.44461662e-02
-9.23719704e-01 -1.19048595e+00 8.20487440e-01 4.28617060e-01
4.37554903e-02 -1.35741341e+00 -5.50607406e-02 -1.85672231e-02
7.56155431e-01 1.20676108e-01 1.13875639e+00 -6.19879782e-01
-7.65286088e-01 -2.97500104e-01 -8.11322272e-01 -4.29556333e-02
-3.44605505e-01 2.50900477e-01 -4.78103876e-01 -5.87900221e-01
-7.55679309e-01 -1.11493908e-01 5.87237537e-01 1.29809871e-01
1.50188565e+00 -3.34134012e-01 -8.39943469e-01 1.52916408e+00
1.65637207e+00 -1.37184232e-01 5.95802844e-01 4.89839882e-01
1.44681156e+00 2.45917123e-02 -1.02814399e-01 1.08197778e-01
5.49215376e-01 4.00027215e-01 7.25455105e-01 -5.77254534e-01
-4.21189480e-02 -3.19952279e-01 -1.60658304e-02 1.10710406e+00
-2.05112204e-01 -4.22013134e-01 -1.33454680e+00 2.10544601e-01
-2.14848375e+00 -5.36894321e-01 -6.56683207e-01 1.76984406e+00
3.66819561e-01 1.92350134e-01 1.79603502e-01 -2.49334678e-01
5.34729242e-01 1.39930859e-01 -5.24206221e-01 -4.47316349e-01
1.47958457e-01 5.19672275e-01 7.24088013e-01 1.40959546e-01
-9.20162737e-01 1.33884466e+00 6.51606798e+00 1.02812779e+00
-1.37478530e+00 8.18322471e-04 4.22817677e-01 1.07744902e-01
-2.39243358e-01 2.28229061e-01 -6.33930862e-01 2.83914506e-01
1.00296426e+00 -3.52010399e-01 4.97186989e-01 1.25321150e+00
-1.78393856e-01 3.86380255e-01 -8.55103552e-01 1.08953965e+00
-3.22086513e-01 -1.76131511e+00 1.69226125e-01 1.09876081e-01
6.16504788e-01 6.09762073e-01 -3.33057225e-01 5.66331446e-01
7.95493722e-01 -1.05124128e+00 2.63582289e-01 -4.87718247e-02
5.44794381e-01 -1.01643944e+00 8.03071022e-01 1.83182657e-01
-1.41428709e+00 1.38523459e-01 -5.10215819e-01 -3.63256931e-01
-1.34913772e-01 6.22562587e-01 -1.00412357e+00 8.45464766e-01
8.28262866e-01 6.83319867e-01 -8.76383424e-01 1.07027495e+00
-2.61066288e-01 7.47896373e-01 -4.16442007e-01 -3.52702707e-01
3.86837572e-01 -4.52389210e-01 2.60857135e-01 1.26735961e+00
2.22061798e-01 -1.88910261e-01 3.65155816e-01 7.45083213e-01
-3.33124936e-01 2.80492276e-01 -5.59976697e-01 -2.33433358e-02
3.74084979e-01 1.60748255e+00 -1.11217892e+00 -3.52744728e-01
-6.64243758e-01 8.50783885e-01 9.11556840e-01 2.81588376e-01
-1.28924060e+00 -7.09591985e-01 7.45358586e-01 1.37525603e-01
7.65772909e-02 -3.29129964e-01 -3.43094558e-01 -1.06506693e+00
-1.79422304e-01 -8.25938463e-01 5.07911146e-01 -6.12788439e-01
-1.14204419e+00 7.63368964e-01 -5.09522974e-01 -6.69881582e-01
9.67270602e-03 -9.25390601e-01 -9.65528369e-01 7.52535522e-01
-1.02050328e+00 -1.39862704e+00 -7.88788795e-01 5.98430634e-01
-2.09663749e-01 -3.69899899e-01 8.35204422e-01 6.83058262e-01
-9.03657436e-01 8.82849157e-01 1.16693430e-01 5.34749687e-01
2.95964271e-01 -1.11476147e+00 1.27846038e+00 8.51398706e-01
-1.02370661e-02 7.18989730e-01 3.83997262e-01 -7.77265310e-01
-1.72025621e+00 -1.44758260e+00 6.00616753e-01 7.54488036e-02
8.27907562e-01 -6.51885450e-01 -9.08770263e-01 1.01981664e+00
-9.68813300e-02 4.34414685e-01 5.66436708e-01 6.87715054e-01
-4.53944087e-01 -3.85973901e-01 -8.79052281e-01 1.09593940e+00
1.42745161e+00 -4.41279233e-01 4.27087516e-01 6.27423286e-01
8.42505038e-01 -9.52366233e-01 -8.06517601e-01 3.43252212e-01
4.73716527e-01 -1.10929930e+00 9.89606202e-01 -6.45337164e-01
2.18680263e-01 -2.71986544e-01 3.62174869e-01 -1.07611644e+00
-6.27628565e-01 -6.22253180e-01 -2.90203482e-01 9.09985900e-01
3.55600715e-01 -8.06104243e-01 1.53442705e+00 2.40051270e-01
-3.54959726e-01 -8.25173497e-01 -3.69585931e-01 -6.48482859e-01
-3.80146831e-01 -4.54900444e-01 1.01205039e+00 1.05586720e+00
-2.11231753e-01 6.15187287e-01 -1.50002874e-02 2.57670730e-01
5.48888505e-01 5.31137409e-03 1.34317625e+00 -1.34138620e+00
-3.56304199e-01 -6.91431999e-01 -8.94942224e-01 -9.50436354e-01
1.11873448e-01 -1.44838810e+00 -5.01333117e-01 -1.67922688e+00
1.34509027e-01 -7.17952490e-01 4.99222353e-02 8.49373519e-01
9.80743580e-03 3.46836060e-01 -6.76617324e-02 -9.82642621e-02
-7.32892573e-01 2.31039077e-01 9.81751740e-01 -2.88999993e-02
-2.97603279e-01 -6.63605407e-02 -4.64819849e-01 7.53946543e-01
7.22540259e-01 -3.71819496e-01 -4.66437459e-01 -9.49984252e-01
5.65743148e-01 -1.77323952e-01 5.68507671e-01 -1.25172102e+00
6.89284384e-01 1.81106880e-01 2.72316545e-01 -6.21386409e-01
-6.72905892e-02 -5.69767654e-01 6.28749371e-01 5.53072333e-01
2.54613072e-01 5.41130364e-01 3.27831089e-01 3.22984546e-01
-5.83001599e-03 1.17860653e-01 4.92797047e-01 -1.52114332e-01
-7.52162099e-01 7.80445218e-01 7.08319172e-02 -2.25440353e-01
1.04844093e+00 -2.55174965e-01 -6.16754353e-01 -1.21039987e-01
-3.80031645e-01 4.32594866e-01 6.50955975e-01 1.62440300e-01
3.69670302e-01 -1.23832953e+00 -5.33891916e-01 4.22117740e-01
5.46698412e-03 1.90863371e-01 3.16460371e-01 8.14177513e-01
-1.64240730e+00 2.97658980e-01 -2.91359067e-01 -6.01405442e-01
-1.38330019e+00 5.97743869e-01 4.26432550e-01 -6.03711903e-01
-9.69218493e-01 1.07798219e+00 3.85093652e-02 -9.18997467e-01
1.30559877e-01 -1.33340642e-01 2.42851540e-01 -5.23450732e-01
1.56042233e-01 6.53600395e-01 5.05730987e-01 -1.35596693e-01
-3.06602985e-01 2.59551406e-01 -2.59600669e-01 6.75380528e-01
1.31949174e+00 5.05660295e-01 -6.90921307e-01 -2.49068454e-01
1.04433429e+00 -1.73202634e-01 -7.45846510e-01 -6.14358857e-02
3.78829092e-02 -1.35804459e-01 2.07904503e-01 -4.93343651e-01
-1.52891839e+00 6.03028595e-01 2.00764075e-01 3.18672597e-01
1.12444401e+00 -8.68194029e-02 9.81108844e-01 3.39162648e-01
4.81444627e-01 -6.58388555e-01 -3.64076316e-01 8.70667219e-01
2.31673628e-01 -9.22019780e-01 7.35395476e-02 -8.64534199e-01
3.12943347e-02 1.24814796e+00 1.07597101e+00 -2.38102540e-01
6.60602331e-01 4.01762068e-01 -1.65821835e-01 -5.26464403e-01
-7.60590553e-01 -1.05663799e-01 3.26545015e-02 6.57940447e-01
4.43227649e-01 1.69613466e-01 -2.22764239e-01 3.92184943e-01
-4.09187645e-01 -8.45620632e-02 1.73352733e-01 7.25092232e-01
6.48547560e-02 -1.21213281e+00 1.78813711e-01 6.95612729e-01
-1.82641327e-01 -3.44195396e-01 -1.69635266e-01 9.72520947e-01
-1.00711495e-01 3.06037098e-01 1.94466561e-01 -6.85478508e-01
2.36903988e-02 -4.52849418e-01 3.06647152e-01 -5.96178651e-01
-1.00707173e+00 -3.16734284e-01 3.09440136e-01 -6.94148779e-01
2.73413777e-01 1.35476246e-01 -1.31098640e+00 -1.19048882e+00
-1.81390047e-01 1.33043677e-01 6.63158000e-01 5.19221425e-01
8.48796606e-01 7.90457129e-01 5.64960614e-02 -8.25319886e-01
1.79203860e-02 -5.97727835e-01 -5.27842820e-01 -1.07834823e-02
-1.54207557e-01 -2.41787300e-01 1.64804924e-02 -6.45222008e-01] | [7.027835845947266, 5.768289566040039] |
d5e56cc7-fc0a-4ee7-99a8-dc9110e99c55 | coevolution-of-camouflage | 2304.11793 | null | https://arxiv.org/abs/2304.11793v2 | https://arxiv.org/pdf/2304.11793v2.pdf | Coevolution of Camouflage | Camouflage in nature seems to arise from competition between predator and prey. To survive, predators must find prey, and prey must avoid being found. This work simulates an abstract model of that adversarial relationship. It looks at crypsis through evolving prey camouflage patterns (as color textures) in competition with evolving predator vision. During their "lifetime" predators learn to better locate camouflaged prey. The environment for this 2D simulation is provided by a set of photographs, typically of natural scenes. This model is based on two evolving populations, one of prey and another of predators. Mutual conflict between these populations can produce both effective prey camouflage and predators skilled at "breaking" camouflage. The result is an open source artificial life model to help study camouflage in nature, and the perceptual phenomenon of camouflage more generally. | ['Craig Reynolds'] | 2023-04-24 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 1.73777919e-02 -3.82464468e-01 6.21480465e-01 6.24078512e-01
6.85719669e-01 -1.02320421e+00 5.44834614e-01 -4.84588534e-01
-6.61697328e-01 8.88921976e-01 -1.01465508e-01 9.84024554e-02
5.19307077e-01 -6.85870886e-01 -6.19938076e-01 -1.10153437e+00
-4.21991944e-01 1.61362402e-02 5.10086298e-01 -4.43531662e-01
2.28649095e-01 5.14407575e-01 -1.83292782e+00 -1.75527304e-01
3.51986080e-01 5.37338853e-01 3.87830526e-01 1.19544721e+00
1.76956326e-01 4.74908859e-01 -1.18669701e+00 -4.84818637e-01
9.19335485e-01 -9.33908582e-01 1.69970527e-01 -5.44903129e-02
6.72535449e-02 -1.06642298e-01 -2.14666530e-01 1.50557494e+00
2.77428031e-01 -3.23384821e-01 7.10201859e-01 -1.34810793e+00
-1.29472136e+00 2.28972077e-01 -4.14468646e-01 4.34656769e-01
2.61662215e-01 1.07229745e+00 2.82100469e-01 -4.75104362e-01
4.48210090e-01 1.45621741e+00 5.94521224e-01 1.25029266e+00
-1.32283199e+00 -7.14336991e-01 -2.38189280e-01 -6.78247884e-02
-1.40234971e+00 -1.15192138e-01 2.76117295e-01 -4.76054162e-01
8.20987761e-01 5.38798332e-01 1.76790631e+00 1.26893699e+00
1.04918408e+00 1.68202788e-01 1.49107969e+00 -1.08957484e-01
5.03687680e-01 1.99360952e-01 -4.27438736e-01 4.27930355e-01
9.43184912e-01 9.91601050e-01 -3.99374217e-01 -5.85255265e-01
9.86690104e-01 5.97953796e-04 -6.68178856e-01 -1.22865468e-01
-6.82103574e-01 6.32252097e-01 4.29897487e-01 -2.04000808e-03
-2.13101342e-01 1.88433707e-01 -1.60505235e-01 8.87019336e-01
1.23593397e-01 8.71989012e-01 -4.31525350e-01 3.64911295e-02
-9.73014086e-02 1.81083918e-01 1.17294168e+00 1.57944337e-01
3.12578887e-01 2.74320930e-01 5.60992479e-01 6.17832601e-01
4.96936023e-01 1.28150666e+00 6.67343915e-01 -8.66901159e-01
-8.15132737e-01 5.92386782e-01 3.23895335e-01 -8.14469814e-01
2.13562220e-01 7.08450079e-02 -3.95738512e-01 1.28221262e+00
3.56199354e-01 -4.49915051e-01 -8.31163466e-01 1.95078599e+00
1.85505420e-01 3.22874188e-01 3.16832215e-01 1.08013821e+00
4.57236707e-01 7.21548736e-01 -1.06567450e-01 -3.05378139e-01
1.17490959e+00 -6.29382312e-01 -4.02375281e-01 1.18227070e-03
1.68520734e-01 -5.58949590e-01 7.77556598e-01 2.99337178e-01
-1.04396331e+00 -8.88099968e-02 -1.10243487e+00 8.60513747e-01
-9.29564357e-01 -7.14584351e-01 2.98139721e-01 7.72425234e-01
-1.46489799e+00 4.18711990e-01 -4.06500340e-01 -5.79989970e-01
2.19967484e-01 3.16256315e-01 -2.91836232e-01 2.81669289e-01
-9.91011739e-01 9.51221406e-01 -1.87771052e-01 -2.51705050e-01
-1.32255471e+00 -5.69385111e-01 -3.93685997e-01 -1.76133409e-01
9.80733633e-02 -9.86215889e-01 1.08972883e+00 -1.86336040e+00
-1.47902799e+00 1.25755548e+00 4.45934474e-01 -5.65979362e-01
5.78562677e-01 -1.19717702e-01 -3.43305498e-01 -7.09564018e-04
-7.29460418e-02 8.11553955e-01 1.12874532e+00 -1.73256898e+00
-5.80466151e-01 -2.01396734e-01 6.93394095e-02 3.50333393e-01
-6.04588203e-02 1.65987477e-01 4.86917496e-01 -9.36131775e-01
-7.38376856e-01 -9.44255233e-01 -3.60456198e-01 9.72817063e-01
9.39316228e-02 3.34140241e-01 1.26827157e+00 -1.19774658e-02
3.88060242e-01 -2.15502381e+00 2.28336021e-01 -4.07090813e-01
4.92231578e-01 6.57218218e-01 -3.67998332e-01 5.03468156e-01
2.33510762e-01 2.62515128e-01 -2.25163057e-01 1.81821316e-01
-3.93543571e-01 6.40416205e-01 -1.02925047e-01 8.48238647e-01
4.81962740e-01 6.25392735e-01 -1.12578952e+00 -7.54938871e-02
-2.61717677e-01 5.67530394e-01 -5.10295451e-01 3.98894489e-01
-3.26857790e-02 5.23801446e-01 -7.34331459e-02 8.44884932e-01
8.63949776e-01 2.40267918e-01 -2.55929351e-01 6.07725084e-01
-6.77050412e-01 -6.97853744e-01 -6.68462098e-01 6.44163370e-01
2.78576687e-02 7.16698885e-01 6.78403318e-01 -1.87189907e-01
4.63036358e-01 -1.73707716e-02 2.18435690e-01 -4.95700210e-01
7.17695594e-01 3.19336772e-01 7.74796903e-01 -2.97451347e-01
-5.21969190e-03 -3.23705792e-01 1.51052445e-01 4.93745178e-01
-9.95242670e-02 -4.02759969e-01 -2.77411044e-02 -2.33078524e-02
1.34961867e+00 -2.60216564e-01 2.39805728e-01 -8.24418724e-01
3.15433770e-01 3.28423440e-01 5.80399215e-01 7.63343871e-01
-6.29581630e-01 1.19126504e-02 1.30177699e-02 -7.00601876e-01
-1.03157723e+00 -1.43639731e+00 1.71189710e-01 5.16723633e-01
8.73517752e-01 1.76599637e-01 -1.00443864e+00 -7.98957720e-02
1.49926871e-01 3.86683732e-01 -1.14991164e+00 -7.93920875e-01
-4.54878360e-01 -6.87775850e-01 5.82049966e-01 -5.50974198e-02
5.90138674e-01 -1.41940665e+00 -1.50977480e+00 -1.57572672e-01
7.90097535e-01 -1.24764666e-01 -5.02639234e-01 1.01919450e-01
-3.06049645e-01 -1.28538084e+00 -1.11851132e+00 -6.02357864e-01
8.36353719e-01 9.19268906e-01 1.02893281e+00 8.29622209e-01
-7.79406071e-01 8.22274327e-01 -6.25200152e-01 -8.01798105e-01
-9.12151992e-01 -1.28610981e+00 7.62702584e-01 -5.30625619e-02
1.66834325e-01 -7.36538172e-01 -8.63704503e-01 4.18204844e-01
-1.10176420e+00 -3.50037307e-01 5.40433347e-01 9.67920959e-01
2.30278984e-01 2.32995480e-01 -1.05055340e-01 -3.32521908e-02
5.14681518e-01 -6.42094910e-01 -8.59504342e-01 1.45854488e-01
1.51491210e-01 -3.88481349e-01 9.15401876e-01 -1.33079863e+00
-7.42772579e-01 -3.74651887e-02 6.89014018e-01 -6.54284239e-01
-1.64175853e-01 -4.82329726e-01 -7.94210006e-03 -8.74256492e-01
8.14382315e-01 2.64001995e-01 1.57803982e-01 -3.57519865e-01
1.39166936e-01 2.01978669e-01 5.19215703e-01 -9.80928838e-02
1.29304218e+00 7.26232767e-01 8.25337321e-02 -1.07836866e+00
1.46844983e-01 2.19775334e-01 -3.06553423e-01 -6.44804299e-01
8.59965265e-01 -5.10885894e-01 -9.23529387e-01 1.06643128e+00
-9.67217684e-01 -6.07832551e-01 -9.27685320e-01 1.97582707e-01
-2.77755558e-01 1.29558429e-01 -6.03320122e-01 -1.17923021e+00
5.09068482e-02 -7.88292170e-01 6.44944787e-01 9.41272736e-01
3.59931916e-01 -8.20407510e-01 7.25482941e-01 -5.32373965e-01
7.74879396e-01 1.98239028e-01 4.84979898e-01 -3.60380083e-01
-6.12116039e-01 4.69279699e-02 -1.44503623e-01 1.06256038e-01
3.13540637e-01 4.51051831e-01 -6.33888602e-01 -3.39593560e-01
4.91959125e-01 -2.35438839e-01 8.10540915e-01 2.96650976e-01
1.01541966e-01 -6.35954738e-01 -4.02179986e-01 4.61308807e-01
1.42245603e+00 8.90783548e-01 3.20764005e-01 1.80335701e-01
2.44533747e-01 8.30784202e-01 3.53755057e-01 4.96069670e-01
-2.21874595e-01 1.55534476e-01 8.49550366e-01 -2.34298557e-01
-3.39055747e-01 4.20910530e-02 6.64580941e-01 4.42749530e-01
-1.36247128e-01 -8.02716911e-01 -4.90151733e-01 4.04258519e-01
-1.30155396e+00 -9.56060767e-01 -1.23750918e-01 2.18265271e+00
4.02798474e-01 -1.18474811e-01 4.41142708e-01 -2.46214703e-01
8.66346896e-01 -6.90512836e-01 -9.64383781e-01 -4.12607878e-01
-8.33258927e-01 1.29576877e-01 9.33324277e-01 4.39153463e-01
-6.53864741e-01 1.11166584e+00 7.29674482e+00 1.80434883e-01
-1.30869615e+00 1.60424281e-02 2.52131730e-01 -1.16647750e-01
-4.09003086e-02 2.49955267e-01 -2.66244352e-01 6.99466407e-01
6.91581190e-01 -3.92495930e-01 7.93321311e-01 5.51056743e-01
1.61547810e-01 -6.20387495e-01 -5.01988411e-01 6.94015801e-01
1.07855871e-01 -8.17963004e-01 1.74954370e-01 4.11640853e-01
5.75059831e-01 -1.59227684e-01 5.49835861e-01 -1.17163390e-01
1.11469138e+00 -8.47102344e-01 7.88748920e-01 5.41778266e-01
5.64518273e-01 -1.22814350e-01 4.52648252e-01 3.81901205e-01
-1.04876781e+00 -5.40072739e-01 -7.17433274e-01 -6.03504002e-01
2.71536380e-01 -3.21022153e-01 -8.47094133e-02 -5.61434865e-01
1.10889924e+00 3.23538333e-01 -7.18828499e-01 1.68800318e+00
1.33210450e-01 1.50948986e-01 -4.40531582e-01 -5.75710654e-01
1.38557568e-01 -2.86415935e-01 1.27335489e+00 7.30650485e-01
4.13393050e-01 5.09730518e-01 4.52292860e-02 1.01023865e+00
3.50740522e-01 -1.15723483e-01 -9.62002218e-01 -3.32630247e-01
3.32241237e-01 8.97404313e-01 -9.59425330e-01 -4.29417878e-01
-1.48433954e-01 1.30574894e+00 -3.76304716e-01 -1.47155467e-02
-7.89950907e-01 1.47222541e-03 1.52863526e+00 1.33751512e-01
1.27943486e-01 -1.68305561e-01 2.56891370e-01 -1.32655406e+00
-9.17859495e-01 -6.88679278e-01 4.00797315e-02 -9.12251770e-01
-1.87449956e+00 7.16857374e-01 -1.18307665e-01 -1.29769242e+00
3.83448780e-01 -9.22297180e-01 -7.48525977e-01 5.81843138e-01
-9.18399870e-01 -9.25375223e-01 -8.51537585e-02 5.44661462e-01
5.72655261e-01 -3.70205194e-01 6.38369739e-01 -2.53832132e-01
-3.66897702e-01 2.45685741e-01 5.25502115e-02 -3.92172962e-01
4.66379493e-01 -1.09932280e+00 5.25527000e-01 1.02267289e+00
-9.66364518e-02 4.12096232e-01 1.09750438e+00 -1.22383273e+00
-1.68266046e+00 -7.07532883e-01 -1.17656559e-01 -6.01504087e-01
9.42309976e-01 -1.85768828e-01 -6.58227444e-01 -1.53877169e-01
4.42608953e-01 -5.47455251e-02 7.04212844e-01 -1.27745891e+00
-2.24012971e-01 5.66863239e-01 -1.70962954e+00 1.30506301e+00
1.30799925e+00 2.22700074e-01 -4.89438891e-01 1.25470355e-01
7.41420329e-01 2.53374904e-01 -1.45664915e-01 -3.79200839e-02
7.20464945e-01 -1.10387111e+00 8.12670887e-01 -6.36214256e-01
-2.95331568e-01 -6.18341565e-01 -1.49457067e-01 -1.51202357e+00
-4.75365877e-01 -8.94835413e-01 3.48323792e-01 6.37476385e-01
9.16643217e-02 -1.15319359e+00 3.47421825e-01 -5.82144409e-03
-3.97961847e-02 1.75346024e-02 -1.00080228e+00 -1.21488976e+00
1.87601149e-01 7.17590332e-01 1.16317771e-01 6.82619452e-01
-3.50016534e-01 -3.39569040e-02 -5.44967055e-01 2.21202359e-01
9.86528516e-01 -2.93345898e-01 5.65949857e-01 -1.62545383e+00
-6.28603041e-01 -8.30105424e-01 -6.25993192e-01 -4.05312210e-01
-2.31832743e-01 -5.54162227e-02 3.37274730e-01 -6.31816685e-01
9.82034504e-02 1.57495271e-02 -8.18366855e-02 2.07789868e-01
1.02421209e-01 7.92867720e-01 4.48630154e-01 4.68971431e-01
-2.70532817e-02 2.02146724e-01 1.35207319e+00 -6.05754461e-03
-2.70453930e-01 -2.82566577e-01 -6.30717576e-01 9.31666911e-01
7.22321093e-01 -7.20363438e-01 -1.81757018e-01 -9.31389928e-02
-1.79038439e-02 -5.25408648e-02 6.78459883e-01 -1.06947041e+00
8.37749429e-03 -4.79540169e-01 2.69545317e-01 3.97565722e-01
7.55242050e-01 -9.90909636e-01 6.57215655e-01 1.54873288e+00
2.13940725e-01 4.81521100e-01 2.89001048e-01 8.98922563e-01
5.35610437e-01 -1.69688970e-01 1.45627153e+00 -6.95781350e-01
-5.25680363e-01 1.04959741e-01 -1.44714236e+00 -9.80160609e-02
1.43948770e+00 -5.65168858e-01 -7.25624442e-01 -2.50027776e-01
-4.30812001e-01 -2.29690298e-01 1.55785239e+00 2.50968099e-01
5.99529982e-01 -8.66352499e-01 -6.11045659e-01 2.97895700e-01
-9.89013463e-02 -1.02660775e+00 -1.53196365e-01 2.43255824e-01
-1.05713606e+00 -3.70138705e-01 -1.14622450e+00 -2.08660170e-01
-1.46751249e+00 9.92219865e-01 6.20095372e-01 4.74292338e-01
-4.06596214e-01 1.48709416e+00 8.12214851e-01 4.98960465e-01
-1.57176003e-01 1.12102166e-01 -1.89264596e-01 -6.39447629e-01
8.53273928e-01 2.50098765e-01 -1.13150656e+00 -1.06211603e+00
-3.25561523e-01 8.56324613e-01 4.00950193e-01 -2.37471491e-01
8.86951864e-01 -1.17022641e-01 -2.56833851e-01 2.97894210e-01
5.15726149e-01 9.12165567e-02 -1.42903221e+00 6.00398839e-01
-4.46882159e-01 -9.60122824e-01 -5.33081830e-01 -9.80361581e-01
-9.50218856e-01 5.54242134e-01 9.80539680e-01 9.52500224e-01
1.16197562e+00 3.83478627e-02 7.92550683e-01 2.33444482e-01
6.68158114e-01 -3.32477123e-01 5.44405878e-01 1.37311041e-01
8.58131230e-01 -7.78492391e-01 -5.79279482e-01 -2.00053930e-01
-3.21553558e-01 5.93514323e-01 8.53067279e-01 -1.03569138e+00
6.77392185e-01 8.43859851e-01 5.31911314e-01 -2.33903140e-01
-1.09032547e+00 -5.17558694e-01 -2.58238077e-01 1.43677676e+00
-4.29468870e-01 5.22932597e-02 -6.18814170e-01 1.88929617e-01
-2.17348501e-01 -7.64504671e-01 7.51275778e-01 9.36106801e-01
-8.54474664e-01 -1.15286851e+00 -8.12786520e-01 -1.80203453e-01
-4.92123365e-01 3.79085280e-02 -1.59146845e+00 7.22369194e-01
6.42906249e-01 9.80093718e-01 1.46157101e-01 -5.03488600e-01
2.08232671e-01 -6.47337377e-01 6.00350022e-01 -4.26149070e-01
-9.21346366e-01 -1.65812653e-02 -5.11293352e-01 -1.80615842e-01
-3.90911788e-01 -5.76426446e-01 -5.54474473e-01 -5.43650925e-01
-3.13069016e-01 2.99496323e-01 3.40512007e-01 2.90785819e-01
2.03690082e-02 -4.76791523e-02 7.76880562e-01 -1.13672853e+00
-2.14115679e-01 -5.02498090e-01 -8.05152357e-01 2.66223133e-01
3.31607699e-01 -9.85213995e-01 -1.00259960e+00 1.59897089e-01] | [8.333562850952148, -0.7586168646812439] |
a8f9c2f9-723d-45d7-9743-252e7384b243 | implicit-bias-of-sgd-in-l-2-regularized | 2305.16038 | null | https://arxiv.org/abs/2305.16038v1 | https://arxiv.org/pdf/2305.16038v1.pdf | Implicit bias of SGD in $L_{2}$-regularized linear DNNs: One-way jumps from high to low rank | The $L_{2}$-regularized loss of Deep Linear Networks (DLNs) with more than one hidden layers has multiple local minima, corresponding to matrices with different ranks. In tasks such as matrix completion, the goal is to converge to the local minimum with the smallest rank that still fits the training data. While rank-underestimating minima can easily be avoided since they do not fit the data, gradient descent might get stuck at rank-overestimating minima. We show that with SGD, there is always a probability to jump from a higher rank minimum to a lower rank one, but the probability of jumping back is zero. More precisely, we define a sequence of sets $B_{1}\subset B_{2}\subset\cdots\subset B_{R}$ so that $B_{r}$ contains all minima of rank $r$ or less (and not more) that are absorbing for small enough ridge parameters $\lambda$ and learning rates $\eta$: SGD has prob. 0 of leaving $B_{r}$, and from any starting point there is a non-zero prob. for SGD to go in $B_{r}$. | ['Arthur Jacot', 'Zihan Wang'] | 2023-05-25 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [-1.49089769e-01 5.06244957e-01 -5.15111051e-02 -3.87356400e-01
-1.07482648e+00 -3.78552079e-01 -1.19100526e-01 1.40080407e-01
-6.29835606e-01 8.91114175e-01 -3.67960222e-02 -3.99744123e-01
-6.49823964e-01 -8.87524545e-01 -9.36568677e-01 -9.80253220e-01
-7.40405500e-01 5.07174850e-01 -1.21932983e-01 -1.97344989e-01
-2.16278613e-01 1.94309250e-01 -1.19127488e+00 1.86642036e-01
6.26236379e-01 1.11877608e+00 -1.57747958e-02 3.25094342e-01
-1.42612532e-01 5.10828614e-01 -3.90456319e-01 -4.33382243e-01
8.58376145e-01 -3.98730844e-01 -6.38761580e-01 -4.46534514e-01
6.91709518e-01 -1.34177163e-01 -2.34327704e-01 1.44808912e+00
2.73247808e-01 3.79211277e-01 5.69742262e-01 -9.16444898e-01
-3.31354469e-01 8.86893511e-01 -9.84200835e-01 -2.17932954e-01
-2.41765812e-01 -3.69244143e-02 1.52370989e+00 -9.60416436e-01
4.11819518e-01 1.09542680e+00 8.37844133e-01 5.82231402e-01
-1.63856828e+00 -9.49987173e-01 3.00571263e-01 -5.45377254e-01
-1.54694843e+00 -2.88963974e-01 6.18037999e-01 -4.67378944e-01
6.41814470e-01 4.00715113e-01 3.88722211e-01 5.39109588e-01
-6.03788421e-02 5.80829084e-01 1.08182728e+00 -2.68588752e-01
2.20472589e-01 -1.54685974e-01 5.19932747e-01 1.08383393e+00
3.73349756e-01 -1.15701079e-01 -7.14593291e-01 -3.44465561e-02
6.99814141e-01 -1.59195870e-01 -9.14042667e-02 -3.52821529e-01
-7.47553706e-01 1.10848641e+00 5.74166477e-01 3.16879302e-01
-2.01898083e-01 4.67210233e-01 1.23619974e-01 4.99955237e-01
1.90567136e-01 6.31042480e-01 -5.85920513e-01 2.57132620e-01
-1.09158170e+00 1.39331505e-01 6.36801779e-01 7.73594677e-01
1.35646725e+00 5.67753613e-02 1.03620581e-01 1.03099465e+00
8.68913010e-02 3.35984528e-01 -5.54435095e-03 -1.12461174e+00
7.98453391e-01 7.48593271e-01 -7.19240233e-02 -1.03347886e+00
-5.78338027e-01 -6.77156687e-01 -1.13024867e+00 5.32291293e-01
7.66768396e-01 -3.17522526e-01 -8.58505905e-01 2.40904331e+00
-2.19936967e-01 -3.18382531e-01 -2.14020222e-01 9.69345152e-01
5.50828040e-01 6.90315664e-01 -1.34430438e-01 -4.06974882e-01
6.51912987e-01 -3.22077662e-01 -1.96191706e-02 -5.30852556e-01
6.82294667e-01 -5.84742129e-01 1.33855438e+00 4.89630491e-01
-1.35089552e+00 -2.25699067e-01 -1.18525827e+00 2.57195570e-02
-3.39362323e-02 8.89406651e-02 6.75423265e-01 4.64814067e-01
-1.19099331e+00 9.42932248e-01 -7.30113328e-01 3.06241691e-01
3.31306428e-01 7.96221077e-01 -5.82013547e-01 -3.34607065e-02
-1.08983350e+00 6.15172327e-01 2.51363277e-01 3.77452224e-01
-9.48909521e-01 -5.70025146e-01 -7.32344329e-01 -1.18284002e-01
3.83626729e-01 -2.82604456e-01 5.42206287e-01 -9.26563382e-01
-7.10190594e-01 1.07201338e+00 -1.50138974e-01 -4.34593737e-01
5.89702308e-01 -2.21380785e-01 -1.94601044e-02 -4.61516649e-01
3.25760841e-01 4.81884211e-01 7.39731133e-01 -1.22655380e+00
-4.14844364e-01 -7.18277693e-01 5.78648932e-02 2.52190650e-01
-4.31276917e-01 -2.19557121e-01 -3.97894382e-01 -1.64660975e-01
9.26125228e-01 -8.52477372e-01 -4.92051959e-01 -9.42093581e-02
-7.76229560e-01 -1.17691293e-01 1.03684753e-01 -4.81320620e-01
1.50914872e+00 -2.01751804e+00 1.53970659e-01 9.07726228e-01
8.40558052e-01 -1.88123761e-03 -1.09299138e-01 1.55948743e-01
-1.84691146e-01 3.29993397e-01 -3.85278672e-01 -4.58646446e-01
-6.49705296e-03 9.90937948e-02 -1.36093214e-01 6.20359302e-01
-2.97906071e-01 2.89310277e-01 -6.36185586e-01 -1.29429504e-01
-1.39354140e-01 1.04847550e-01 -8.35227311e-01 -2.61524469e-01
-2.36654952e-01 -1.54827489e-02 -3.49618077e-01 1.70584440e-01
7.56181777e-01 -3.59536588e-01 2.96108037e-01 4.96334136e-02
-3.42892736e-01 3.31788719e-01 -1.88561034e+00 1.16966176e+00
-1.86824396e-01 3.55788201e-01 7.08396673e-01 -1.20795143e+00
9.82082963e-01 -2.00279534e-01 7.17723727e-01 -4.67731714e-01
-5.78234009e-02 4.05826539e-01 1.27800494e-01 1.58638448e-01
2.11395383e-01 -4.33361471e-01 1.86612271e-02 6.04179502e-01
-1.31705344e-01 3.63065392e-01 1.75334916e-01 1.59909144e-01
1.40732646e+00 -4.03610468e-01 -4.18849885e-01 -5.50023139e-01
4.07799006e-01 -2.87091941e-01 8.71390104e-01 1.02861714e+00
-2.12160330e-02 6.12625718e-01 1.15767491e+00 -3.30104381e-01
-1.13298225e+00 -1.29781616e+00 -2.46358112e-01 1.40903342e+00
-1.13708898e-01 -3.41259390e-01 -6.40029967e-01 -2.66346723e-01
-1.50129246e-02 3.29779774e-01 -7.60076404e-01 -1.83683321e-01
-7.87063062e-01 -1.05773377e+00 4.35403734e-01 1.65569201e-01
5.47733903e-01 -9.72154915e-01 -1.67425990e-01 1.37673333e-01
-1.29415244e-02 -2.31060550e-01 -5.12868822e-01 8.32982898e-01
-1.14163339e+00 -9.56139565e-01 -6.52512729e-01 -8.29329431e-01
1.20659792e+00 -1.94374576e-01 1.09864092e+00 1.47298783e-01
-1.29466608e-01 -3.46947104e-01 2.73794740e-01 4.22555208e-02
-1.43855289e-01 2.34293312e-01 3.11307490e-01 1.95613932e-02
1.22529872e-01 -5.52512765e-01 -9.54475045e-01 2.79726803e-01
-7.10354030e-01 -2.00932980e-01 3.90778095e-01 8.00384939e-01
8.32093179e-01 2.26439282e-01 3.82228613e-01 -8.75962675e-01
6.51007533e-01 -2.84261554e-01 -7.22666204e-01 2.54114438e-02
-6.17522001e-01 4.16087210e-01 7.71623969e-01 -2.29235649e-01
-3.84269565e-01 7.48797692e-03 -1.05199009e-01 -5.00821710e-01
4.55178767e-01 7.25851834e-01 -1.88616902e-01 3.05376589e-01
1.10611570e+00 5.70949800e-02 8.63370001e-02 -7.46753633e-01
3.15753192e-01 1.63454682e-01 2.95528352e-01 -6.60654426e-01
7.20243692e-01 4.19366986e-01 2.01976523e-01 -7.52711236e-01
-9.34004724e-01 -1.63659960e-01 -3.87563437e-01 -1.11039914e-01
4.57916826e-01 -8.03377748e-01 -9.68393981e-01 2.73753971e-01
-5.82560241e-01 -5.40424347e-01 -5.07609963e-01 4.51531440e-01
-3.75377148e-01 1.12170056e-01 -5.21979570e-01 -7.96234012e-01
-2.38908991e-01 -9.81314600e-01 3.64923984e-01 -4.01104949e-02
-3.00767720e-01 -8.57675552e-01 -1.98572837e-02 1.80286542e-01
2.06053793e-01 1.81876287e-01 1.27785063e+00 -6.44222975e-01
-5.85672975e-01 -3.52684766e-01 -1.27342477e-01 7.91184783e-01
-4.52334583e-02 -1.70327246e-01 -5.82335711e-01 -5.77364504e-01
3.19744548e-04 -3.08568627e-01 1.23242426e+00 7.01933742e-01
8.11129689e-01 -8.00566852e-01 -1.95353344e-01 6.96164608e-01
1.36493850e+00 6.73449188e-02 5.26452184e-01 1.71863392e-01
5.92424214e-01 4.78549063e-01 -8.52891654e-02 2.24210322e-01
-9.16031301e-02 2.75686741e-01 4.20615047e-01 -2.13027984e-01
1.45771191e-01 -2.57528901e-01 4.52993959e-01 6.28296018e-01
5.20257838e-02 2.02954307e-01 -9.51054215e-01 5.27574718e-01
-1.66030216e+00 -9.13739562e-01 -1.29949406e-01 2.97523594e+00
1.25108063e+00 7.39181578e-01 1.03062816e-01 -3.91297787e-03
6.74327314e-01 3.13186288e-01 -6.13327980e-01 -2.51527011e-01
-3.31723481e-01 6.01018190e-01 7.29725242e-01 9.31911469e-01
-9.50196266e-01 8.81726146e-01 5.93323946e+00 8.82238150e-01
-1.03772640e+00 -1.78031161e-01 8.20865571e-01 -4.78770435e-01
-5.55234849e-01 1.20336533e-01 -9.08596396e-01 3.66495281e-01
5.63473761e-01 2.32512102e-01 6.01048231e-01 9.20049429e-01
1.50900319e-01 -1.51681855e-01 -1.22317898e+00 7.67932355e-01
-3.20708632e-01 -1.35806596e+00 -1.42580256e-01 3.09453100e-01
8.71086955e-01 1.50488511e-01 3.06234002e-01 2.95223355e-01
8.72574210e-01 -1.41272080e+00 5.79467356e-01 1.44379660e-01
1.00191033e+00 -9.77212965e-01 1.87132910e-01 5.90489030e-01
-1.03317070e+00 -1.12798020e-01 -6.11898661e-01 -9.26626101e-02
-1.73816442e-01 1.07357109e+00 -5.58088720e-01 -8.06166902e-02
7.38384426e-01 2.83947796e-01 -1.51902989e-01 6.45226121e-01
-4.35717851e-02 3.97818238e-01 -6.52190983e-01 -3.43852341e-02
3.02752346e-01 -7.86217093e-01 5.01094401e-01 8.99997056e-01
1.43057793e-01 2.10066184e-01 1.91023156e-01 9.35107172e-01
-4.89560515e-01 2.30817601e-01 -6.24012768e-01 2.01608300e-01
2.17268229e-01 8.73696625e-01 -6.30245984e-01 -2.49360111e-02
-2.37838477e-01 5.54915190e-01 6.40347123e-01 4.36552912e-01
-5.51663756e-01 -5.78714669e-01 1.02121031e+00 4.47628230e-01
-5.12476228e-02 -3.68709981e-01 -7.43198633e-01 -9.98829305e-01
1.30020469e-01 -7.22132564e-01 6.02284789e-01 -3.63992602e-01
-1.21562254e+00 5.84462702e-01 -2.10852772e-01 -8.39331031e-01
3.60179916e-02 -7.22768724e-01 -3.23185116e-01 1.07696891e+00
-7.32884169e-01 -4.25303102e-01 2.90001929e-01 7.78163433e-01
9.05508175e-03 -2.10347116e-01 5.75774312e-01 4.96185333e-01
-5.25930285e-01 8.87389064e-01 3.74911100e-01 5.42480588e-01
2.92483985e-01 -1.12790191e+00 -1.88627660e-01 7.47629046e-01
1.71007097e-01 1.11936748e+00 8.23825538e-01 -6.20404184e-01
-1.02373838e+00 -8.29926252e-01 9.41013575e-01 -2.88379520e-01
6.37960613e-01 -4.32949364e-01 -9.38593566e-01 9.49370801e-01
-4.44554985e-01 -1.42622948e-01 4.68420744e-01 4.88690823e-01
-5.74726760e-01 -4.09930646e-01 -1.08279610e+00 9.52308714e-01
1.02814424e+00 -4.62029845e-01 -6.53325468e-02 3.83072585e-01
1.77476913e-01 -2.31630534e-01 -7.41250873e-01 6.36232555e-01
5.07816553e-01 -1.10865259e+00 1.07182026e+00 -7.80163169e-01
4.46999937e-01 -1.32806242e-01 -3.58987600e-01 -9.24250245e-01
-2.34962553e-01 -9.31127667e-01 1.55801341e-01 6.90350115e-01
9.35181081e-01 -5.60929477e-01 1.29214525e+00 7.95957685e-01
-1.15689941e-01 -8.26842487e-01 -1.38657963e+00 -7.20770121e-01
7.09443688e-01 -4.56235558e-01 -4.81510237e-02 7.66533375e-01
-1.16869815e-01 3.86742234e-01 -4.37232316e-01 -7.24580605e-03
8.91418576e-01 7.87589177e-02 4.74532604e-01 -1.23964989e+00
-4.54751045e-01 -8.50085199e-01 1.38013378e-01 -1.26630926e+00
-1.78771377e-01 -1.10797238e+00 4.86492440e-02 -1.42335463e+00
3.59574199e-01 -1.23285055e+00 -5.71051180e-01 6.66957974e-01
1.40967771e-01 3.08173567e-01 8.33827928e-02 2.67487437e-01
-4.72094059e-01 1.52874529e-01 1.15979528e+00 -2.40990222e-02
-6.43817663e-01 1.89874232e-01 -8.71214032e-01 8.82121384e-01
5.60882449e-01 -6.78994060e-01 -1.62598923e-01 -4.00960773e-01
8.99752140e-01 3.71118397e-01 5.19463420e-02 -7.19696164e-01
1.07090831e-01 -5.98072894e-02 3.75379175e-01 -4.55459118e-01
4.80931759e-01 -3.55205983e-01 1.90296665e-01 4.76230890e-01
-8.44683349e-01 -1.24802366e-01 -6.88591674e-02 2.57671386e-01
2.71438241e-01 -5.64474165e-01 1.16871870e+00 -3.18816423e-01
4.90311943e-02 4.47275579e-01 -1.09599546e-01 3.34841907e-01
5.76581299e-01 -1.69357568e-01 -2.15806775e-02 -5.31998932e-01
-1.11242366e+00 3.01857919e-01 2.58066267e-01 -1.95766632e-02
4.29273546e-01 -1.24886119e+00 -7.27343976e-01 3.65924001e-01
-1.89135328e-01 4.92442220e-01 8.62693787e-02 5.17692208e-01
-4.78570789e-01 1.88324541e-01 1.11740887e-01 -4.64720041e-01
-8.51960421e-01 -6.77897260e-02 7.73597240e-01 -4.86028284e-01
-5.14051318e-01 1.52469695e+00 2.16372758e-01 -5.36977232e-01
5.63729227e-01 -1.71142429e-01 1.78553984e-01 1.32368818e-01
2.30455175e-01 5.28857887e-01 2.37186983e-01 -5.07935286e-01
-2.37942562e-01 4.22376513e-01 -1.69913292e-01 -4.75613654e-01
1.37302220e+00 2.21959665e-01 -5.61107814e-01 3.76753211e-01
1.54256999e+00 2.52073318e-01 -1.49390435e+00 -2.40781650e-01
-1.63946413e-02 -2.46392637e-01 -2.37400293e-01 -6.91100061e-01
-1.36820471e+00 8.88519287e-01 6.07113242e-01 2.92263389e-01
6.92598224e-01 1.62299618e-01 4.90593106e-01 6.29143715e-01
3.74136299e-01 -1.39176357e+00 4.07877475e-01 8.95500600e-01
7.98312366e-01 -1.06345689e+00 9.88781676e-02 3.09650928e-01
-4.13783103e-01 7.71627843e-01 5.64762533e-01 -3.16124350e-01
8.44250143e-01 9.56032309e-04 7.92114064e-03 -3.54013175e-01
-4.09014702e-01 4.01300378e-02 1.87166780e-01 3.35781537e-02
3.42916578e-01 3.14017572e-02 -2.87592083e-01 5.85274875e-01
-6.35953486e-01 -5.24842024e-01 2.14678362e-01 4.94800955e-01
-8.97910476e-01 -1.12652659e+00 -2.59440005e-01 9.73870158e-01
-5.91463447e-01 -3.81337315e-01 -1.25901192e-01 7.66984284e-01
1.14766978e-01 6.99627757e-01 1.40729025e-01 -2.43173301e-01
1.85413480e-01 2.45421544e-01 3.79052430e-01 -6.92676902e-01
-2.97179103e-01 2.12722480e-01 -1.63304061e-01 -3.84677112e-01
3.07851523e-01 -5.90434074e-01 -1.68833566e+00 -5.58948815e-01
-9.78060290e-02 2.65276395e-02 4.63899940e-01 6.45639062e-01
-1.61477432e-01 -5.42016141e-02 7.00609624e-01 -1.60664633e-01
-6.92926407e-01 -6.60519361e-01 -9.10875022e-01 3.83099705e-01
6.48366392e-01 -3.06319475e-01 -7.99700558e-01 -3.80869508e-01] | [7.768314361572266, 3.79764461517334] |
21a3c6a2-51dd-463a-895e-52cd5165ff52 | multi-modal-relational-graph-for-cross-modal | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zeng_Multi-Modal_Relational_Graph_for_Cross-Modal_Video_Moment_Retrieval_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zeng_Multi-Modal_Relational_Graph_for_Cross-Modal_Video_Moment_Retrieval_CVPR_2021_paper.pdf | Multi-Modal Relational Graph for Cross-Modal Video Moment Retrieval | Given an untrimmed video and a query sentence, cross-modal video moment retrieval aims to rank a video moment from pre-segmented video moment candidates that best matches the query sentence. Pioneering work typically learns the representations of the textual and visual content separately and then obtains the interactions or alignments between different modalities. However, the task of cross-modal video moment retrieval is not yet thoroughly addressed as it needs to further identify the fine-grained differences of video moment candidates with high repeatability and similarity. Moveover, the relation among objects in both video and query sentence is intuitive and efficient for understanding semantics but is rarely considered. Toward this end, we contribute a multi-modal relational graph to capture the interactions among objects from the visual and textual content to identify the differences among similar video moment candidates. Specifically, we first introduce a visual relational graph and a textual relational graph to form relation-aware representations via message propagation. Thereafter, a multi-task pre-training is designed to capture domain-specific knowledge about objects and relations, enhancing the structured visual representation after explicitly defined relation. Finally, the graph matching and boundary regression are employed to perform the cross-modal retrieval. We conduct extensive experiments on two datasets about daily activities and cooking activities, demonstrating significant improvements over state-of-the-art solutions. | ['Zheng Qin', 'Zhou Zhao', 'Meng Liu', 'Xiaochi Wei', 'Da Cao', 'Yawen Zeng'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['moment-retrieval'] | ['computer-vision'] | [ 1.65455416e-01 -4.57341939e-01 -5.76530874e-01 -3.64739269e-01
-9.88478780e-01 -6.03122771e-01 7.22388208e-01 5.95634818e-01
-7.54187778e-02 9.39509496e-02 4.53276426e-01 2.70437002e-01
-4.03142571e-01 -4.89546746e-01 -7.39842772e-01 -3.60891074e-01
-3.20454240e-01 1.88175887e-01 3.49344879e-01 -1.27454132e-01
2.27080643e-01 1.13218352e-01 -1.63821948e+00 7.11860359e-01
4.85041380e-01 1.21536529e+00 3.17278892e-01 6.21924400e-01
-2.03807682e-01 1.07760608e+00 -3.02374154e-01 -2.31431261e-01
-8.86632726e-02 -5.82130969e-01 -8.70123148e-01 5.66845477e-01
6.84866726e-01 -2.44972721e-01 -8.00368905e-01 9.90762949e-01
1.99561492e-01 4.94415283e-01 6.49243653e-01 -1.32179630e+00
-7.86181986e-01 6.46750629e-01 -7.47945070e-01 4.18785721e-01
9.89849150e-01 -1.25838071e-01 1.17033923e+00 -6.90259993e-01
9.41580713e-01 1.24701715e+00 9.89909843e-02 8.21004659e-02
-9.02400494e-01 -4.78721142e-01 4.58552301e-01 6.18475199e-01
-1.63531244e+00 -2.63193429e-01 1.13519347e+00 -5.09459615e-01
7.33002782e-01 3.33058268e-01 8.41521561e-01 9.84524071e-01
-1.43998981e-01 1.14757299e+00 4.85250056e-01 -1.12190917e-01
-3.39892626e-01 -1.52235359e-01 1.94322363e-01 8.69268060e-01
-3.93765360e-01 -3.79032165e-01 -8.81322563e-01 9.59422886e-02
6.40623271e-01 4.84101176e-01 -5.00157356e-01 -5.65325677e-01
-1.54328585e+00 5.57188034e-01 4.11494315e-01 5.34364045e-01
-4.56382871e-01 1.17643058e-01 7.58022189e-01 3.09094787e-01
3.47761363e-01 1.04397953e-01 -1.65234223e-01 -2.56339777e-02
-1.03152549e+00 3.99197936e-02 5.39290190e-01 1.18134570e+00
9.20463622e-01 -4.23585445e-01 -5.15743911e-01 8.19932759e-01
4.82670933e-01 4.65096146e-01 4.12170589e-01 -7.21242189e-01
8.69376600e-01 9.78340924e-01 -2.00045183e-01 -1.81876624e+00
-6.92914277e-02 1.51361655e-02 -6.61945820e-01 -7.51281738e-01
2.24552870e-01 4.51512188e-01 -7.43424654e-01 1.50934184e+00
3.34843874e-01 4.82582808e-01 -2.70819943e-02 1.14419758e+00
1.31392610e+00 8.79100263e-01 2.42214352e-01 -3.89806628e-01
1.66610885e+00 -9.95025754e-01 -7.97775388e-01 -2.00430125e-01
5.08393407e-01 -7.63017893e-01 1.01872635e+00 -1.08249307e-01
-1.07283425e+00 -6.73499942e-01 -9.59104836e-01 -3.57092261e-01
-3.82429600e-01 6.49527013e-02 2.53285170e-01 -1.70627475e-01
-5.88532448e-01 2.96588182e-01 -6.02441132e-01 -4.55371886e-01
1.24915957e-01 3.15305777e-02 -5.35921276e-01 -3.07438672e-01
-1.40204668e+00 4.92624342e-01 4.57472235e-01 -1.98373361e-03
-8.97951305e-01 -5.33340275e-01 -1.26008427e+00 -1.26715496e-01
6.99235618e-01 -6.72140896e-01 8.56885016e-01 -1.06304812e+00
-1.08667779e+00 1.12185574e+00 -2.31738582e-01 -2.71834224e-01
2.59545475e-01 -3.08566004e-01 -4.99243587e-01 8.12770844e-01
2.64019489e-01 5.07770181e-01 9.30337727e-01 -1.31707394e+00
-7.09984660e-01 -3.06634873e-01 5.38181782e-01 5.85338533e-01
-3.80829483e-01 9.77653787e-02 -1.39493263e+00 -7.34512150e-01
2.88599908e-01 -7.51115739e-01 2.37218872e-01 -1.98154017e-01
-3.77575070e-01 -3.06191295e-01 9.86157596e-01 -7.48831630e-01
1.42108822e+00 -2.23392248e+00 4.52199459e-01 1.61192775e-01
2.07468778e-01 -4.17111516e-01 -3.44412595e-01 5.04955173e-01
-5.52774407e-04 -1.13152973e-01 1.61374763e-01 -1.38264984e-01
-5.24498411e-02 1.51140407e-01 -1.87373340e-01 5.98365009e-01
-3.38318832e-02 1.00615585e+00 -1.30581081e+00 -1.06016874e+00
2.37144992e-01 5.61667681e-01 -1.45913303e-01 4.19317752e-01
-2.39670783e-01 2.64643222e-01 -7.48027861e-01 8.13657641e-01
1.58359915e-01 -5.75155199e-01 1.91692248e-01 -8.64387870e-01
1.44460708e-01 -6.04406651e-03 -9.45415199e-01 2.20496535e+00
-2.84842908e-01 7.07708478e-01 -2.94987500e-01 -1.20479774e+00
8.42262864e-01 2.12723508e-01 9.37171280e-01 -9.09596443e-01
1.33355707e-01 -3.05088401e-01 -6.62555397e-01 -1.05030441e+00
7.06934750e-01 3.02191168e-01 -3.01243901e-01 3.24610651e-01
2.59399749e-02 2.07995977e-02 4.21643078e-01 5.54389536e-01
7.96136737e-01 3.60054523e-01 2.17871159e-01 1.74291432e-01
7.31065571e-01 1.13482475e-02 2.84986168e-01 4.49683249e-01
-6.87170997e-02 7.82650352e-01 4.03116941e-01 -3.10208738e-01
-5.14785886e-01 -1.01185989e+00 3.82368326e-01 1.47881794e+00
1.02072263e+00 -7.95432210e-01 -2.66735703e-01 -8.03097904e-01
-1.84741884e-01 1.63569793e-01 -7.14865625e-01 -9.23488289e-02
-5.28349757e-01 -1.37398824e-01 2.25342542e-01 3.35241944e-01
4.67530042e-01 -8.64400148e-01 -4.45026129e-01 -2.59763420e-01
-7.38370240e-01 -1.34107769e+00 -9.92840707e-01 -3.13702822e-01
-6.16931736e-01 -1.45370412e+00 -6.97854459e-01 -1.05944800e+00
6.82279766e-01 6.71230674e-01 1.27543437e+00 2.37153739e-01
-1.56664267e-01 9.04625952e-01 -7.15054095e-01 3.32435697e-01
6.71076775e-02 -1.45251513e-01 -2.43636101e-01 2.88599849e-01
2.92558402e-01 -3.13998729e-01 -7.96404898e-01 3.89635444e-01
-1.05061316e+00 1.35055035e-01 4.34072942e-01 5.42269886e-01
1.00218201e+00 -1.02971509e-01 -4.24792953e-02 -6.46047056e-01
3.57476979e-01 -7.25376129e-01 -1.13408864e-01 7.15045750e-01
-2.19221655e-02 -3.92415300e-02 3.01120430e-01 -6.81202531e-01
-7.75714338e-01 1.46589652e-01 5.72537124e-01 -1.08362341e+00
1.87441595e-02 7.74339497e-01 -1.56471550e-01 3.84930760e-01
2.69778132e-01 4.28981870e-01 -3.14939886e-01 -1.98689669e-01
6.14861667e-01 3.64034355e-01 6.94585979e-01 -6.50835752e-01
6.17621779e-01 5.42469025e-01 -1.64256230e-01 -8.28308761e-01
-1.00794387e+00 -9.83027220e-01 -7.34799802e-01 -6.92878962e-01
1.13423777e+00 -9.79352415e-01 -5.90205789e-01 -9.63501856e-02
-1.12774909e+00 -7.75079504e-02 -2.00525522e-02 4.58999336e-01
-6.58034801e-01 7.18995392e-01 -4.88328844e-01 -3.43362331e-01
-1.49366394e-01 -1.06179106e+00 1.40037823e+00 1.80057317e-01
-1.98221102e-01 -9.26005244e-01 -5.23441983e-03 6.00959241e-01
-2.89634645e-01 2.93537110e-01 8.17262948e-01 -6.32022500e-01
-8.10321808e-01 -2.61153907e-01 -4.52284783e-01 -2.34698221e-01
2.65514046e-01 1.58483177e-01 -3.98139477e-01 -1.97701156e-01
-3.20578188e-01 -3.41951072e-01 8.13564003e-01 2.22407758e-01
1.16011357e+00 -2.44403139e-01 -5.48641503e-01 5.46947002e-01
1.06101894e+00 1.05860218e-01 3.54480565e-01 4.16989863e-01
1.03150880e+00 7.56900132e-01 9.39441025e-01 3.93944681e-01
6.87048972e-01 7.46124446e-01 3.15326899e-01 2.46792480e-01
-4.35138717e-02 -4.75220472e-01 3.23016733e-01 1.08563328e+00
-6.93676919e-02 -2.21348554e-01 -7.55640984e-01 6.51182890e-01
-2.27503848e+00 -1.34426117e+00 5.37408069e-02 1.93439817e+00
6.34522736e-01 -1.42883182e-01 1.66969612e-01 -1.70636684e-01
9.48696554e-01 6.24777436e-01 -5.24234056e-01 3.02878737e-01
-8.02307948e-02 -6.36436880e-01 1.85916945e-02 1.83172911e-01
-1.38097262e+00 7.83425033e-01 5.31485844e+00 9.24343646e-01
-9.66876686e-01 2.10944097e-02 4.07420158e-01 -1.70676559e-01
-3.31324160e-01 -1.04396179e-01 -4.11544144e-01 3.21422607e-01
4.45246875e-01 -2.12124377e-01 3.81406307e-01 7.28187621e-01
1.16997562e-01 1.74873192e-02 -1.32909203e+00 1.42190266e+00
5.98497808e-01 -1.30996299e+00 3.31077754e-01 -5.21321774e-01
6.13551855e-01 -2.18190432e-01 -6.99537173e-02 3.43996793e-01
-3.07567626e-01 -6.21734917e-01 8.81297886e-01 9.30925608e-01
6.11426115e-01 -6.31176054e-01 3.80708486e-01 9.51555446e-02
-1.87942266e+00 1.66668609e-01 -1.21649556e-01 3.75926077e-01
1.79823399e-01 1.43411294e-01 -3.46934825e-01 9.18078363e-01
9.36721563e-01 1.44918168e+00 -6.81652844e-01 7.31315553e-01
-3.20661068e-02 -9.87711246e-04 -6.84178174e-02 1.13170713e-01
2.54794002e-01 -2.05219105e-01 4.88862604e-01 1.52164078e+00
1.42189547e-01 1.36355862e-01 6.20640635e-01 4.49720591e-01
-1.30527020e-01 3.15579653e-01 -5.99438727e-01 -1.83486268e-01
3.42515826e-01 1.25812209e+00 -9.00508285e-01 -4.25220191e-01
-6.83275521e-01 1.04608858e+00 3.24795783e-01 5.88614404e-01
-9.81838107e-01 -1.92895502e-01 5.31902969e-01 5.01416437e-02
2.77246535e-01 -2.36247778e-01 2.95639455e-01 -1.39977574e+00
6.94297627e-02 -8.21819305e-01 9.33879018e-01 -1.03466475e+00
-1.48220968e+00 5.46653330e-01 3.33836436e-01 -1.60970283e+00
-3.11393917e-01 -7.90888667e-02 -2.55006045e-01 2.81187505e-01
-1.34301114e+00 -1.36508644e+00 -6.40674531e-01 1.04950631e+00
8.87971461e-01 1.56898290e-01 3.55237305e-01 4.78117734e-01
-4.71575141e-01 4.09625441e-01 -2.35576421e-01 4.53476846e-01
7.42855728e-01 -8.26036572e-01 -3.11857224e-01 7.48818755e-01
5.35300791e-01 6.15886688e-01 5.29489219e-01 -7.93475926e-01
-1.77334285e+00 -1.14199197e+00 6.70124650e-01 -3.03341866e-01
9.08927262e-01 -7.40548447e-02 -9.63068366e-01 5.76396704e-01
2.85486817e-01 2.24766508e-01 5.49171984e-01 -1.02828778e-01
-5.23109317e-01 -1.26486942e-01 -3.93722296e-01 6.42235160e-01
1.22345340e+00 -1.22458375e+00 -8.09311628e-01 4.20803905e-01
7.52383649e-01 -5.94006121e-01 -9.49007750e-01 3.67196470e-01
6.06372774e-01 -7.57331908e-01 1.34928727e+00 -6.28593743e-01
6.91088200e-01 -5.19686759e-01 -3.89148235e-01 -8.32959354e-01
-5.26464991e-02 -5.61293304e-01 -3.46755683e-01 1.48762405e+00
1.50058314e-01 2.97436714e-01 5.22973359e-01 4.00754720e-01
3.16854045e-02 -7.02103853e-01 -4.72704083e-01 -5.20843148e-01
-7.21427619e-01 -6.13808572e-01 3.27404588e-01 1.15551317e+00
2.84573197e-01 4.77879524e-01 -5.19998550e-01 3.08534801e-01
3.87467653e-01 7.91494250e-01 7.46327162e-01 -7.67185807e-01
-7.40813315e-02 -4.24087971e-01 -6.32165790e-01 -1.43677807e+00
4.02047485e-01 -1.01162326e+00 1.23094544e-01 -1.67540812e+00
6.32583022e-01 6.55971915e-02 -4.90099609e-01 1.56247124e-01
-1.51813909e-01 3.71657729e-01 4.06418920e-01 4.85926777e-01
-1.58774757e+00 5.27718306e-01 1.37706804e+00 -6.22849941e-01
-2.28454098e-01 -3.65048259e-01 -3.01688731e-01 6.54436409e-01
2.68595695e-01 -2.84217834e-01 -7.24272311e-01 -5.20538032e-01
3.81378621e-01 5.48028290e-01 4.37501341e-01 -7.55768657e-01
4.48910773e-01 -3.39998722e-01 2.79938847e-01 -9.33617890e-01
3.91081005e-01 -8.55106294e-01 1.26292765e-01 -6.35816436e-03
-7.09637165e-01 2.84217983e-01 -1.39765501e-01 1.02123404e+00
-6.62768662e-01 1.14225768e-01 2.31493562e-01 -6.31935298e-02
-1.14511442e+00 7.17816293e-01 -1.45423785e-01 2.70139873e-01
1.14148772e+00 -2.75140077e-01 -3.02922875e-01 -5.91396868e-01
-8.69354129e-01 5.06768286e-01 4.09711868e-01 8.16309214e-01
9.92123663e-01 -1.58754253e+00 -3.59172225e-01 -1.60051242e-01
5.40801167e-01 -1.17192701e-01 5.81075013e-01 8.78285408e-01
-2.90685922e-01 1.39577031e-01 -6.96887970e-02 -8.77380013e-01
-1.60220540e+00 8.89502466e-01 1.36997504e-02 -8.51094574e-02
-6.81219578e-01 7.14218497e-01 3.22696149e-01 2.03159302e-01
4.15864915e-01 -2.75686771e-01 -5.88074803e-01 5.11543095e-01
5.63338637e-01 2.29651153e-01 -4.37132329e-01 -1.17385745e+00
-4.41688001e-01 1.01905739e+00 1.23764619e-01 1.67824730e-01
1.04401898e+00 -6.48420751e-01 -2.15397596e-01 7.13996112e-01
1.66415608e+00 -2.19222888e-01 -1.13996351e+00 -4.88055855e-01
7.27854446e-02 -5.62294006e-01 -1.42931938e-01 -9.35048833e-02
-1.25011790e+00 7.43806183e-01 2.91469842e-01 2.95163870e-01
1.20144057e+00 4.99282241e-01 8.89719605e-01 4.53150034e-01
4.04880978e-02 -9.30307865e-01 6.59931898e-01 4.29346979e-01
9.38516498e-01 -1.25068665e+00 2.13727534e-01 -4.73319620e-01
-8.55249286e-01 1.15295804e+00 5.54810226e-01 -6.72396496e-02
6.19029284e-01 -2.48460665e-01 -1.75494179e-02 -5.33436000e-01
-5.77685595e-01 -4.94729787e-01 9.85873640e-01 3.88098925e-01
3.34871203e-01 -1.66728914e-01 -7.64037892e-02 4.75532234e-01
2.80945182e-01 -2.96085924e-01 -1.33666232e-01 8.13002110e-01
-2.65254527e-01 -7.02318907e-01 -2.43943974e-01 2.82079041e-01
-3.24113756e-01 7.42809847e-02 -4.47163105e-01 7.04585969e-01
-8.64342675e-02 9.43936348e-01 2.64143527e-01 -6.19599819e-01
3.68158519e-01 -6.81913793e-02 4.23572868e-01 -4.38906312e-01
-4.30936277e-01 2.64090329e-01 -9.90219116e-02 -8.84728253e-01
-9.24327910e-01 -6.09695792e-01 -1.33743501e+00 -1.75138582e-02
-7.71693587e-02 1.26791745e-01 2.63563544e-01 1.09374738e+00
3.11233550e-01 5.38847744e-01 5.61452210e-01 -9.48558986e-01
2.07354978e-01 -5.24840295e-01 -3.84803236e-01 1.08309519e+00
3.37590903e-01 -7.56243527e-01 -1.37667105e-01 4.87851560e-01] | [10.182877540588379, 0.8766908645629883] |
8dbdf16a-ff1f-4d51-815d-4a2fd990096c | neuse-neural-se-3-equivariant-embedding-for | 2303.07308 | null | https://arxiv.org/abs/2303.07308v2 | https://arxiv.org/pdf/2303.07308v2.pdf | NeuSE: Neural SE(3)-Equivariant Embedding for Consistent Spatial Understanding with Objects | We present NeuSE, a novel Neural SE(3)-Equivariant Embedding for objects, and illustrate how it supports object SLAM for consistent spatial understanding with long-term scene changes. NeuSE is a set of latent object embeddings created from partial object observations. It serves as a compact point cloud surrogate for complete object models, encoding full shape information while transforming SE(3)-equivariantly in tandem with the object in the physical world. With NeuSE, relative frame transforms can be directly derived from inferred latent codes. Our proposed SLAM paradigm, using NeuSE for object shape and pose characterization, can operate independently or in conjunction with typical SLAM systems. It directly infers SE(3) camera pose constraints that are compatible with general SLAM pose graph optimization, while also maintaining a lightweight object-centric map that adapts to real-world changes. Our approach is evaluated on synthetic and real-world sequences featuring changed objects and shows improved localization accuracy and change-aware mapping capability, when working either standalone or jointly with a common SLAM pipeline. | ['John J. Leonard', 'Joshua B. Tenenbaum', 'Kurran Singh', 'Yilun Du', 'Jiahui Fu'] | 2023-03-13 | null | null | null | null | ['object-slam'] | ['computer-vision'] | [-8.60901643e-03 -4.63612638e-02 2.39990093e-02 -5.88930726e-01
-4.27142799e-01 -7.87584424e-01 9.74852324e-01 1.17709942e-01
-3.40479612e-01 3.48715752e-01 1.39748991e-01 2.02491462e-01
-3.08956534e-01 -4.88077670e-01 -1.03764737e+00 -4.11667019e-01
-1.93223551e-01 1.04577446e+00 3.50193888e-01 -4.09522951e-02
1.80019855e-01 1.08015168e+00 -1.54169309e+00 -2.56267428e-01
3.77278656e-01 6.70123219e-01 5.64178526e-01 8.55833411e-01
5.74875399e-02 6.03984773e-01 -5.77239506e-02 6.90887868e-02
4.70914066e-01 1.81685060e-01 -4.49445903e-01 1.98731020e-01
1.07762837e+00 -1.10432558e-01 -6.61519647e-01 7.93211639e-01
9.43666250e-02 7.08793327e-02 4.17332977e-01 -1.59921813e+00
-6.96141660e-01 6.59252778e-02 -1.55031038e-02 -1.12819485e-01
4.76446211e-01 2.01379016e-01 9.11080718e-01 -1.20013201e+00
9.61538255e-01 1.32048404e+00 9.80423391e-01 8.87455195e-02
-1.54077542e+00 -3.08721185e-01 7.78639317e-02 3.41598034e-01
-1.70810997e+00 -6.01045072e-01 6.62327588e-01 -4.80388314e-01
1.33623099e+00 3.36467057e-01 8.19947958e-01 9.97804284e-01
5.97729325e-01 2.95460552e-01 6.14772558e-01 -6.19570576e-02
3.81828934e-01 7.40411133e-02 -1.37916282e-01 7.99523711e-01
5.00098705e-01 -8.32129344e-02 -1.08085787e+00 -2.72291183e-01
7.08483338e-01 1.14434697e-01 -2.82088995e-01 -1.54487538e+00
-1.85556173e+00 5.53650796e-01 7.98663259e-01 -2.35698432e-01
-3.99189353e-01 8.82125974e-01 -7.16618299e-02 1.24462843e-01
1.74656913e-01 4.74446893e-01 -4.23970044e-01 -2.59435892e-01
-7.39762664e-01 3.18292141e-01 7.14180291e-01 1.67887700e+00
1.25595915e+00 2.03686982e-01 1.63233921e-01 2.06078723e-01
6.35401070e-01 1.02664423e+00 1.02380171e-01 -1.20575273e+00
1.46326184e-01 5.26217163e-01 2.54930913e-01 -1.30527866e+00
-5.23605824e-01 -3.98667693e-01 -4.04343307e-01 1.92114070e-01
-2.12563574e-01 5.79810739e-01 -7.67511785e-01 1.85211957e+00
3.08629423e-01 5.17313659e-01 -1.33277923e-02 7.84490585e-01
3.49329293e-01 4.57318842e-01 -3.09095472e-01 1.93681955e-01
1.17563748e+00 -7.90230811e-01 -4.90930706e-01 -5.87489605e-01
4.93430138e-01 -5.50894618e-01 7.73150146e-01 1.22235138e-02
-6.77319527e-01 -3.29717636e-01 -1.21655786e+00 -5.20226300e-01
-4.61107224e-01 -8.08264390e-02 7.88969398e-01 2.40387917e-01
-1.48450100e+00 2.89898694e-01 -1.30399966e+00 -8.18304896e-01
-1.95087492e-02 4.71258581e-01 -9.33147013e-01 -3.43324356e-02
-5.46030462e-01 1.31361830e+00 5.10969043e-01 3.26894641e-01
-1.02280474e+00 -7.23006070e-01 -1.40601015e+00 -1.33486524e-01
2.87209034e-01 -9.65604782e-01 8.25794220e-01 -1.54562861e-01
-1.30312562e+00 8.41706932e-01 -1.85066715e-01 -5.92398405e-01
5.08936644e-01 -3.76859218e-01 -8.38907808e-02 -6.63110539e-02
2.31737018e-01 1.03426719e+00 7.30818927e-01 -1.45731997e+00
-3.13723207e-01 -4.37048912e-01 -1.26730921e-02 5.10183811e-01
2.10166588e-01 -4.75827128e-01 -4.86023933e-01 -1.49861723e-01
1.01588309e+00 -1.43297672e+00 -3.21771838e-02 6.68157518e-01
-4.39125337e-02 2.90128201e-01 1.35021317e+00 -4.55718517e-01
3.72309774e-01 -1.98606944e+00 6.06338024e-01 7.30513632e-02
2.36188665e-01 -3.48397374e-01 -2.50672996e-01 5.24437487e-01
6.62388206e-02 -4.30562407e-01 -1.35270059e-01 -7.74832249e-01
4.13662970e-01 7.03255296e-01 -2.97784567e-01 1.09099042e+00
1.17124528e-01 1.30296171e+00 -1.07698488e+00 -1.79471359e-01
5.66598296e-01 5.07878006e-01 -6.34806514e-01 1.97866298e-02
-2.30952427e-01 4.23317850e-01 1.42912865e-01 6.44070864e-01
7.77603924e-01 -5.55074699e-02 -3.46634723e-02 -4.21951324e-01
-2.36177519e-01 2.63505802e-02 -1.33738947e+00 2.52803946e+00
-4.96084124e-01 1.01499724e+00 7.49968514e-02 -4.16646719e-01
9.61159468e-01 -1.99907318e-01 5.39716363e-01 -1.87316954e-01
-4.48132679e-02 1.77768171e-01 -4.74580258e-01 -1.19735301e-01
1.01253879e+00 2.86658019e-01 -2.07391456e-01 1.98926628e-01
5.76587439e-01 -5.54041088e-01 -3.46846074e-01 3.81937742e-01
1.08109319e+00 6.19445920e-01 4.90890294e-01 -5.25834858e-01
2.51989901e-01 -2.11417070e-03 3.03402692e-01 6.27434134e-01
7.65521731e-03 5.43632746e-01 -3.80323827e-02 -5.28219461e-01
-1.28354096e+00 -1.59572899e+00 -7.84151554e-02 5.33208609e-01
6.11701727e-01 -4.34639543e-01 -3.57472599e-02 -2.89114207e-01
4.79211479e-01 6.74715102e-01 -5.46238601e-01 -3.61696571e-01
-5.50116777e-01 -1.10552974e-01 3.12641710e-01 3.05999309e-01
2.50988275e-01 -7.15690851e-01 -9.96276200e-01 2.72447973e-01
-1.05161048e-01 -1.33257246e+00 -3.42106998e-01 3.99125814e-01
-5.44454157e-01 -9.93474424e-01 -4.14948389e-02 -6.01769030e-01
6.01782680e-01 5.62570751e-01 9.18213606e-01 -2.25254163e-01
-3.02315742e-01 1.00697768e+00 -3.01947802e-01 -2.55257338e-01
-3.50088149e-01 -2.00424626e-01 7.17653334e-01 1.16885059e-01
1.84957355e-01 -8.42978120e-01 -1.70451775e-01 2.58958966e-01
-5.78639388e-01 2.98403919e-01 3.29093754e-01 6.09849334e-01
6.90832436e-01 -6.05099201e-01 -7.28909895e-02 -3.31752539e-01
-1.38394013e-01 -2.66396075e-01 -8.66164267e-01 1.84736520e-01
-5.91517150e-01 2.42248192e-01 -8.74359813e-03 -3.19925308e-01
-6.33273900e-01 4.76572335e-01 3.19923848e-01 -7.97979057e-01
-3.82129997e-02 2.60887682e-01 -2.79249579e-01 -6.33685172e-01
6.74501896e-01 4.64873880e-01 1.18152253e-01 -5.13331890e-01
6.33879244e-01 3.20565879e-01 9.71342862e-01 -4.42929238e-01
1.33713591e+00 8.43873739e-01 2.68908978e-01 -8.29779565e-01
-4.05310899e-01 -6.47580147e-01 -1.21313632e+00 -1.72513112e-01
7.57847488e-01 -1.06968188e+00 -7.05087483e-01 3.71600032e-01
-1.30185699e+00 -2.59371132e-01 -4.48135585e-01 7.14671016e-01
-9.73682702e-01 3.74370188e-01 -1.37676358e-01 -5.20694137e-01
2.79741079e-01 -9.87638831e-01 1.56755173e+00 -3.00546855e-01
-2.12662891e-01 -9.30208445e-01 9.89249423e-02 -1.61760822e-01
2.16088474e-01 5.45642376e-01 5.31035066e-01 -2.13075921e-01
-1.38209486e+00 -1.04547389e-01 -8.62477645e-02 -1.35980010e-01
1.75067276e-01 -1.71925217e-01 -7.61195421e-01 -7.65174210e-01
-8.99756793e-03 6.61588535e-02 5.26121736e-01 -1.04219474e-01
4.81811732e-01 -2.32872084e-01 -3.55885029e-01 1.24175549e+00
1.65753257e+00 -3.54554951e-01 2.44951934e-01 5.65330744e-01
1.04264939e+00 2.42510483e-01 5.63170433e-01 4.91805971e-01
7.06008852e-01 1.00592387e+00 1.03940821e+00 2.91819185e-01
-2.00463101e-01 -6.31735742e-01 4.98452008e-01 9.01582718e-01
5.09563029e-01 -8.75049308e-02 -1.11755466e+00 4.97039586e-01
-2.01066947e+00 -4.60948914e-01 -2.03804567e-01 2.00951385e+00
2.13683903e-01 -1.75433323e-01 -6.06150568e-01 -4.46059734e-01
3.70632440e-01 4.33021665e-01 -6.67037189e-01 6.13891818e-02
-3.81189853e-01 -2.32890576e-01 9.56519961e-01 8.64065886e-01
-9.47886705e-01 1.11159194e+00 6.11655664e+00 3.44684981e-02
-1.16579783e+00 3.06757987e-01 -6.92975581e-01 -1.04579993e-01
-5.37266076e-01 5.67609131e-01 -6.64187014e-01 3.72437239e-02
8.17002654e-01 -2.20783919e-01 4.69249010e-01 8.58456075e-01
-7.03127012e-02 2.76116468e-03 -1.66058326e+00 1.31675863e+00
5.45991898e-01 -1.69268489e+00 1.19547524e-01 2.77628988e-01
5.75779498e-01 5.15301406e-01 -9.44896191e-02 2.35780269e-01
5.78671433e-02 -6.15319312e-01 1.50574982e+00 6.18967354e-01
9.00973201e-01 -1.84923396e-01 4.58591074e-01 6.56480134e-01
-1.35978925e+00 1.29621327e-01 -4.01417106e-01 -7.11283684e-02
4.56468999e-01 -2.87065329e-03 -1.18182266e+00 7.13107467e-01
6.66801274e-01 1.05209255e+00 -9.34057415e-01 8.76168728e-01
-4.65122759e-02 -1.56399593e-01 -7.67090142e-01 3.26809376e-01
2.37041667e-01 -2.17268571e-01 1.15719497e+00 1.02582788e+00
4.53408539e-01 -3.41041744e-01 2.38577947e-01 1.11903787e+00
2.07413197e-01 -3.65043491e-01 -8.26211214e-01 2.62526512e-01
6.88188374e-01 1.00864089e+00 -7.19502330e-01 -9.37150121e-02
-6.53180480e-02 1.25046599e+00 3.83867294e-01 2.62925595e-01
-8.91431630e-01 1.62102863e-01 1.01029885e+00 -1.34280458e-01
2.06952751e-01 -1.08197749e+00 -5.90328053e-02 -1.38843787e+00
2.28794649e-01 -2.75015175e-01 -2.18237311e-01 -1.28513646e+00
-7.37312198e-01 3.27870131e-01 2.55887598e-01 -1.42285132e+00
-1.93305358e-01 -6.55271292e-01 3.61589529e-02 6.81740403e-01
-1.45771885e+00 -1.66614127e+00 -7.47597814e-01 4.98950273e-01
4.24100995e-01 -1.06062084e-01 9.72991884e-01 -1.00076776e-02
1.20747969e-01 2.79086292e-01 1.65211573e-01 -3.04511189e-01
5.12899756e-01 -1.26493144e+00 6.38844252e-01 1.04573452e+00
8.45882833e-01 7.68268764e-01 9.63073194e-01 -7.00028539e-01
-2.09099126e+00 -1.28503406e+00 7.83848941e-01 -1.12030578e+00
6.84376478e-01 -1.07028210e+00 -8.05049539e-01 1.23963416e+00
-3.65813017e-01 3.09730530e-01 4.76210266e-02 3.41095254e-02
-5.23336112e-01 -1.03920758e-01 -9.18314338e-01 5.38950086e-01
1.43560064e+00 -9.78201926e-01 -6.52216077e-01 5.27838886e-01
1.11640263e+00 -1.01204324e+00 -7.41517365e-01 4.92041737e-01
5.34181654e-01 -7.72857726e-01 1.19218135e+00 -3.08596969e-01
-4.80486810e-01 -8.20654631e-01 -7.17242241e-01 -1.08220065e+00
-6.22907043e-01 -4.88818735e-01 -4.10911024e-01 8.08328271e-01
-1.72427759e-01 -6.43871427e-01 9.31244195e-01 4.10550892e-01
-4.52624261e-01 -1.52596489e-01 -1.34926462e+00 -1.13727450e+00
-6.50932670e-01 -6.54232860e-01 7.12972939e-01 8.94656360e-01
-7.00188875e-01 -4.06998098e-02 -4.47469413e-01 9.38211679e-01
7.63026774e-01 4.49000373e-02 1.31870520e+00 -1.19158661e+00
-9.94389653e-02 -7.57188573e-02 -1.37747884e+00 -1.19166660e+00
4.99589741e-01 -1.11979377e+00 2.98776984e-01 -1.54344380e+00
-4.48283888e-02 -4.45787758e-01 -3.87105793e-02 5.93290567e-01
4.98443872e-01 2.69212782e-01 4.30051565e-01 4.90725249e-01
-6.47707701e-01 8.97541761e-01 5.34344673e-01 -1.52468681e-01
2.60723550e-02 -6.23185396e-01 -4.36321199e-02 7.28991151e-01
1.61222726e-01 -4.78081852e-01 -3.93843561e-01 -8.47857475e-01
2.74128109e-01 -1.28573209e-01 9.29355800e-01 -1.35687983e+00
5.36886930e-01 -2.28337497e-01 8.89372975e-02 -9.75448072e-01
9.40930247e-01 -1.43716276e+00 8.60330105e-01 3.62290114e-01
-5.61017133e-02 2.63352394e-01 1.47038296e-01 9.60539520e-01
-1.25391275e-01 8.44828114e-02 4.38768923e-01 1.12258039e-01
-1.43353033e+00 5.65509856e-01 1.33228958e-01 -4.80869889e-01
1.09617770e+00 -6.72753990e-01 -1.35758787e-01 -3.16194773e-01
-7.04384744e-01 2.18905300e-01 1.18719423e+00 9.36541259e-01
7.87133992e-01 -1.64903855e+00 -6.33665740e-01 6.45830691e-01
8.19363832e-01 2.78108299e-01 1.65419906e-01 8.20162714e-01
-1.09421813e+00 4.82068330e-01 -3.96602482e-01 -1.38313150e+00
-1.04144263e+00 4.86112028e-01 2.32023850e-01 4.39542562e-01
-9.46082950e-01 8.18063259e-01 1.05859093e-01 -8.99306059e-01
1.45353228e-01 -4.51031059e-01 5.17962098e-01 -2.11087987e-01
7.65064657e-02 1.53536335e-01 7.98255876e-02 -1.23017108e+00
-8.04187834e-01 6.32679224e-01 5.96285760e-01 -2.67592788e-01
1.48862040e+00 -5.93955934e-01 -4.50606525e-01 1.00795031e+00
1.26976204e+00 -6.79600518e-03 -1.49980605e+00 -5.34176826e-01
2.42511675e-01 -7.80803323e-01 -5.27413860e-02 -2.81078577e-01
-3.25204551e-01 6.22790754e-01 7.18669415e-01 -5.60747266e-01
4.48472172e-01 1.26465648e-01 2.56131947e-01 8.31957936e-01
1.24236810e+00 -5.92206538e-01 -2.49809567e-02 7.46206760e-01
1.26688886e+00 -1.09103739e+00 1.88574627e-01 -2.16411889e-01
-3.39964658e-01 1.05181301e+00 4.36333269e-01 -2.05611944e-01
5.14070690e-01 1.64787322e-01 -1.05894431e-01 -2.85382479e-01
-6.71306431e-01 1.36403561e-01 2.76124001e-01 7.69885480e-01
-4.21678722e-01 1.75656304e-01 5.34061849e-01 -2.89315999e-01
-4.14638847e-01 -4.89386022e-01 5.15994966e-01 1.06846797e+00
-5.35711408e-01 -8.30388606e-01 -4.21400785e-01 2.43245326e-02
7.60889471e-01 1.14742860e-01 -1.21774964e-01 8.82203698e-01
2.56099075e-01 2.75693804e-01 2.67902106e-01 -3.36531341e-01
2.78792500e-01 8.92375633e-02 6.52972102e-01 -8.24357331e-01
1.80144921e-01 -2.72823662e-01 -1.49411663e-01 -1.12037408e+00
-3.27352315e-01 -9.70239222e-01 -1.17615020e+00 -1.80991173e-01
-1.53969243e-01 -2.09773362e-01 1.03899539e+00 6.91507161e-01
5.47778010e-01 2.42145658e-01 1.21269159e-01 -1.31156194e+00
-4.33632791e-01 -7.14751363e-01 -6.11530602e-01 4.36538845e-01
6.79209590e-01 -1.10006118e+00 -2.28690699e-01 7.60037005e-02] | [7.396956920623779, -2.331049680709839] |
f1145e32-a01e-4230-afe8-acbf10951099 | automatic-design-of-semantic-similarity | 2307.00925 | null | https://arxiv.org/abs/2307.00925v1 | https://arxiv.org/pdf/2307.00925v1.pdf | Automatic Design of Semantic Similarity Ensembles Using Grammatical Evolution | Semantic similarity measures are widely used in natural language processing to catalyze various computer-related tasks. However, no single semantic similarity measure is the most appropriate for all tasks, and researchers often use ensemble strategies to ensure performance. This research work proposes a method for automatically designing semantic similarity ensembles. In fact, our proposed method uses grammatical evolution, for the first time, to automatically select and aggregate measures from a pool of candidates to create an ensemble that maximizes correlation to human judgment. The method is evaluated on several benchmark datasets and compared to state-of-the-art ensembles, showing that it can significantly improve similarity assessment accuracy and outperform existing methods in some cases. As a result, our research demonstrates the potential of using grammatical evolution to automatically compare text and prove the benefits of using ensembles for semantic similarity tasks. | ['Jorge Martinez-Gil'] | 2023-07-03 | null | null | null | null | ['semantic-textual-similarity', 'semantic-similarity'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.68341833e-01 -3.75507087e-01 2.86228567e-01 -3.86365175e-01
-3.60761464e-01 -3.33216935e-01 6.94126546e-01 6.19301975e-01
-4.65733081e-01 5.70602894e-01 1.14164211e-01 1.37654275e-01
-4.16853637e-01 -1.02411366e+00 1.41470850e-01 -4.21372294e-01
3.84958208e-01 5.01814663e-01 1.98581159e-01 -6.28299594e-01
1.00363457e+00 8.69355872e-02 -2.19529223e+00 1.77562743e-01
1.68487704e+00 7.19190121e-01 3.42296273e-01 1.04845025e-01
-4.76552427e-01 2.89523393e-01 -7.34705806e-01 -5.71351111e-01
6.09509386e-02 -7.53358245e-01 -8.51828218e-01 -3.53194803e-01
2.34726146e-01 4.58858669e-01 2.50497639e-01 1.23548782e+00
7.52740204e-01 5.20138144e-01 7.91021109e-01 -9.86962974e-01
-6.62659228e-01 6.56158268e-01 -1.45321980e-01 2.04210162e-01
6.87261045e-01 -1.60274476e-01 1.12732375e+00 -6.73614860e-01
4.10555810e-01 1.26379943e+00 7.50461280e-01 3.31945151e-01
-9.99444485e-01 -8.14900339e-01 -1.56091571e-01 3.33109677e-01
-1.17181206e+00 -2.18166709e-01 8.54994953e-01 -2.23902002e-01
1.00245643e+00 2.82371759e-01 3.91362071e-01 7.97399998e-01
1.74430773e-01 3.50042135e-01 1.49583995e+00 -6.80321872e-01
3.63195240e-01 1.50336593e-01 3.75057489e-01 6.23579919e-01
3.72126520e-01 -2.47418672e-01 -3.77335489e-01 -3.15446734e-01
6.23661019e-02 -1.26070052e-01 -1.21970586e-01 -1.43016949e-02
-1.07672799e+00 8.86111319e-01 1.61747545e-01 8.92352462e-01
-3.96588087e-01 -4.03124303e-01 5.24909616e-01 2.71068394e-01
5.54782510e-01 1.18026888e+00 2.98754573e-02 -1.72607362e-01
-8.85369956e-01 2.41078556e-01 7.14380622e-01 3.53320122e-01
6.10284925e-01 -4.70801651e-01 -3.29199135e-01 1.19503152e+00
1.31031632e-01 3.29232782e-01 1.03861344e+00 -7.83377945e-01
2.15107083e-01 1.11562335e+00 -1.63032293e-01 -1.20383847e+00
-2.48410210e-01 -4.41950560e-01 -5.45181692e-01 1.98796187e-02
-6.60511758e-03 9.67074111e-02 -4.63018239e-01 1.61875010e+00
2.77874738e-01 2.75653034e-01 1.67749822e-01 7.31784344e-01
8.56447756e-01 3.67650002e-01 3.43708307e-01 -8.98361392e-03
1.32107484e+00 -6.59270883e-01 -5.97862363e-01 -1.29584461e-01
7.31467724e-01 -8.58150899e-01 1.08152127e+00 3.33469510e-01
-7.44610786e-01 -7.74664581e-01 -1.26812816e+00 4.49415147e-01
-5.95390499e-01 -6.51311800e-02 5.88531256e-01 7.75158167e-01
-9.29913998e-01 1.02127492e+00 -2.44688690e-01 -7.63450563e-01
6.62548840e-02 1.94418550e-01 -1.15564667e-01 4.61184569e-02
-1.54566646e+00 1.33122706e+00 8.70711148e-01 -3.41013938e-01
-8.40843320e-02 -2.98176229e-01 -5.82652688e-01 1.59805134e-01
7.22143725e-02 -1.10959375e+00 9.22921836e-01 -1.08908463e+00
-1.35110605e+00 8.73582602e-01 -1.02191381e-01 -3.38177323e-01
1.48258641e-01 -6.41395524e-02 -5.93371630e-01 -5.45681603e-02
1.31111234e-01 3.81124109e-01 4.31437820e-01 -9.90562975e-01
-7.32233047e-01 -4.50553238e-01 -2.13247095e-03 6.37562156e-01
-9.01517153e-01 2.40243256e-01 1.06434479e-01 -6.02452934e-01
7.45531395e-02 -8.34893525e-01 -9.63146761e-02 -6.07002556e-01
2.26404071e-01 -7.49849319e-01 5.54543257e-01 -5.64056575e-01
1.66027498e+00 -1.75918543e+00 1.45636365e-01 3.82484168e-01
6.43158704e-02 6.56639874e-01 -1.73438951e-01 5.18052459e-01
2.19481125e-01 1.65725961e-01 -3.50379288e-01 -1.27141133e-01
6.85563758e-02 6.24686480e-02 1.45994231e-01 -2.84994721e-01
-1.39993712e-01 6.34607375e-01 -1.17957509e+00 -6.75613284e-01
1.88673228e-01 2.65648514e-01 -2.71127313e-01 2.25153416e-01
-5.52453808e-02 1.34263068e-01 -6.36597216e-01 2.09756240e-01
2.83579886e-01 -3.87579128e-02 3.32524478e-01 -2.39739284e-01
1.32690087e-01 3.02695870e-01 -1.19796062e+00 1.68828464e+00
-6.99311554e-01 3.29686165e-01 -7.51593411e-01 -1.18974304e+00
1.45777822e+00 1.35817438e-01 4.37830150e-01 -7.13416219e-01
4.02551442e-01 5.05575478e-01 2.68798709e-01 -5.88419795e-01
6.54438138e-01 2.01518089e-02 -2.13754456e-02 7.74795115e-01
-3.40681821e-02 -3.65603358e-01 6.34261906e-01 7.80196860e-02
9.70265448e-01 -1.91843789e-02 5.86504936e-01 -3.54124606e-01
9.40335214e-01 -9.40715745e-02 3.37381184e-01 7.05689430e-01
-2.59260356e-01 1.49574876e-01 -1.78915054e-01 -2.33566150e-01
-9.51790512e-01 -7.95860767e-01 3.65456268e-02 1.06700420e+00
2.29635313e-01 -5.77448726e-01 -1.13169920e+00 -7.91734278e-01
-4.56679985e-02 1.15075707e+00 -4.62212563e-01 -4.37802166e-01
-2.88682222e-01 -7.12339461e-01 5.37837625e-01 2.21814796e-01
8.55711401e-01 -1.23101294e+00 -6.23118460e-01 4.02528495e-01
-4.79649723e-01 -7.41754174e-01 -2.58002281e-01 -4.54640985e-01
-1.01389515e+00 -9.71123040e-01 -3.05744112e-01 -7.13593304e-01
4.68683034e-01 4.82776582e-01 1.24101925e+00 4.27791655e-01
-1.66932017e-01 2.61732638e-01 -7.88911223e-01 -5.65752506e-01
-7.64806569e-01 2.34471142e-01 3.62306647e-02 3.23775783e-02
7.37832546e-01 -5.49316049e-01 -5.08885443e-01 3.40354413e-01
-8.19091320e-01 -1.78894609e-01 4.28465486e-01 6.93657041e-01
8.76685977e-02 2.87993759e-01 8.36478233e-01 -6.40977502e-01
1.39460874e+00 -4.63183552e-01 -7.06799850e-02 7.05932438e-01
-1.16549706e+00 3.63136709e-01 6.97773218e-01 -1.50669858e-01
-1.06679845e+00 -4.90492314e-01 1.95023388e-01 1.49209157e-01
-1.80309653e-01 5.16966343e-01 1.50670841e-01 -1.99062824e-01
6.96911395e-01 2.07064927e-01 1.92410946e-01 -2.10968792e-01
1.65520862e-01 1.09179544e+00 1.35894418e-01 -7.74860620e-01
2.00845793e-01 1.08269066e-01 -9.92332771e-02 -5.36599636e-01
-9.39281821e-01 -5.60209394e-01 -6.43854260e-01 -3.43039066e-01
5.75306475e-01 -4.49265182e-01 -4.50798690e-01 3.74421567e-01
-1.03951275e+00 5.14118433e-01 9.17631462e-02 5.00045955e-01
-4.00892943e-01 4.52571541e-01 8.92666075e-03 -7.01322019e-01
-9.49189067e-01 -9.48948324e-01 9.23881829e-01 5.81932247e-01
-7.39224553e-01 -1.02375519e+00 2.16878086e-01 6.31390989e-01
7.52766728e-01 1.26556000e-02 8.16292763e-01 -9.44084287e-01
3.86698805e-02 -1.43980056e-01 1.32963136e-01 3.95994395e-01
2.70845205e-01 1.10407338e-01 -6.98949277e-01 -7.28287101e-02
-5.33805899e-02 3.66628207e-02 7.59117901e-01 3.76723297e-02
1.10833383e+00 3.72985490e-02 -3.99264663e-01 1.26323387e-01
1.20941460e+00 4.30693924e-01 5.81101716e-01 5.79269052e-01
3.02758485e-01 7.65002608e-01 8.62176239e-01 3.45411003e-01
4.77971673e-01 7.22415745e-01 -1.81744415e-02 4.21176255e-01
4.06268872e-02 -3.38515043e-02 5.33812605e-02 1.02210307e+00
-3.79503876e-01 -1.89329848e-01 -8.97629321e-01 2.69456536e-01
-1.85064793e+00 -1.16102707e+00 1.66211035e-02 2.28666902e+00
6.76776350e-01 2.35945620e-02 1.22334212e-01 4.64512527e-01
9.27539408e-01 -9.11924243e-02 -1.88575149e-01 -7.38049030e-01
-1.05359338e-01 7.63467014e-01 -1.11215837e-01 2.00010464e-01
-9.44082081e-01 8.88832271e-01 5.97171497e+00 9.09185350e-01
-8.14301789e-01 2.03457084e-02 3.49888265e-01 3.02403629e-01
-4.61005211e-01 6.96708709e-02 -5.16494930e-01 6.20738745e-01
8.57259870e-01 -8.43624830e-01 3.27067822e-01 4.18851465e-01
1.57568619e-01 -1.08435795e-01 -6.64820492e-01 9.07576740e-01
4.36054230e-01 -8.86282265e-01 2.98628986e-01 -2.92491943e-01
8.72308314e-01 -4.09224808e-01 -2.19741449e-01 8.01580548e-02
4.64636922e-01 -9.65726137e-01 2.63173521e-01 7.15468943e-01
1.37767792e-01 -9.29410398e-01 1.01459336e+00 2.66454816e-01
-1.00346637e+00 -1.69855937e-01 -2.35738590e-01 -9.60264876e-02
-7.18402714e-02 6.09100819e-01 -9.36230898e-01 8.23090255e-01
4.95276183e-01 5.70212364e-01 -9.11745310e-01 1.05195618e+00
-2.30900854e-01 3.80158782e-01 -1.47467777e-01 -7.41870284e-01
2.03765258e-01 -4.28594947e-01 5.12333035e-01 1.09782445e+00
8.19942057e-01 3.35135832e-02 5.90251246e-03 7.27532923e-01
2.28187278e-01 8.14955711e-01 -7.49058664e-01 -4.53740880e-02
9.14957523e-01 1.27458084e+00 -8.89549255e-01 -3.63398701e-01
-7.33408853e-02 9.74183083e-01 2.96610147e-01 -1.91900015e-01
-7.52560914e-01 -6.30010664e-01 5.08918524e-01 -2.44325802e-01
-1.51560768e-01 -7.80897588e-02 -6.43655300e-01 -8.65084171e-01
-1.15117252e-01 -1.06430888e+00 6.58090591e-01 -8.10629785e-01
-1.55837536e+00 7.07686543e-01 -9.48189646e-02 -1.21180511e+00
-2.03880593e-01 -3.66444767e-01 -6.90342665e-01 8.46882105e-01
-8.39679658e-01 -7.47100711e-01 -7.60737956e-01 2.69266456e-01
4.41277623e-01 -3.64593685e-01 1.10849535e+00 3.96577679e-02
-2.58381009e-01 4.76452291e-01 1.33069664e-01 -2.31124103e-01
8.19894791e-01 -1.13555324e+00 3.96191478e-01 6.97357297e-01
3.86280537e-01 7.87388384e-01 8.22007835e-01 -6.49339378e-01
-7.19838440e-01 -5.59596896e-01 1.13884032e+00 -3.52855295e-01
4.17773694e-01 1.85158491e-01 -9.72550392e-01 -1.81346729e-01
4.15481329e-01 -6.95042849e-01 8.93070817e-01 1.05325043e-01
-2.80004799e-01 -1.45065561e-01 -1.30020940e+00 6.96244359e-01
1.23198414e+00 -2.88409859e-01 -1.09662747e+00 -1.49768749e-02
2.80514419e-01 -8.02766755e-02 -9.59391117e-01 5.27858675e-01
6.49149001e-01 -1.17887807e+00 7.93538809e-01 -4.46248084e-01
5.36453843e-01 -3.16938460e-01 -1.04006015e-01 -1.93945348e+00
-3.31894189e-01 -2.71638483e-01 2.64354557e-01 1.36127222e+00
3.62882018e-01 -9.79578078e-01 3.24988335e-01 4.60832864e-01
-4.04220745e-02 -5.77214301e-01 -4.69866663e-01 -7.94169664e-01
-5.03210723e-02 -9.13878381e-02 1.02388692e+00 1.17880905e+00
9.99761894e-02 4.41760689e-01 2.32984662e-01 -3.54531229e-01
5.87172687e-01 3.06743175e-01 5.37281454e-01 -1.68533742e+00
-6.38676137e-02 -1.06312954e+00 -6.84853911e-01 -7.90299103e-02
3.85947615e-01 -1.21612477e+00 -1.51710674e-01 -1.63857067e+00
3.15823108e-01 -6.06949985e-01 -6.44605875e-01 1.25147209e-01
-8.55109990e-01 7.46614998e-03 2.85611033e-01 9.40209907e-03
-4.89676148e-01 5.11318564e-01 1.17925608e+00 -8.26246291e-02
-6.50780797e-02 -1.54023275e-01 -1.04517543e+00 5.82694888e-01
1.16348016e+00 -3.69684726e-01 -6.08681500e-01 -8.24668482e-02
8.16919208e-02 -6.41925752e-01 5.29451147e-02 -1.33478749e+00
9.51634422e-02 -2.25584671e-01 1.60011684e-03 -5.21266311e-02
-1.79631040e-01 -5.79577148e-01 2.82556921e-01 3.52193326e-01
-4.53310370e-01 3.42860729e-01 -3.02344076e-02 1.82711735e-01
-3.79711241e-01 -7.40539312e-01 5.16594470e-01 -1.36805892e-01
-9.62231755e-01 -7.26826638e-02 1.17476650e-01 1.28309831e-01
1.13343620e+00 -3.54979277e-01 -2.67319560e-01 -1.84298396e-01
-2.46700481e-01 1.14256211e-01 5.17620504e-01 8.77994597e-01
4.99865502e-01 -1.23569918e+00 -9.43761289e-01 -8.60176310e-02
3.58153164e-01 -6.33783162e-01 8.61736983e-02 4.64684486e-01
-4.01381075e-01 2.85630971e-01 -3.39886665e-01 -4.97214645e-01
-1.61831415e+00 1.71474651e-01 1.70858636e-01 -4.24996555e-01
-2.54142702e-01 5.90333521e-01 -4.60055113e-01 -4.23263371e-01
-2.22259521e-01 2.25215301e-01 -6.91321671e-01 1.11829571e-01
3.30663741e-01 5.67209423e-01 3.54250193e-01 -5.75058579e-01
-2.89377838e-01 7.38477290e-01 1.77459717e-01 -1.12521842e-01
1.22047746e+00 -1.98677946e-02 -3.20986807e-01 3.09040010e-01
9.95136619e-01 -1.94173425e-01 -3.26729715e-01 -2.19385639e-01
4.19775784e-01 -5.15597582e-01 -1.57395497e-01 -8.81689727e-01
-6.26643717e-01 6.60506845e-01 6.54465497e-01 3.18847924e-01
1.32922423e+00 -4.16587919e-01 8.24065685e-01 4.32422310e-01
5.60692370e-01 -1.47083688e+00 5.68899699e-02 6.19105458e-01
6.46573544e-01 -1.20693851e+00 -6.33065728e-03 -2.63063937e-01
-7.74328589e-01 1.20598984e+00 7.11710036e-01 -1.12989349e-02
3.29440713e-01 -9.32804942e-02 -4.64630052e-02 -1.24418780e-01
-5.77778637e-01 -3.76038879e-01 6.46249771e-01 4.44215268e-01
9.32656109e-01 1.80972561e-01 -1.19499397e+00 2.53034860e-01
-3.96489739e-01 2.23465753e-03 -5.19710295e-02 8.11086953e-01
-6.98158383e-01 -1.54843032e+00 -3.31458926e-01 7.03699112e-01
-1.65407658e-01 -1.29459813e-01 -7.04817712e-01 3.89541686e-01
2.76401669e-01 1.22529554e+00 -6.37879521e-02 -6.24405801e-01
3.30742776e-01 3.71781200e-01 6.51438653e-01 -4.99206632e-01
-9.75975156e-01 -6.95774853e-01 2.54811525e-01 -3.39340448e-01
-7.00011671e-01 -6.73664033e-01 -1.14417768e+00 -3.64325255e-01
-4.18492079e-01 5.05864620e-01 7.26024985e-01 1.16866410e+00
5.62357962e-01 4.43940490e-01 7.80691445e-01 -3.29173237e-01
-7.02742755e-01 -1.14915085e+00 -1.13464825e-01 1.13238788e+00
-7.04506457e-01 -1.02640533e+00 -4.24682945e-01 -2.68941969e-01] | [10.266336441040039, 8.860600471496582] |
01e51062-4783-460c-9fb8-fa13f2fd5f1b | active-universal-domain-adaptation | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Ma_Active_Universal_Domain_Adaptation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Ma_Active_Universal_Domain_Adaptation_ICCV_2021_paper.pdf | Active Universal Domain Adaptation | Most unsupervised domain adaptation methods rely on rich prior knowledge about the source-target label set relationship, and they cannot recognize categories beyond the source classes, which limits their applicability in practical scenarios. This paper proposes a new paradigm for unsupervised domain adaptation, termed as Active Universal Domain Adaptation (AUDA), which removes all label set assumptions and aims for not only recognizing target samples from source classes but also inferring those from target-private classes by using active learning to annotate a small budget of target data. For AUDA, it is challenging to jointly adapt the model to the target domain and select informative target samples for annotations under a large domain gap and significant semantic shift. To address the problems, we propose an Active Universal Adaptation Network (AUAN). Specifically, we first introduce Adversarial and Diverse Curriculum Learning (ADCL), which progressively aligns source and target domains to classify whether target samples are from source classes. Then, we propose a Clustering Non-transferable Gradient Embedding (CNTGE) strategy, which utilizes the clues of transferability, diversity, and uncertainty to annotate target informative sample, making it possible to infer labels for target samples of target-private classes. Finally, we propose to jointly train ADCL and CNTGE with target supervision to promote domain adaptation and target-private class recognition. Extensive experiments demonstrate that the proposed AUDA model equipped with ADCL and CNTGE achieves significant results on four popular benchmarks. | ['Changsheng Xu', 'Junyu Gao', 'Xinhong Ma'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['universal-domain-adaptation'] | ['computer-vision'] | [ 4.95484471e-01 1.77848563e-01 -6.83245361e-01 -5.07469893e-01
-9.81056035e-01 -8.15212429e-01 4.33111161e-01 1.08243123e-01
-2.49315485e-01 8.69216502e-01 4.83475588e-02 -3.96917900e-03
-5.46863489e-02 -8.44918907e-01 -6.79952085e-01 -9.04791951e-01
2.51986355e-01 5.68928242e-01 3.73166859e-01 -3.65642644e-02
-2.17778385e-01 3.54465395e-01 -1.11364257e+00 1.92608729e-01
1.22115719e+00 1.05711687e+00 1.05025291e-01 9.65811759e-02
-3.10767382e-01 6.20610118e-01 -5.59404612e-01 -3.13116461e-01
2.09149644e-01 -4.78037804e-01 -7.16000438e-01 1.36107892e-01
1.69401422e-01 -1.97116464e-01 -8.72747302e-02 1.03200734e+00
4.13338512e-01 2.13480130e-01 1.08520353e+00 -1.37788987e+00
-9.07024443e-01 4.79968548e-01 -3.89961004e-01 6.28009439e-02
6.55063689e-02 2.65205652e-01 9.77021217e-01 -9.95312393e-01
4.71158385e-01 1.12503839e+00 5.45129418e-01 9.09407854e-01
-1.35517883e+00 -1.07121778e+00 5.46391904e-01 1.78743109e-01
-1.26509249e+00 -4.62184250e-01 1.16931891e+00 -5.68356574e-01
2.00443283e-01 1.40583470e-01 2.48297825e-01 1.56840241e+00
-3.64987463e-01 1.06280386e+00 1.06706214e+00 -3.69820744e-01
5.38157105e-01 5.47835708e-01 3.13728213e-01 2.56080925e-01
1.76423416e-01 9.71258208e-02 -3.06644469e-01 -2.69722164e-01
5.16029894e-01 1.29722431e-01 -3.23552340e-01 -9.48052406e-01
-1.21194780e+00 9.78673875e-01 4.26597178e-01 -1.57627966e-02
-1.99662134e-01 -4.59831953e-01 4.48717505e-01 3.82836103e-01
5.43576121e-01 1.85822412e-01 -5.19768894e-01 4.30352658e-01
-3.10164362e-01 -1.69154555e-01 5.98643124e-01 1.28867972e+00
1.05268466e+00 5.71097545e-02 -2.33730882e-01 1.02436829e+00
4.01244819e-01 5.59182644e-01 6.93905115e-01 -5.13755143e-01
5.21831155e-01 9.42577600e-01 8.23775530e-02 -7.18107104e-01
1.26684427e-01 -3.55651736e-01 -7.92971492e-01 2.92224377e-01
3.85678321e-01 -1.92315266e-01 -9.61070418e-01 1.89008737e+00
6.45967782e-01 4.10843104e-01 4.46847260e-01 7.44620383e-01
7.65854239e-01 7.41400540e-01 3.66946340e-01 -2.50754535e-01
9.24586594e-01 -8.79673243e-01 -4.36732262e-01 -4.58680063e-01
6.47322595e-01 -4.22973901e-01 1.20001459e+00 3.02669983e-02
-4.14033622e-01 -7.78155565e-01 -1.20581853e+00 1.77100569e-01
-5.00926495e-01 -6.42858213e-03 3.79133493e-01 4.60197955e-01
-3.57491821e-01 1.50038585e-01 -6.45403445e-01 -1.39797717e-01
7.58636177e-01 4.51443762e-01 -2.73856074e-01 -2.68320113e-01
-1.25479627e+00 4.73527431e-01 7.91142344e-01 -3.41833442e-01
-1.26870465e+00 -7.26981580e-01 -1.06792259e+00 7.20768049e-02
4.63925779e-01 -3.04108262e-01 9.96567547e-01 -1.44375193e+00
-1.75902629e+00 7.42278278e-01 1.60049811e-01 -3.47941816e-01
2.89868236e-01 5.48915341e-02 -6.23137832e-01 -3.93499881e-02
7.06944317e-02 6.07680678e-01 9.38298047e-01 -1.34990966e+00
-6.56965852e-01 -4.18786466e-01 -1.81496460e-02 3.67360801e-01
-8.84755313e-01 -3.34642023e-01 -4.03096005e-02 -6.41673386e-01
-4.02522385e-02 -6.66176319e-01 -1.22934669e-01 -1.91856716e-02
-2.81329840e-01 -5.71613491e-01 1.11358750e+00 -4.08084095e-01
9.97941792e-01 -2.38561153e+00 2.41694272e-01 2.29219601e-01
1.03824392e-01 4.20287162e-01 -3.35398704e-01 -1.02502862e-02
-1.00497775e-01 -2.54856437e-01 -5.56058526e-01 -1.38243198e-01
5.22820093e-02 4.77169037e-01 -6.71740890e-01 1.98732749e-01
5.24905026e-01 6.22345626e-01 -1.12722683e+00 -4.77640182e-01
1.14993319e-01 2.19729021e-01 -4.24415946e-01 4.92875695e-01
-4.13399428e-01 7.30652273e-01 -7.87159801e-01 6.97804928e-01
6.73130274e-01 -2.79289871e-01 2.02322558e-01 -1.72241796e-02
4.03026730e-01 1.78043842e-02 -1.20769751e+00 1.49877012e+00
-3.61996710e-01 1.31114289e-01 -5.97095937e-02 -1.50710797e+00
1.38262641e+00 2.19136491e-01 4.13665414e-01 -5.59986949e-01
-2.06164569e-02 3.68153483e-01 -1.40615582e-01 -2.24369973e-01
1.20672444e-03 -9.35338661e-02 -4.16935176e-01 2.08350450e-01
3.33450615e-01 1.82008430e-01 -1.69656575e-01 1.33949056e-01
8.40152442e-01 2.63606757e-01 3.96244586e-01 -9.35361236e-02
8.61059010e-01 1.32240474e-01 1.07132602e+00 3.90416741e-01
-5.53487539e-01 3.75001043e-01 2.46692091e-01 -1.77653760e-01
-8.48417163e-01 -1.41294694e+00 -7.67578781e-02 1.34265792e+00
4.06866491e-01 2.83521861e-01 -4.86978948e-01 -1.44593263e+00
1.11266542e-02 7.36585796e-01 -6.61377788e-01 -6.50211394e-01
-4.95423973e-01 -6.03864849e-01 1.60366550e-01 6.09048069e-01
4.47902083e-01 -9.16715801e-01 7.67874643e-02 1.79481879e-01
-4.80426252e-02 -8.38883996e-01 -6.13527298e-01 3.48945916e-01
-7.50810444e-01 -1.19879019e+00 -8.99153411e-01 -1.13602412e+00
8.28241348e-01 9.00231488e-03 9.12633896e-01 -6.64953649e-01
2.29587704e-01 4.04907495e-01 -5.09402931e-01 -3.71147573e-01
-5.59216380e-01 1.17698856e-01 2.26973295e-01 4.21078086e-01
6.53232038e-01 -5.67089677e-01 -3.67516011e-01 5.95132411e-01
-7.96924770e-01 -1.39533892e-01 7.30836630e-01 9.72896695e-01
7.79131711e-01 -7.71198049e-02 1.12895370e+00 -1.23016036e+00
3.73038828e-01 -9.10712779e-01 -4.58769321e-01 3.98591906e-01
-5.76092243e-01 -2.27365151e-01 1.09593725e+00 -9.49092984e-01
-1.21764815e+00 4.27194417e-01 1.33279994e-01 -6.79973066e-01
-4.01919305e-01 2.51947075e-01 -7.92338192e-01 3.91415805e-02
1.00517225e+00 4.26216334e-01 2.44417861e-02 -3.87677073e-01
2.23010048e-01 8.48618686e-01 4.94327247e-01 -7.03291237e-01
1.01148450e+00 3.17423642e-01 -4.92741346e-01 -3.37961346e-01
-1.04291153e+00 -5.99130869e-01 -8.88640463e-01 1.03784621e-01
6.57519639e-01 -1.17477822e+00 -6.66525364e-02 4.28337872e-01
-7.21276820e-01 -5.33100188e-01 -5.91295600e-01 6.12127304e-01
-5.60815156e-01 2.59615421e-01 -6.26414493e-02 -4.55105573e-01
-1.58576414e-01 -1.23651576e+00 6.53724372e-01 3.67185563e-01
-9.93070081e-02 -1.23341072e+00 -8.97392572e-04 3.04662198e-01
1.89052060e-01 3.64602000e-01 8.72863770e-01 -1.32150412e+00
-4.75121796e-01 -1.57867610e-01 -3.31302173e-02 7.36610830e-01
4.81146365e-01 -5.11568725e-01 -9.55198705e-01 -4.88577813e-01
-1.55259237e-01 -7.26414740e-01 6.91891313e-01 5.74642699e-03
1.19818437e+00 -4.07800078e-01 -4.19378787e-01 5.78485131e-01
1.25697076e+00 3.48812908e-01 2.97040999e-01 3.21124464e-01
8.23830485e-01 4.48144794e-01 9.26284015e-01 2.40778580e-01
2.58158147e-01 5.47838926e-01 2.76739717e-01 8.76500364e-03
-1.93968162e-01 -3.19194287e-01 6.14938259e-01 1.03049970e+00
4.59752202e-01 -9.92505401e-02 -9.45306122e-01 8.41794491e-01
-1.60694325e+00 -5.69582820e-01 3.09745431e-01 2.23533845e+00
1.31394172e+00 2.50080287e-01 1.10138282e-01 2.71598808e-03
9.79487538e-01 -6.25864267e-02 -1.13108623e+00 -5.17947450e-02
-3.21546867e-02 1.24183014e-01 3.79294723e-01 3.35389048e-01
-1.29206085e+00 7.15900362e-01 5.18336821e+00 1.24507833e+00
-1.09337068e+00 2.60118335e-01 6.49863183e-01 3.76389414e-01
-3.38220775e-01 -8.76502469e-02 -9.36587572e-01 7.49188066e-01
7.07636833e-01 -2.47022197e-01 9.48668644e-02 1.29157293e+00
-3.03252310e-01 4.87178564e-01 -1.30055690e+00 6.06255293e-01
1.87078509e-02 -9.71401334e-01 2.21542716e-01 -1.26677111e-01
9.12474215e-01 -3.83525252e-01 1.45026982e-01 8.50888431e-01
6.88682973e-01 -6.61917329e-01 3.40694636e-01 1.95184961e-01
9.65089679e-01 -8.12174499e-01 5.19182563e-01 5.23496091e-01
-9.82722402e-01 -3.62822473e-01 -5.71216226e-01 3.85272264e-01
-4.26341653e-01 3.80445063e-01 -9.60089505e-01 5.35770297e-01
5.24694622e-01 1.01145518e+00 -4.41366494e-01 8.30265999e-01
-1.97310343e-01 9.02304411e-01 -1.78857088e-01 1.65791109e-01
3.87951098e-02 -1.58162981e-01 5.05104303e-01 9.40578997e-01
2.61742115e-01 1.46482885e-01 4.50486034e-01 7.68770456e-01
-1.53115228e-01 1.08007953e-01 -4.85379845e-01 -1.25599131e-01
8.88333499e-01 9.57801342e-01 -4.21607405e-01 -4.54403341e-01
-4.05898482e-01 1.06823969e+00 4.54245806e-01 5.90838730e-01
-7.50014544e-01 -5.74048817e-01 5.85528731e-01 -1.84599221e-01
2.77401447e-01 1.39483675e-01 -2.58244544e-01 -1.30200756e+00
-8.65105614e-02 -8.42309535e-01 8.42185438e-01 -4.17644590e-01
-1.71135283e+00 1.97985277e-01 -8.86255428e-02 -1.68928933e+00
-5.43560460e-02 -5.67081332e-01 -7.08441317e-01 9.57650959e-01
-1.72605896e+00 -1.23631477e+00 -3.22937161e-01 1.05285406e+00
5.44894636e-01 -6.40877664e-01 9.89439428e-01 3.26049685e-01
-6.39001071e-01 1.08450985e+00 6.83407664e-01 5.21973848e-01
1.01270998e+00 -1.19772589e+00 -3.16641666e-02 6.54633999e-01
-2.02119932e-01 4.24531698e-01 2.75165111e-01 -4.24511790e-01
-1.08229876e+00 -1.55313075e+00 4.63191658e-01 -3.85538518e-01
5.60934961e-01 -4.45503384e-01 -1.16837084e+00 9.69089329e-01
-1.24694690e-01 1.99891388e-01 9.20300007e-01 1.09661698e-01
-6.79633617e-01 -4.40099299e-01 -1.32569170e+00 2.84219921e-01
7.75373936e-01 -4.64363068e-01 -8.24637592e-01 4.18638527e-01
9.97564137e-01 -2.44817063e-01 -1.01052165e+00 6.64976299e-01
-2.11755354e-02 -4.91097212e-01 1.19551182e+00 -5.49705446e-01
3.53538692e-01 -2.57216811e-01 -7.15022832e-02 -1.52963114e+00
-4.63484824e-01 6.87087849e-02 -2.96751559e-01 1.64613223e+00
2.72915542e-01 -8.90167892e-01 8.24258029e-01 2.49034330e-01
-4.94673669e-01 -5.55445373e-01 -8.72627497e-01 -9.35380578e-01
3.92823339e-01 4.32186723e-02 6.38489425e-01 1.64613426e+00
-2.54716605e-01 3.84446949e-01 -8.08928087e-02 3.07986856e-01
8.05941761e-01 3.02846014e-01 7.47661293e-01 -1.41901243e+00
-2.04269081e-01 -2.35940352e-01 -3.62805694e-01 -1.09887660e+00
4.40440595e-01 -1.10031319e+00 8.09782520e-02 -1.12204933e+00
1.58973739e-01 -9.89241779e-01 -7.29448080e-01 6.88148320e-01
-3.18042248e-01 1.48403630e-01 -2.35153094e-01 4.91761327e-01
-6.59298241e-01 8.57352734e-01 1.30212915e+00 -5.34463406e-01
-2.50380844e-01 1.09132305e-02 -8.17715287e-01 5.72525263e-01
7.08522916e-01 -3.93461913e-01 -8.07117999e-01 -2.79683292e-01
-4.45178211e-01 -1.67955965e-01 3.19794148e-01 -1.05928242e+00
4.14414182e-02 -4.98054415e-01 6.65992856e-01 -3.01779896e-01
1.70238644e-01 -1.03545165e+00 -3.30530614e-01 2.44472831e-01
-6.20475113e-01 -8.60745192e-01 1.15432844e-01 9.50472414e-01
-3.16499501e-01 -2.12235108e-01 1.11981618e+00 1.18231133e-01
-1.19853318e+00 5.18939555e-01 3.02770622e-02 3.16120386e-01
1.37702477e+00 -2.91114658e-01 -2.82516956e-01 1.09096028e-01
-9.52087164e-01 4.92706418e-01 3.09884757e-01 4.79474843e-01
5.23260951e-01 -1.79268122e+00 -7.52926528e-01 4.39213604e-01
6.37219071e-01 4.74338979e-01 2.88796157e-01 3.52886558e-01
-2.30809627e-03 1.45861179e-01 -2.67237693e-01 -7.94725239e-01
-9.17671978e-01 7.53649056e-01 1.60242677e-01 -1.88136995e-01
-2.95482516e-01 9.70061600e-01 6.50210679e-01 -1.09554529e+00
2.93890357e-01 1.07370555e-01 -4.33695912e-01 2.67779063e-02
3.77963513e-01 1.56724602e-02 -2.53720224e-01 -5.40165782e-01
-1.76522657e-01 3.31991076e-01 -3.23530138e-01 4.62955743e-01
1.07791913e+00 -2.65897155e-01 3.76356542e-01 5.10835886e-01
1.32048559e+00 -1.43130451e-01 -1.48672116e+00 -7.79308259e-01
5.41337505e-02 -2.86601484e-01 -3.44634116e-01 -9.08135712e-01
-8.86627913e-01 9.96737480e-01 8.67529094e-01 -1.15215316e-01
1.26334882e+00 -5.97796440e-02 7.89320886e-01 2.66998142e-01
2.02808768e-01 -1.14151561e+00 3.95603597e-01 4.85696822e-01
5.30632079e-01 -1.48266113e+00 -2.97000319e-01 -4.16440994e-01
-7.17025399e-01 9.16682124e-01 1.07560563e+00 -8.53545666e-02
5.06950319e-01 -2.51079172e-01 8.68797675e-02 3.14234972e-01
-4.70383793e-01 2.29603928e-02 3.23961973e-01 9.43687499e-01
8.08685869e-02 1.89392686e-01 1.18948348e-01 9.23709035e-01
2.44734704e-01 -1.36968851e-01 1.64572239e-01 7.09155023e-01
-5.88031769e-01 -1.25550163e+00 -4.18597221e-01 2.35684648e-01
2.19108704e-02 2.09426105e-01 -4.44577903e-01 6.70855701e-01
3.42682004e-01 6.50698483e-01 7.74906725e-02 -4.57589805e-01
2.63913840e-01 3.47024500e-01 1.66037887e-01 -9.15327013e-01
-9.30817798e-02 -1.06216311e-01 -2.33318001e-01 -8.83673280e-02
-3.41073930e-01 -4.73306626e-01 -1.24212658e+00 1.94188222e-01
-1.90033704e-01 3.88148129e-01 1.11208856e-01 8.55788112e-01
3.13662261e-01 4.75792855e-01 9.13101017e-01 -3.49847853e-01
-8.82676899e-01 -9.11787093e-01 -5.37613392e-01 7.38051891e-01
2.82618523e-01 -8.66560578e-01 -4.65607226e-01 2.55871236e-01] | [10.347002029418945, 3.093291997909546] |
17b532ed-5318-4c90-9fdf-bc31bad56aab | learning-a-general-clause-to-clause | 2208.13549 | null | https://arxiv.org/abs/2208.13549v2 | https://arxiv.org/pdf/2208.13549v2.pdf | Learning a General Clause-to-Clause Relationships for Enhancing Emotion-Cause Pair Extraction | Emotion-cause pair extraction (ECPE) is an emerging task aiming to extract potential pairs of emotions and corresponding causes from documents. Previous approaches have focused on modeling the pair-to-pair relationship and achieved promising results. However, the clause-to-clause relationship, which fundamentally symbolizes the underlying structure of a document, has still been in its research infancy. In this paper, we define a novel clause-to-clause relationship. To learn it applicably, we propose a general clause-level encoding model named EA-GAT comprising E-GAT and Activation Sort. E-GAT is designed to aggregate information from different types of clauses; Activation Sort leverages the individual emotion/cause prediction and the sort-based mapping to propel the clause to a more favorable representation. Since EA-GAT is a clause-level encoding model, it can be broadly integrated with any previous approach. Experimental results show that our approach has a significant advantage over all current approaches on the Chinese and English benchmark corpus, with an average of $2.1\%$ and $1.03\%$. | ['Xiang Li', 'Xinyu Yang', 'Hang Chen'] | 2022-08-29 | null | null | null | null | ['emotion-cause-pair-extraction'] | ['natural-language-processing'] | [ 2.31925547e-01 2.87869334e-01 -4.71511126e-01 -9.13356066e-01
-1.06470346e+00 -5.13305187e-01 4.88638103e-01 4.33631361e-01
9.72292423e-02 6.13225996e-01 3.54098350e-01 3.60177383e-02
-7.25759938e-02 -8.00328553e-01 -3.65381777e-01 -4.93272930e-01
-3.86419833e-01 2.74383724e-01 -3.58228564e-01 -3.16432148e-01
2.20806003e-01 -5.77198938e-02 -1.65571523e+00 8.11854482e-01
8.79764318e-01 1.30700743e+00 -2.18976930e-01 2.04476908e-01
-5.65091789e-01 1.17869818e+00 -5.38889408e-01 -6.82766914e-01
-1.70773208e-01 -5.93740165e-01 -9.57044423e-01 -2.27097079e-01
-2.56311774e-01 1.63696110e-01 3.85605603e-01 9.43357050e-01
3.64528120e-01 -1.73373744e-01 6.56460881e-01 -1.40529573e+00
-4.29903209e-01 9.83157992e-01 -6.21542990e-01 -4.18827720e-02
7.73667336e-01 -5.64787745e-01 1.70248878e+00 -9.11681175e-01
6.21518850e-01 1.14569283e+00 3.93316329e-01 5.12917817e-01
-8.08954298e-01 -8.65186274e-01 5.69994748e-01 2.80643404e-01
-1.25449419e+00 -3.84238958e-02 1.08660972e+00 -1.71219349e-01
1.39405906e+00 4.22328562e-01 6.71203792e-01 9.77952957e-01
1.63135350e-01 1.28853738e+00 1.32651651e+00 -4.21057135e-01
2.53320754e-01 4.60858829e-03 4.22312707e-01 4.17528778e-01
-9.13851559e-02 -2.43062094e-01 -8.40260267e-01 -2.60841310e-01
4.69605159e-03 -3.56871933e-01 -2.84834743e-01 1.15459554e-01
-8.78404617e-01 9.38543439e-01 3.96089464e-01 3.46125931e-01
-4.24692005e-01 -8.75775144e-03 3.96053761e-01 1.40308306e-01
6.08889461e-01 5.33142090e-01 -6.22468710e-01 -4.40960675e-01
-7.76767612e-01 5.49844682e-01 8.51427078e-01 1.06485200e+00
6.85130239e-01 -3.42030942e-01 -2.33593449e-01 8.63008797e-01
6.58483431e-02 6.55534863e-02 3.39018285e-01 -4.38913226e-01
4.30387825e-01 1.11230147e+00 -1.72224849e-01 -1.10415542e+00
-4.46321577e-01 -4.01392341e-01 -7.54184783e-01 -4.40486938e-01
-3.25024635e-01 -2.40440443e-01 -5.05605221e-01 1.86626601e+00
3.63762796e-01 1.44030616e-01 2.38859743e-01 7.60309279e-01
1.01550174e+00 8.77537370e-01 1.05332300e-01 -4.68942732e-01
1.41913402e+00 -7.80482948e-01 -1.06450367e+00 -5.23046315e-01
7.11734176e-01 -6.03361726e-01 9.18894589e-01 5.77802241e-01
-1.05800223e+00 -1.07911257e-02 -9.92954969e-01 -5.53008728e-02
-4.19775128e-01 4.31205221e-02 1.03226531e+00 3.97400409e-01
-7.15015471e-01 2.33439803e-01 -3.62825066e-01 7.29455873e-02
1.83215797e-01 3.31429601e-01 -1.87163204e-01 6.92060068e-02
-1.51155472e+00 5.81372261e-01 6.15891039e-01 -1.24353291e-02
-1.79276645e-01 -6.30898833e-01 -9.04277205e-01 2.69425690e-01
5.22615790e-01 -4.07744408e-01 1.12335122e+00 -8.99686635e-01
-1.19603086e+00 8.32712233e-01 -4.19628739e-01 -2.53524542e-01
-4.48476225e-01 -3.95415097e-01 -6.90734386e-01 -4.28731777e-02
5.52092977e-02 7.39319682e-01 4.67170656e-01 -1.36274219e+00
-8.23581457e-01 -3.08319926e-01 6.56487932e-03 3.80858958e-01
-4.28324163e-01 4.22283113e-01 -5.34282327e-01 -7.79636681e-01
1.45736694e-01 -7.05809474e-01 -7.24854171e-02 -7.23376513e-01
-3.86886597e-01 -6.50341690e-01 5.52901387e-01 -3.56889248e-01
1.84952128e+00 -2.05943227e+00 1.40891939e-01 2.08569571e-01
1.90119162e-01 -1.71083331e-01 -5.55367507e-02 5.55501282e-01
-4.39368010e-01 2.79672533e-01 -2.25112319e-01 -3.17315102e-01
3.30066383e-01 2.08207428e-01 -4.78103280e-01 -1.55625254e-01
8.13747704e-01 9.28023934e-01 -9.19597447e-01 -6.54743373e-01
-3.58440578e-01 2.95351863e-01 -9.18909967e-01 5.08885264e-01
-2.66017288e-01 -2.20569536e-01 -6.05474770e-01 8.46385717e-01
6.72097325e-01 -9.44179967e-02 4.90210533e-01 -1.62149414e-01
-6.74090087e-02 7.31614411e-01 -8.90306056e-01 1.33324909e+00
-2.36690059e-01 3.01644623e-01 2.94635240e-02 -1.01606131e+00
1.35984123e+00 3.08654934e-01 6.08367503e-01 -8.22976828e-01
1.85523123e-01 2.07490832e-01 1.60798291e-03 -4.55057770e-01
6.24659657e-01 -3.80860269e-01 -8.05175781e-01 3.35560828e-01
1.21891499e-04 -1.59897476e-01 3.11655104e-01 3.42581332e-01
1.19621921e+00 6.39360324e-02 5.89698195e-01 -2.87744164e-01
4.06333715e-01 -3.17080915e-02 1.02234173e+00 2.77888507e-01
-1.19634166e-01 3.64398509e-01 1.01955497e+00 -2.57640481e-01
-3.72835666e-01 -7.48452127e-01 -4.80561424e-03 1.06495273e+00
4.69697304e-02 -1.05007493e+00 -6.56469405e-01 -7.24973440e-01
-3.86348009e-01 9.03661489e-01 -6.92070067e-01 -2.27100581e-01
-6.38408303e-01 -9.35094774e-01 4.90310311e-01 5.93745172e-01
3.71093959e-01 -1.04094231e+00 -4.87651885e-01 3.16576213e-01
-7.42634654e-01 -1.10128033e+00 -8.71366709e-02 6.78420484e-01
-5.53496718e-01 -8.92895818e-01 2.84428638e-03 -8.89277756e-01
3.76276970e-01 -1.96468875e-01 1.51694441e+00 -9.95727330e-02
-2.66032182e-02 9.10244584e-02 -8.95530820e-01 -7.57154405e-01
9.27544907e-02 2.29591150e-02 -2.52295792e-01 8.05182382e-02
9.87198293e-01 -5.00395060e-01 -3.65070194e-01 -6.26274049e-02
-7.12036431e-01 9.11040157e-02 7.01266289e-01 7.84731090e-01
7.63235569e-01 2.72653610e-01 6.68075204e-01 -1.09920907e+00
9.89930809e-01 -7.06110358e-01 -8.78207106e-03 2.45352685e-01
-7.83442438e-01 5.54365851e-03 4.51887041e-01 -9.12323371e-02
-1.08193827e+00 2.97456570e-02 -2.62199372e-01 -1.02922678e-01
-2.14135453e-01 1.04110754e+00 -4.95920777e-01 6.93621635e-01
5.12051908e-03 1.28418818e-01 -5.80522239e-01 -1.83392331e-01
3.80416662e-01 7.02206612e-01 6.01729274e-01 -9.02670205e-01
3.18187803e-01 9.21605602e-02 -3.01571578e-01 -1.63665906e-01
-1.06107175e+00 -6.20667219e-01 -2.80239165e-01 -2.45599091e-01
8.31940293e-01 -1.04008293e+00 -6.40913069e-01 7.25650787e-02
-1.26584435e+00 -1.52537692e-02 -4.40106913e-02 2.50445843e-01
-2.17158630e-01 -8.21145698e-02 -6.30842149e-01 -9.17005777e-01
-4.51342672e-01 -8.30590606e-01 1.26918471e+00 1.48868561e-01
-6.98683977e-01 -7.11314321e-01 -5.26746223e-03 1.53817147e-01
1.50731415e-01 5.12231886e-01 1.27020109e+00 -7.44688749e-01
-1.36881903e-01 -1.49822161e-01 -1.48185104e-01 3.40539329e-02
9.46800411e-02 5.76114208e-02 -9.29233372e-01 3.66398245e-02
3.03300887e-01 -6.02338791e-01 6.35305345e-01 -1.46554830e-02
1.40264368e+00 -5.95450342e-01 -2.99353838e-01 3.06147248e-01
1.32924902e+00 5.03298461e-01 6.52391016e-01 2.57804185e-01
3.64889950e-01 9.25638497e-01 1.10870206e+00 7.40667164e-01
8.76801610e-01 5.88921487e-01 5.60006499e-01 -1.09932445e-01
3.27864438e-01 -1.69199586e-01 5.26173830e-01 1.14110970e+00
7.51017407e-02 -4.04770166e-01 -9.06005561e-01 5.27261078e-01
-1.75352478e+00 -1.01075542e+00 -3.00710469e-01 1.45585310e+00
1.21902645e+00 2.24121720e-01 -5.93603924e-02 3.89725983e-01
5.19105077e-01 2.34812304e-01 -1.41074672e-01 -9.20035958e-01
-2.48373941e-01 4.83878225e-01 -3.65967423e-01 2.07461879e-01
-1.03490424e+00 8.78660619e-01 6.08826351e+00 8.69580925e-01
-1.15248799e+00 -2.37862781e-01 8.15614283e-01 -8.65998119e-02
-6.41976595e-01 2.31421646e-02 -8.30262899e-01 2.89562225e-01
7.81139195e-01 -2.57203072e-01 2.48749435e-01 8.18281889e-01
-8.21979120e-02 2.88064051e-02 -1.34667170e+00 1.01367927e+00
2.74775714e-01 -9.59710598e-01 1.25086069e-01 -5.68142161e-02
5.52222610e-01 -4.06712651e-01 1.46738044e-03 6.59035087e-01
1.86043903e-02 -1.06843007e+00 8.31603289e-01 1.62133247e-01
5.98962069e-01 -1.25606167e+00 9.19534743e-01 2.25978523e-01
-1.56295478e+00 -1.26188308e-01 -5.91601320e-02 -3.76852393e-01
1.22450098e-01 8.90987635e-01 -6.39237404e-01 8.35531056e-01
9.67058361e-01 8.46534610e-01 -4.26419020e-01 3.33255917e-01
-4.91288602e-01 7.55701125e-01 2.14558486e-02 -3.77601832e-01
3.67317617e-01 -1.72079280e-01 3.40832502e-01 1.68202198e+00
3.28602225e-01 6.10345364e-01 2.66175885e-02 1.00901198e+00
-1.96764469e-01 3.06390047e-01 -3.16698819e-01 -1.54840380e-01
6.32916152e-01 1.44826984e+00 -6.52850270e-01 -3.82941842e-01
-3.18242967e-01 8.47067833e-01 6.00404501e-01 -2.50873901e-02
-1.03053272e+00 -6.41126752e-01 6.14839554e-01 -3.94808441e-01
3.62843454e-01 2.98556000e-01 -3.66732746e-01 -1.02178133e+00
2.82418549e-01 -1.04104698e+00 6.58882320e-01 -8.12503278e-01
-1.29280198e+00 9.32250977e-01 9.62471738e-02 -1.04745853e+00
-5.08231401e-01 -5.69658101e-01 -8.31237912e-01 4.63266194e-01
-1.55504000e+00 -1.06876111e+00 -1.33705780e-01 4.78466094e-01
3.26019913e-01 1.32821992e-01 1.17100954e+00 1.92913234e-01
-7.12063491e-01 6.62995875e-01 -5.44841111e-01 1.75190598e-01
6.11147583e-01 -1.43689215e+00 -4.00158428e-02 8.66455972e-01
8.16565454e-02 8.52157235e-01 5.70940554e-01 -5.43936491e-01
-1.34559250e+00 -9.78951097e-01 1.59307122e+00 -3.48086029e-01
5.43991089e-01 -6.45300567e-01 -8.57433677e-01 5.29090941e-01
6.64300025e-01 -2.67484933e-01 1.23933482e+00 6.19890928e-01
-5.89208543e-01 -1.53272435e-01 -8.80731583e-01 5.04026532e-01
8.55891764e-01 -4.66700375e-01 -8.19811940e-01 -1.31985769e-01
6.90517128e-01 -2.73214132e-01 -9.28716481e-01 5.96200347e-01
2.91842699e-01 -1.10546112e+00 6.84849620e-01 -5.32121181e-01
1.03550804e+00 -1.13406703e-01 -3.45864475e-01 -1.26723635e+00
-4.85874385e-01 -6.81993365e-01 -2.47667015e-01 1.91470444e+00
5.51858962e-01 -1.93681613e-01 4.71201301e-01 8.06134999e-01
-3.94673109e-01 -1.41657948e+00 -8.14214408e-01 -4.37451124e-01
1.60006344e-01 -7.90037215e-01 1.11077225e+00 1.23390138e+00
6.89610541e-01 6.45599544e-01 -2.40038410e-01 2.46861409e-02
2.95675009e-01 6.81153417e-01 3.75098258e-01 -9.84346628e-01
-2.67597884e-01 -7.23078668e-01 1.53247984e-02 -8.58527541e-01
3.66730213e-01 -1.06426311e+00 2.58247107e-01 -1.48541677e+00
4.21263367e-01 -4.33781087e-01 -4.71442133e-01 8.01345348e-01
-5.77489913e-01 -5.44754416e-02 4.38619852e-02 -2.09195420e-01
-8.07855546e-01 6.34706497e-01 7.65671849e-01 -6.83623701e-02
-8.27114806e-02 -4.45729047e-01 -1.12339044e+00 7.00214267e-01
8.25032413e-01 -5.47420740e-01 -3.72238368e-01 -1.83905110e-01
7.52622604e-01 -1.31348208e-01 3.38054970e-02 -4.85670835e-01
1.86571136e-01 -3.43447149e-01 -5.36133498e-02 -7.63799310e-01
3.11804324e-01 -6.38473094e-01 -1.60844775e-03 7.03239739e-02
-4.44662720e-01 2.43536070e-01 7.34090582e-02 2.55918235e-01
-7.76549339e-01 -8.52193777e-03 2.55467117e-01 1.27710462e-01
-7.50208616e-01 -2.23480873e-02 -2.04243764e-01 1.82714313e-01
8.75116825e-01 -5.24252374e-03 -2.43477270e-01 -3.25545222e-01
-3.13346118e-01 3.33229542e-01 -2.82732360e-02 5.45434296e-01
8.03782344e-01 -1.58067834e+00 -6.55081809e-01 7.09770545e-02
5.17521083e-01 -1.34963496e-03 -1.13564327e-01 6.49693727e-01
2.64029920e-01 4.93001014e-01 8.92148092e-02 -2.01021522e-01
-1.29266191e+00 4.97681558e-01 2.51750778e-02 -6.68373168e-01
-2.70748854e-01 1.07789624e+00 2.04977080e-01 -3.26604337e-01
1.87128425e-01 -4.93955791e-01 -5.08824646e-01 2.90130705e-01
4.73903745e-01 -2.87308320e-02 4.74051349e-02 -6.10214412e-01
-6.74990892e-01 3.32750350e-01 -1.42288640e-01 4.01622318e-02
1.52302551e+00 -1.46815842e-02 -6.19400978e-01 4.91448700e-01
1.16287255e+00 5.64662069e-02 -5.89958787e-01 -4.73221345e-03
3.74001712e-01 -1.10965937e-01 4.28399034e-02 -1.01978731e+00
-1.00515068e+00 6.97528303e-01 5.19191287e-02 2.36940116e-01
1.58547008e+00 2.30240986e-01 7.52069533e-01 1.26262158e-01
4.15710598e-01 -1.21587336e+00 7.34405890e-02 6.42445624e-01
9.69845057e-01 -1.07513261e+00 -1.50791690e-01 -8.98471296e-01
-7.40987122e-01 1.03867412e+00 8.47270131e-01 1.12119995e-01
4.47480351e-01 5.90244591e-01 -6.62449598e-02 -5.21213770e-01
-1.33274686e+00 -2.64612138e-01 4.75555658e-01 2.39835173e-01
1.06503475e+00 2.35974297e-01 -6.68960452e-01 1.54351199e+00
-4.05550867e-01 -1.69854686e-01 1.68370172e-01 1.13080537e+00
-2.27518305e-01 -1.52313125e+00 -1.72098726e-01 5.00813127e-01
-6.72786534e-01 -3.65729034e-01 -9.57811356e-01 6.56812370e-01
4.77477968e-01 1.12046504e+00 5.53877875e-02 -6.70377016e-01
2.69234329e-01 1.73323244e-01 1.68332264e-01 -5.44617534e-01
-9.03543174e-01 2.43322015e-01 4.37713563e-01 -7.79917121e-01
-8.06917548e-01 -7.30254114e-01 -1.61847544e+00 -1.49576524e-02
-2.11293742e-01 4.36357588e-01 3.59610677e-01 9.54642355e-01
4.54542905e-01 6.80873930e-01 6.95853770e-01 -3.62117738e-01
2.87833251e-02 -6.92306817e-01 -5.17497838e-01 4.83933926e-01
-5.97728789e-02 -5.26151836e-01 -1.40164107e-01 -6.63417652e-02] | [12.63239860534668, 6.213926792144775] |
633c7235-3815-44bc-b9a5-85f92d3202e7 | complex-relation-extraction-challenges-and | 2012.04821 | null | https://arxiv.org/abs/2012.04821v1 | https://arxiv.org/pdf/2012.04821v1.pdf | Complex Relation Extraction: Challenges and Opportunities | Relation extraction aims to identify the target relations of entities in texts. Relation extraction is very important for knowledge base construction and text understanding. Traditional binary relation extraction, including supervised, semi-supervised and distant supervised ones, has been extensively studied and significant results are achieved. In recent years, many complex relation extraction tasks, i.e., the variants of simple binary relation extraction, are proposed to meet the complex applications in practice. However, there is no literature to fully investigate and summarize these complex relation extraction works so far. In this paper, we first report the recent progress in traditional simple binary relation extraction. Then we summarize the existing complex relation extraction tasks and present the definition, recent progress, challenges and opportunities for each task. | ['Yanghua Xiao', 'Li Wang', 'Deqing Yang', 'Qiao Cheng', 'Qiaoben Bao', 'Haiyun Jiang'] | 2020-12-09 | null | null | null | null | ['binary-relation-extraction'] | ['natural-language-processing'] | [ 2.34979123e-01 6.16600096e-01 -8.23764622e-01 -2.66334236e-01
-1.38748825e-01 -4.15906608e-01 8.07258427e-01 7.07452655e-01
-2.06521481e-01 1.41522872e+00 -8.03763717e-02 -5.89746773e-01
-3.48977447e-01 -1.13507342e+00 2.71662660e-02 -3.37235838e-01
-2.62677252e-01 8.30693126e-01 3.09932798e-01 -2.21767351e-01
-4.77382503e-02 5.97184598e-01 -1.25244427e+00 2.19337314e-01
8.36460888e-01 8.95778656e-01 -3.91125232e-01 4.63205874e-01
-4.96465057e-01 1.15874910e+00 -7.43128896e-01 -8.74337316e-01
-3.29694569e-01 -3.97973567e-01 -1.49689543e+00 -2.25989565e-01
-5.21315277e-01 2.04779148e-01 -3.98596197e-01 9.08591330e-01
2.28209049e-01 -1.52972266e-01 8.71105909e-01 -1.40706968e+00
-3.96973252e-01 1.34448075e+00 -5.01337171e-01 4.77055341e-01
6.40934408e-01 -7.57971168e-01 1.24545908e+00 -6.57710612e-01
7.03482926e-01 9.17066097e-01 5.24550200e-01 6.30423501e-02
-6.88753426e-01 -7.99023867e-01 2.70798393e-02 5.54851055e-01
-1.71519351e+00 -3.93304199e-01 5.28491557e-01 -2.82022506e-01
1.51740336e+00 4.96075273e-01 5.46608865e-01 4.60096300e-01
-3.36337253e-03 6.67001247e-01 1.07628500e+00 -7.83284605e-01
-3.27403009e-01 2.86034673e-01 7.52911687e-01 6.58264995e-01
8.33119154e-01 -1.41681775e-01 -5.78761995e-01 -1.53096303e-01
4.34917271e-01 -3.70645732e-01 -3.05403262e-01 4.82163467e-02
-9.64087963e-01 6.27150059e-01 8.96006301e-02 7.98286915e-01
-3.81500944e-02 -4.91094589e-01 4.33088034e-01 2.72662461e-01
6.30592942e-01 4.48472053e-01 -9.76409912e-01 6.91919029e-03
-6.42638206e-01 9.28554758e-02 1.51459503e+00 1.50445795e+00
5.07354558e-01 -5.11034489e-01 -1.32413074e-01 5.33578575e-01
2.68254489e-01 7.81343207e-02 2.46668816e-01 1.32111146e-03
7.98661113e-01 1.00793171e+00 -2.88800895e-01 -1.11034799e+00
-7.74920881e-01 -2.50034511e-01 -1.00626850e+00 -4.85431999e-01
9.54910293e-02 -3.55587721e-01 -5.79189420e-01 8.58232558e-01
4.63257760e-01 5.63295335e-02 5.03204644e-01 1.21327020e-01
1.87050474e+00 2.84524977e-01 2.26544514e-01 -9.25624907e-01
1.67860651e+00 -9.94069397e-01 -1.53999925e+00 -1.58175260e-01
8.18917871e-01 -9.89911556e-01 8.64690766e-02 1.16630569e-01
-7.95663655e-01 -1.18772335e-01 -1.20714951e+00 -1.06521167e-01
-9.18169320e-01 2.10451987e-02 1.41739368e+00 7.29498446e-01
-2.72177458e-01 5.22334039e-01 -6.63547873e-01 -3.94545078e-01
6.93491101e-01 7.60248363e-01 -6.37266099e-01 1.95135340e-01
-1.75524986e+00 1.32209194e+00 9.54692006e-01 8.45135152e-02
1.58330798e-01 -1.75133765e-01 -1.03626776e+00 -1.63677007e-01
9.54036295e-01 -6.17564857e-01 1.12315154e+00 4.51328866e-02
-1.27949524e+00 9.03788686e-01 -2.81885684e-01 -5.39265215e-01
-1.00225866e-01 -3.83573681e-01 -7.95513272e-01 -1.21245973e-01
4.31076847e-02 -2.13543564e-01 9.07493383e-02 -9.13534284e-01
-9.43920732e-01 -3.09021622e-01 6.08698130e-02 2.95645565e-01
-2.30915740e-01 6.25370085e-01 -3.86790633e-01 -5.62038302e-01
1.68034226e-01 -5.84989130e-01 -2.47564122e-01 -6.57033563e-01
-8.96463096e-01 -8.29696536e-01 8.47795844e-01 -3.46412152e-01
1.79965568e+00 -1.53415430e+00 -1.84967831e-01 1.75731763e-01
6.01050019e-01 4.94006604e-01 6.44894958e-01 5.95928192e-01
-4.84004855e-01 3.37739348e-01 -1.42616808e-01 -1.64174456e-02
-3.18735212e-01 3.36495638e-01 -6.81972727e-02 1.28412694e-01
4.31565434e-01 1.28650987e+00 -1.01142001e+00 -1.18571115e+00
5.21287546e-02 1.26569435e-01 3.98953855e-01 3.49191070e-01
7.37930909e-02 9.22135189e-02 -6.62187040e-01 8.22024882e-01
4.65466648e-01 -3.21067303e-01 6.02856874e-01 -3.92001480e-01
-1.11940809e-01 6.74055040e-01 -1.12180567e+00 9.75699306e-01
-4.24124338e-02 6.04822934e-01 -4.18329835e-01 -1.24690723e+00
9.77458000e-01 7.41899312e-01 5.17515123e-01 -1.89675465e-01
3.69421035e-01 2.85644352e-01 2.29967758e-01 -7.88262725e-01
4.76310372e-01 -2.59719223e-01 -3.13620269e-02 1.63659289e-01
2.65643001e-01 -2.71158487e-01 6.08577311e-01 1.70550555e-01
1.14544368e+00 1.33221507e-01 1.48010600e+00 1.57653794e-01
7.10620105e-01 2.14982107e-01 4.75904673e-01 2.67544299e-01
-2.44685989e-02 1.34493425e-01 6.89331651e-01 -2.60914296e-01
-4.75423068e-01 -6.78085625e-01 -4.37884480e-01 5.48835158e-01
2.74123579e-01 -1.02952421e+00 -2.77605683e-01 -1.16478431e+00
-1.59394354e-01 1.91687629e-01 -3.81015390e-01 7.96793103e-02
-6.16298676e-01 -1.05580688e+00 7.84834027e-01 4.35894132e-01
7.34689415e-01 -1.18084073e+00 -9.52775124e-03 2.96016484e-01
-4.34875846e-01 -1.66711783e+00 4.19027984e-01 6.75845802e-01
-7.62058020e-01 -1.32647121e+00 -7.85152316e-02 -1.11897957e+00
2.97835320e-01 5.93768395e-02 1.41118300e+00 9.63128433e-02
-9.94696468e-02 -5.04471362e-01 -8.32709610e-01 -6.07726932e-01
-1.02336206e-01 7.06127524e-01 -5.43906875e-02 -6.25529945e-01
8.71683240e-01 -4.96693105e-01 4.83181588e-02 2.13320598e-01
-5.65273106e-01 3.16121317e-02 9.17144179e-01 6.70728564e-01
4.58468944e-01 8.06903780e-01 4.19535279e-01 -1.70867753e+00
8.37331474e-01 -5.17166257e-01 2.31475551e-02 6.48124456e-01
-8.44547927e-01 7.56788999e-02 3.30112278e-01 -2.27120385e-01
-1.06903160e+00 -1.38888266e-02 -1.54965401e-01 5.43164074e-01
-4.02654558e-01 1.10729277e+00 -3.61654639e-01 -8.24089944e-02
6.39847994e-01 -2.28724480e-01 -6.09333456e-01 -2.61953741e-01
4.23576355e-01 9.80358124e-01 2.43841365e-01 -4.21149313e-01
8.73106122e-01 1.27891883e-01 2.39062563e-01 -6.48364127e-01
-1.31827700e+00 -7.23226309e-01 -1.23968601e+00 2.41907001e-01
7.49099910e-01 -5.78718543e-01 -6.13177240e-01 2.33211026e-01
-1.37729084e+00 6.10007979e-02 -1.95152208e-01 4.67894822e-01
-1.41092911e-01 3.85192096e-01 -7.34280825e-01 -9.40040410e-01
-5.58597505e-01 -5.93671858e-01 6.32606328e-01 5.74544907e-01
-5.81009686e-01 -1.04730535e+00 1.53200448e-01 2.83141792e-01
-1.12991840e-01 3.60074252e-01 7.38209546e-01 -1.01063943e+00
-1.78920761e-01 -3.85663301e-01 -4.85074401e-01 -2.41812572e-01
7.19546199e-01 3.19882645e-03 -7.07001805e-01 3.11362505e-01
-1.78132460e-01 -2.31147945e-01 5.38504481e-01 -1.08298749e-01
6.02997839e-01 -2.82881558e-01 -1.16001010e+00 2.94430584e-01
9.74411130e-01 5.62920630e-01 7.20701456e-01 3.25200826e-01
7.88625717e-01 7.74320126e-01 1.15789580e+00 6.80002198e-02
6.22488320e-01 5.48119426e-01 -3.00272256e-01 -2.44099677e-01
-5.35608418e-02 3.75721194e-02 -3.50604624e-01 1.01790845e+00
-7.06238031e-01 -3.24603647e-01 -1.09503150e+00 1.85023814e-01
-1.89093637e+00 -7.73047805e-01 -7.15793133e-01 1.58698833e+00
1.44334173e+00 6.11946285e-01 -1.40949696e-01 7.94715524e-01
6.71550333e-01 1.60677743e-03 -4.73193405e-03 -2.58651018e-01
-3.78984660e-01 8.67401838e-01 4.77452904e-01 3.93836200e-01
-1.67144549e+00 1.36227000e+00 6.71451044e+00 8.50421309e-01
-5.25805235e-01 -1.73051581e-02 3.67830813e-01 5.92643797e-01
1.46753833e-01 3.71282965e-01 -1.15806806e+00 -2.91316032e-01
8.14419091e-01 -2.95014888e-01 -3.06004792e-01 6.03249371e-01
-3.19053292e-01 -2.15354607e-01 -1.14616275e+00 1.01511300e+00
-2.04962511e-02 -1.17921746e+00 -8.56318995e-02 -2.66684722e-02
5.40762067e-01 -5.64590156e-01 -4.41261858e-01 3.50012034e-01
3.62712443e-01 -1.29495275e+00 -5.53986356e-02 3.41250628e-01
7.38856554e-01 -8.00695837e-01 1.34977782e+00 2.62870908e-01
-1.76639771e+00 2.82100171e-01 8.49984493e-03 -5.40382981e-01
4.27831173e-01 1.13172984e+00 -8.15981507e-01 1.37820947e+00
5.94192207e-01 1.00077188e+00 -5.28913617e-01 7.80758440e-01
-6.97126508e-01 6.17556274e-01 -2.80600160e-01 -2.67658740e-01
-2.52573133e-01 -1.07768312e-01 2.04710513e-01 1.30752647e+00
-3.99361223e-01 7.08849251e-01 9.21805799e-02 1.52670562e-01
1.00367025e-01 4.33141798e-01 -6.94997966e-01 -3.13529521e-01
4.93688881e-01 1.33308911e+00 -1.04370904e+00 -4.87755179e-01
-6.53845191e-01 5.97648978e-01 5.78675568e-01 4.40261476e-02
-6.76836491e-01 -8.62666786e-01 2.26315811e-01 -1.49087727e-01
7.27205351e-02 -4.24645871e-01 -5.73300660e-01 -1.21028399e+00
-5.69313914e-02 -6.30378783e-01 7.73146749e-01 -2.00542524e-01
-1.23888505e+00 9.85618830e-01 3.26668262e-01 -9.54436243e-01
-1.44181386e-01 -4.87919182e-01 -3.80115122e-01 7.37927914e-01
-1.52739191e+00 -1.17670441e+00 -2.64364153e-01 3.37935328e-01
1.23505130e-01 -3.85084003e-02 1.09090459e+00 5.14026582e-01
-9.64363158e-01 6.77116036e-01 -5.36281586e-01 5.74414790e-01
6.44574821e-01 -1.27834916e+00 3.68543446e-01 5.62926292e-01
4.09400672e-01 8.08213890e-01 5.62470615e-01 -9.27654445e-01
-7.92993784e-01 -7.86446333e-01 1.67608750e+00 -4.35590208e-01
9.55868900e-01 -1.81918085e-01 -8.07862878e-01 7.13633537e-01
2.86144972e-01 1.00640349e-01 1.11888003e+00 5.16033590e-01
-2.34818265e-01 1.00324182e-02 -9.44774210e-01 5.02230465e-01
1.39590931e+00 -2.75384933e-01 -8.52776170e-01 3.96504641e-01
7.65077770e-01 -4.70964849e-01 -1.36630952e+00 1.06734908e+00
4.52593446e-01 -4.85026985e-01 1.04220271e+00 -6.42211676e-01
3.11516583e-01 -3.08340669e-01 3.54612857e-01 -7.09433079e-01
-2.05463812e-01 -6.39950812e-01 -8.82980645e-01 1.70219743e+00
8.78431797e-01 -5.89048147e-01 9.43322659e-01 4.12283182e-01
3.02324921e-01 -1.22220612e+00 -6.43085837e-01 -5.38537502e-01
-1.64489850e-01 -1.93940133e-01 5.59496045e-01 1.35879958e+00
6.52977645e-01 1.40249741e+00 -1.57198712e-01 1.52094932e-02
4.19335872e-01 4.71565247e-01 6.05256140e-01 -1.55201912e+00
-1.17455617e-01 -5.18688262e-01 -6.08897507e-01 -1.02682686e+00
1.98672906e-01 -6.91708922e-01 -2.64565319e-01 -1.91274667e+00
2.67212421e-01 -7.18635380e-01 1.65240109e-01 6.69053912e-01
-6.08736455e-01 1.56196102e-01 -3.99273276e-01 3.57533157e-01
-6.26078188e-01 1.73492610e-01 1.17444849e+00 -2.76307136e-01
-3.47821653e-01 3.80042434e-01 -9.55869079e-01 7.50349998e-01
9.29711282e-01 -5.28060317e-01 -3.95619661e-01 2.52390593e-01
4.19287980e-01 -8.13329369e-02 -6.00938737e-01 -5.73231101e-01
3.76356065e-01 -3.29912841e-01 2.32659474e-01 -9.76385593e-01
-9.60257370e-03 -7.22617030e-01 4.05412763e-02 2.20509529e-01
-1.60380045e-03 -1.31667450e-01 -1.74989298e-01 1.99085280e-01
-6.89462066e-01 -4.13020134e-01 2.57542998e-01 1.06939003e-02
-5.89080393e-01 4.06321734e-01 -1.71708927e-01 8.17414373e-02
1.33886051e+00 -3.02949511e-02 -2.61357903e-01 1.12715221e-04
-8.97233009e-01 2.49833450e-01 -4.54034209e-01 2.45147035e-01
4.47024196e-01 -1.15892470e+00 -6.89767003e-01 -3.61082882e-01
2.43349731e-01 6.34491622e-01 -4.74266738e-01 7.55398333e-01
-4.48066443e-01 7.14980125e-01 2.62081325e-01 3.59799862e-02
-1.67836964e+00 8.36479247e-01 6.73856214e-03 -1.09557498e+00
-4.25010949e-01 8.53682101e-01 -2.65543401e-01 -1.69782266e-01
8.73056725e-02 -3.19285601e-01 -1.24914944e+00 2.17246190e-01
4.27173376e-01 1.66797310e-01 2.76608258e-01 -6.52910709e-01
-6.68379068e-01 4.60976422e-01 -1.55447751e-01 2.21997842e-01
1.09595346e+00 -1.37410283e-01 -7.26204097e-01 5.66660523e-01
8.76809895e-01 5.69526777e-02 4.91424948e-02 -5.05007267e-01
7.44058430e-01 -1.13214038e-01 -3.78748477e-01 -5.07590413e-01
-8.88836682e-01 3.53236586e-01 -3.16392034e-01 8.17774177e-01
1.02463436e+00 5.63723922e-01 6.43518090e-01 5.82747161e-01
5.06908655e-01 -7.63181746e-01 -5.31759977e-01 8.78028631e-01
5.04384696e-01 -1.24660766e+00 7.76577055e-01 -1.68104613e+00
-3.07546616e-01 1.07862651e+00 6.74361467e-01 3.37883204e-01
1.24849033e+00 9.08377826e-01 -9.04567540e-02 -3.50495577e-01
-4.88033593e-01 -7.85934627e-01 4.28649753e-01 7.88078010e-01
1.09253013e+00 1.80036291e-01 -1.02556360e+00 8.94260347e-01
-5.70612252e-01 -2.54169762e-01 -7.41052181e-02 1.22818160e+00
-1.53188974e-01 -1.67755806e+00 -1.55676469e-01 8.35007727e-01
-6.35567427e-01 -4.69255209e-01 -8.89448941e-01 8.70298088e-01
3.87361735e-01 1.32163239e+00 -2.85821378e-01 -5.75005770e-01
5.50417602e-01 -1.02089562e-01 5.77185214e-01 -1.01515508e+00
-4.92616624e-01 -3.97360414e-01 8.82790804e-01 5.12466542e-02
-1.01766503e+00 -4.39008802e-01 -1.47770607e+00 -2.50057459e-01
-1.22028756e+00 6.03201389e-01 4.48862500e-02 1.39107609e+00
-2.00614020e-01 7.20608532e-01 3.24277610e-01 -2.07447812e-01
3.26710671e-01 -1.21188641e+00 -5.86591184e-01 1.25130326e-01
-1.22694187e-01 -1.01120722e+00 1.71381142e-02 9.30920690e-02] | [9.153907775878906, 8.709939956665039] |
2283d8a5-4180-4890-b72a-f21ad2497089 | modeling-4d-fmri-data-via-spatio-temporal | 1805.12564 | null | http://arxiv.org/abs/1805.12564v3 | http://arxiv.org/pdf/1805.12564v3.pdf | Modeling 4D fMRI Data via Spatio-Temporal Convolutional Neural Networks (ST-CNN) | Simultaneous modeling of the spatio-temporal variation patterns of brain
functional network from 4D fMRI data has been an important yet challenging
problem for the field of cognitive neuroscience and medical image analysis.
Inspired by the recent success in applying deep learning for functional brain
decoding and encoding, in this work we propose a spatio-temporal convolutional
neural network (ST-CNN)to jointly learn the spatial and temporal patterns of
targeted network from the training data and perform automatic, pin-pointing
functional network identification. The proposed ST-CNN is evaluated by the task
of identifying the Default Mode Network (DMN) from fMRI data. Results show that
while the framework is only trained on one fMRI dataset,it has the sufficient
generalizability to identify the DMN from different populations of data as well
as different cognitive tasks. Further investigation into the results show that
the superior performance of ST-CNN is driven by the jointly-learning scheme,
which capture the intrinsic relationship between the spatial and temporal
characteristic of DMN and ensures the accurate identification. | ['Wei zhang', 'Yu Zhao', 'Mo Zhang', 'Tianming Liu', 'Shijie Zhao', 'Quanzheng Li', 'Milad Makkie', 'Xiang Li'] | 2018-05-31 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [ 2.27836475e-01 -3.39688361e-01 1.24526015e-02 -4.83315349e-01
-1.12285160e-01 -3.44296128e-01 5.42018294e-01 -2.36170873e-01
-5.10631144e-01 2.84280658e-01 1.80591360e-01 -5.37427031e-02
-6.58135235e-01 -2.87390292e-01 -4.57586259e-01 -7.97165155e-01
-4.72135007e-01 2.99657941e-01 2.25207508e-01 2.26623878e-01
2.46583164e-01 7.99972177e-01 -1.23904228e+00 4.58101422e-01
6.81068838e-01 1.24155974e+00 4.87557471e-01 1.19979367e-01
1.04666561e-01 6.02914453e-01 -2.43183300e-01 1.49354771e-01
2.01976568e-01 -4.20297742e-01 -8.50961566e-01 -1.15690738e-01
1.05032414e-01 -1.72070801e-01 -5.64178228e-01 1.02195799e+00
8.20653439e-01 1.48510337e-01 6.85078025e-01 -7.68117249e-01
-3.70269030e-01 4.28806484e-01 -2.21431300e-01 9.40557241e-01
-1.87502310e-01 3.65358353e-01 8.31179500e-01 -6.66022599e-01
5.36152184e-01 1.15373743e+00 4.74154681e-01 4.44538772e-01
-1.24181104e+00 -7.43304372e-01 5.22223860e-02 2.18672037e-01
-1.39667189e+00 -4.32089239e-01 6.11804664e-01 -8.52318048e-01
8.49867582e-01 -2.60480732e-01 8.82818580e-01 1.30863380e+00
4.90386456e-01 5.68211794e-01 1.25158525e+00 1.78412031e-02
1.32911101e-01 -4.83360320e-01 2.14305446e-01 5.60321271e-01
-1.66624576e-01 1.50313109e-01 -5.30314028e-01 -1.03018805e-01
1.15553415e+00 1.30283713e-01 -2.30647922e-01 -2.41951868e-01
-1.52354407e+00 6.85231388e-01 7.30472982e-01 7.32862413e-01
-6.43902659e-01 1.78304002e-01 5.25418043e-01 5.95224015e-02
4.10518587e-01 4.61778551e-01 -4.70455050e-01 8.45771357e-02
-1.30238593e+00 8.36880803e-02 6.02564253e-02 2.71812946e-01
4.71698135e-01 9.34110358e-02 -4.76352215e-01 8.93220186e-01
3.18539143e-01 1.63340271e-01 7.73204267e-01 -9.20683503e-01
1.04258902e-01 8.20218682e-01 -1.81446791e-01 -9.07529414e-01
-8.84351730e-01 -5.34157336e-01 -9.93770242e-01 1.32174939e-01
5.05828917e-01 -1.33843184e-01 -6.85829401e-01 1.86626315e+00
-2.52591204e-02 4.88152308e-03 -5.29289603e-01 8.80164623e-01
6.53583884e-01 6.88627884e-02 5.53776957e-02 -1.01716518e-01
1.34431648e+00 -4.15111899e-01 -5.26886761e-01 -3.36300671e-01
4.33380514e-01 -1.50885656e-02 6.27088845e-01 7.96854571e-02
-8.58702242e-01 -6.25864148e-01 -7.53226519e-01 1.63934663e-01
-1.61276445e-01 2.99725950e-01 5.47109187e-01 1.64393336e-01
-1.12411284e+00 4.54490840e-01 -1.03587234e+00 -4.88989234e-01
9.43137705e-01 6.84985697e-01 -6.05358899e-01 -3.54968272e-02
-1.18173039e+00 8.86002779e-01 6.25294566e-01 3.71802032e-01
-1.20281351e+00 -8.46363902e-01 -4.32199508e-01 1.45131484e-01
-1.59435406e-01 -7.38390267e-01 6.54372633e-01 -1.13264108e+00
-1.18270636e+00 9.99312520e-01 -2.05909088e-01 -2.89914608e-01
2.73140728e-01 2.55297869e-01 -3.51979196e-01 4.00282115e-01
1.65751934e-01 8.79564703e-01 7.30970562e-01 -6.10420287e-01
-1.73089042e-01 -7.50141978e-01 -1.81165591e-01 -7.28510916e-02
-3.70516390e-01 1.52522013e-01 -1.84859887e-01 -6.27543747e-01
2.13281617e-01 -6.90139234e-01 -1.05752438e-01 1.95474997e-01
-2.80538619e-01 -1.73567221e-01 4.47446078e-01 -6.64085567e-01
9.85224545e-01 -2.33078957e+00 4.49126512e-01 2.49135718e-01
4.89701688e-01 1.42312616e-01 1.92620864e-05 1.97875835e-02
-3.54572952e-01 1.16252188e-04 -3.88138384e-01 1.06904104e-01
-1.32901698e-01 -8.15195590e-02 2.34587729e-01 6.96360886e-01
4.12576616e-01 1.04558861e+00 -6.75483823e-01 -2.32371256e-01
2.72870492e-02 4.78700936e-01 -4.26947832e-01 1.75499499e-01
-6.00305051e-02 1.00540507e+00 -4.57276911e-01 3.73688728e-01
3.95818084e-01 -4.03383613e-01 3.79869670e-01 -4.74302232e-01
-1.55598402e-01 1.53807670e-01 -7.21171677e-01 1.78888702e+00
3.20017524e-02 8.47021937e-01 1.09664120e-01 -1.45607352e+00
8.80480409e-01 5.29551625e-01 8.32554162e-01 -1.02864695e+00
3.09568822e-01 1.52517840e-01 6.75954401e-01 -8.39211524e-01
-5.18856645e-01 -1.58124477e-01 2.79584259e-01 5.14416575e-01
4.15580511e-01 6.95128083e-01 4.04741168e-02 -2.30555445e-01
1.32653987e+00 -1.85040340e-01 6.74732402e-02 -7.67074525e-01
4.40278322e-01 -2.48487517e-01 5.17205596e-01 4.86973166e-01
-5.08886635e-01 3.44819397e-01 7.10042834e-01 -4.45469558e-01
-8.78080010e-01 -8.20655704e-01 -5.08512080e-01 9.27613497e-01
-3.05150807e-01 2.77076483e-01 -6.38799131e-01 -4.59825367e-01
5.94912097e-02 5.88515252e-02 -9.11957085e-01 -2.77412832e-01
-5.58741868e-01 -9.50179219e-01 6.55841112e-01 5.02699435e-01
7.37377822e-01 -1.22583389e+00 -8.01314950e-01 9.33172032e-02
-2.06289366e-02 -1.20998204e+00 -3.81931514e-01 2.72149563e-01
-9.75948870e-01 -1.27979279e+00 -6.95009887e-01 -7.81621397e-01
5.20428717e-01 4.28758329e-03 6.89453483e-01 -1.58832014e-01
-4.43967879e-01 1.77703321e-01 8.08183774e-02 1.09834701e-01
-6.00164197e-03 1.64219111e-01 -1.42177586e-02 4.14845228e-01
2.18398347e-01 -9.57801759e-01 -9.38579321e-01 3.98780107e-01
-7.99122751e-01 -9.07166488e-03 6.94521308e-01 6.93090856e-01
5.12518942e-01 2.47169495e-01 7.87095487e-01 -6.53131366e-01
4.53866392e-01 -6.53773844e-01 -4.32689011e-01 9.08803120e-02
-4.08683389e-01 1.38850138e-01 3.07137340e-01 -3.99498552e-01
-8.63307953e-01 3.09908427e-02 -1.02517776e-01 -4.41253662e-01
-3.09309781e-01 6.30421877e-01 -1.95796952e-01 -2.16666669e-01
4.93876487e-01 3.37611526e-01 2.37534791e-01 -5.23882687e-01
-1.72748759e-01 1.75187349e-01 3.99081171e-01 -2.57491261e-01
8.76185298e-02 6.34809136e-01 9.65802595e-02 -7.05264747e-01
-6.81627333e-01 -3.48096281e-01 -1.13956952e+00 -2.44297042e-01
1.22377479e+00 -8.20430160e-01 -8.89510512e-01 5.60283601e-01
-1.07752085e+00 -5.92927039e-01 2.76271909e-01 5.66182673e-01
-2.72804081e-01 1.87583774e-01 -4.92317438e-01 -4.91117865e-01
-3.70856136e-01 -1.24483073e+00 7.79454947e-01 -2.27405295e-01
-1.52732491e-01 -1.28426957e+00 9.93276089e-02 1.89077497e-01
4.82682645e-01 3.38742763e-01 1.29812086e+00 -5.85700870e-01
-3.95425797e-01 -1.46502405e-01 -4.10690427e-01 1.72046572e-01
1.23376049e-01 -1.76477537e-01 -9.24237013e-01 -1.28510103e-01
9.33789760e-02 -6.78857416e-02 1.01917660e+00 7.98657000e-01
1.30959547e+00 4.24516983e-02 -2.98518181e-01 6.99936450e-01
1.15807986e+00 1.09826900e-01 5.05263686e-01 1.87452540e-01
7.32935071e-01 7.65458882e-01 -1.01128243e-01 2.98131943e-01
1.84315845e-01 6.85857594e-01 5.44100165e-01 -3.20142065e-03
-1.10024922e-01 2.78647393e-01 3.84008259e-01 4.01436716e-01
-5.31936362e-02 1.25707164e-01 -1.11540031e+00 5.84805071e-01
-1.90523028e+00 -1.06209612e+00 -3.29399928e-02 1.90205503e+00
6.06629670e-01 -2.13136133e-02 2.95809150e-01 -1.48701817e-01
7.40131915e-01 9.78698060e-02 -8.92058313e-01 1.58371195e-01
-2.83727467e-01 2.17680693e-01 1.64722979e-01 -4.87903468e-02
-8.98128092e-01 5.23307860e-01 6.81976938e+00 6.23643219e-01
-1.40383089e+00 4.01073933e-01 6.99744403e-01 -2.89307117e-01
6.39449060e-02 -4.08580422e-01 -4.81946498e-01 6.79588079e-01
9.41869915e-01 2.15397865e-01 7.02923656e-01 1.86923921e-01
5.62396705e-01 3.65760587e-02 -9.86261785e-01 9.72375989e-01
-3.07873428e-01 -1.33789384e+00 -1.79930240e-01 1.46006793e-01
4.82642055e-01 4.48248327e-01 8.82196203e-02 -1.26403376e-01
-1.84660122e-01 -1.29542470e+00 6.21394277e-01 8.84441376e-01
7.93885291e-01 -4.66727406e-01 6.21927679e-01 5.23363173e-01
-9.41412628e-01 -3.75571221e-01 -2.78246462e-01 8.23482219e-03
-2.86111027e-01 7.48353899e-01 -3.71763617e-01 7.04269037e-02
7.74748862e-01 1.09110069e+00 -6.36729717e-01 1.08284152e+00
7.44725615e-02 5.04191697e-01 1.04520462e-01 3.49630773e-01
5.23364209e-02 6.80457056e-03 1.96379393e-01 1.13535643e+00
3.37952405e-01 -1.00730032e-01 -4.40917835e-02 1.44192374e+00
-1.74247939e-02 -1.51729658e-01 -3.94358784e-01 -4.01555061e-01
2.37093002e-01 1.50209463e+00 -8.38202357e-01 1.51044726e-01
-1.25693589e-01 6.08428478e-01 7.25462198e-01 3.71254146e-01
-5.09360790e-01 1.13581143e-01 5.57431996e-01 1.38431430e-01
3.56229365e-01 -4.89821166e-01 -4.95034218e-01 -9.63101089e-01
-1.33300558e-01 -5.64138055e-01 2.94890642e-01 -5.71328819e-01
-1.40468156e+00 6.67979360e-01 4.47234437e-02 -7.36462772e-01
1.29436493e-01 -9.61513996e-01 -7.74844348e-01 1.01023197e+00
-1.36895490e+00 -9.01419938e-01 -2.70653218e-01 8.73091042e-01
1.60920590e-01 -2.97101766e-01 7.83039749e-01 4.97244269e-01
-1.04489553e+00 2.00820655e-01 1.13397524e-01 3.48830879e-01
5.50976872e-01 -8.90506983e-01 -7.13815764e-02 6.63362443e-01
-2.57194549e-01 7.65212536e-01 9.61566567e-02 -4.65795070e-01
-1.09993804e+00 -1.09395695e+00 6.48313999e-01 -2.92677701e-01
6.17675245e-01 -5.57054877e-01 -8.96050096e-01 4.78197813e-01
3.36858183e-02 3.04772705e-01 6.02544546e-01 -1.45718426e-01
-2.06657186e-01 -2.33162418e-01 -1.11356020e+00 1.07693471e-01
1.06118894e+00 -7.52790630e-01 -3.16514403e-01 4.09444332e-01
1.72357291e-01 5.03262468e-02 -1.08667350e+00 2.95207471e-01
4.76716757e-01 -1.05074561e+00 9.62561548e-01 -4.80273575e-01
3.80128324e-01 -7.52256289e-02 2.37893257e-02 -1.14554870e+00
-7.58599758e-01 -5.07284105e-02 -4.31245044e-02 1.04941130e+00
2.02418238e-01 -5.28315663e-01 4.65138227e-01 4.79868352e-01
-1.63078994e-01 -6.94602609e-01 -1.20961332e+00 -4.55605567e-01
9.02746767e-02 -3.10777545e-01 2.81259984e-01 8.11650217e-01
-1.68667346e-01 3.84485006e-01 -1.37380168e-01 1.09771520e-01
3.89575779e-01 -9.78887603e-02 -8.79479945e-02 -1.40322161e+00
-1.15148649e-01 -7.14857876e-01 -4.46726888e-01 -6.88752294e-01
5.60078919e-01 -1.25941098e+00 -5.29609881e-02 -1.39518082e+00
4.40209538e-01 -2.13130459e-01 -7.06599474e-01 5.12185335e-01
1.96329579e-02 1.88014925e-01 -2.33909234e-01 3.87366086e-01
-5.15328765e-01 5.31069458e-01 1.36022103e+00 -3.29752453e-02
1.42654255e-01 -3.85225192e-02 -4.46674377e-01 3.86946946e-01
7.05193579e-01 -3.52050096e-01 -3.84472489e-01 -5.31599045e-01
3.50646786e-02 9.70469862e-02 8.52499723e-01 -1.11253726e+00
1.77589491e-01 1.64149806e-01 7.74430335e-01 -1.92950115e-01
1.66383848e-01 -7.45663106e-01 -3.14931870e-02 5.78911245e-01
-5.23615360e-01 8.78540650e-02 2.22680539e-01 5.31327784e-01
-4.22937199e-02 7.19666183e-02 1.07442391e+00 -4.17469829e-01
-6.80902064e-01 7.19074070e-01 -6.43612385e-01 2.00766250e-02
6.18121028e-01 -2.83066034e-01 -2.03748181e-01 8.66146535e-02
-8.97547603e-01 2.24310756e-02 -9.79297049e-03 1.29091308e-01
6.08475745e-01 -1.31240642e+00 -4.36066896e-01 4.38396513e-01
-1.06676094e-01 -3.53459150e-01 6.07594192e-01 1.37743318e+00
-1.08669937e-01 5.55308044e-01 -7.53431261e-01 -7.17447162e-01
-6.38896406e-01 4.16190356e-01 8.93015981e-01 -2.02697799e-01
-5.98810196e-01 7.88400292e-01 4.87731844e-01 -3.78057152e-01
7.41279647e-02 -2.10659593e-01 -5.30100822e-01 9.20875221e-02
2.75270492e-01 2.14544311e-01 1.29344374e-01 -6.65445209e-01
-5.88047266e-01 2.31679305e-01 1.41467795e-01 1.02522738e-01
1.62216794e+00 -1.19769052e-01 -5.05372107e-01 5.39130449e-01
1.21531296e+00 -7.04655647e-01 -1.44092703e+00 -3.71191770e-01
1.77852422e-01 -8.93295631e-02 3.62619340e-01 -7.56688058e-01
-1.37547159e+00 1.16881347e+00 8.85305047e-01 -9.71117765e-02
9.29464459e-01 -4.11273912e-02 5.50227165e-01 9.05167982e-02
1.56382799e-01 -8.67671371e-01 3.66489381e-01 5.91330469e-01
7.14238584e-01 -1.23725426e+00 -2.66370952e-01 8.12278166e-02
-4.06332672e-01 1.31417239e+00 6.26036763e-01 -2.88439304e-01
9.40418005e-01 -1.51787549e-01 -3.49512279e-01 -8.19489062e-01
-6.07119441e-01 -1.05230510e-01 5.83723903e-01 5.25619030e-01
5.96664608e-01 -1.33815497e-01 9.42333341e-02 8.02371800e-01
2.05608845e-01 5.90274595e-02 -1.63220286e-01 5.49080253e-01
-3.76176089e-01 -7.96863139e-01 5.91180585e-02 7.34953463e-01
-5.85878670e-01 -4.70010638e-02 -3.14890951e-01 6.93351030e-01
3.95047992e-01 5.15170097e-01 1.92953646e-01 -2.32193738e-01
7.46698976e-02 2.87510246e-01 5.95726013e-01 -5.81482410e-01
-7.61934400e-01 1.47692621e-01 -3.40893596e-01 -5.01578450e-01
-6.77013278e-01 -7.48632908e-01 -1.29909265e+00 -5.95511980e-02
1.00220986e-01 -2.23805383e-01 3.52675527e-01 1.33987498e+00
6.57419026e-01 8.24665606e-01 4.27854538e-01 -8.94829750e-01
-1.48348317e-01 -1.08024108e+00 -8.52955699e-01 3.52225691e-01
3.89414042e-01 -9.71117318e-01 -2.19211634e-02 -1.11649558e-01] | [12.526934623718262, 3.342341423034668] |
7084d139-de17-48de-86e1-b60e47b18c45 | demograsp-few-shot-learning-for-robotic | 2112.02849 | null | https://arxiv.org/abs/2112.02849v1 | https://arxiv.org/pdf/2112.02849v1.pdf | DemoGrasp: Few-Shot Learning for Robotic Grasping with Human Demonstration | The ability to successfully grasp objects is crucial in robotics, as it enables several interactive downstream applications. To this end, most approaches either compute the full 6D pose for the object of interest or learn to predict a set of grasping points. While the former approaches do not scale well to multiple object instances or classes yet, the latter require large annotated datasets and are hampered by their poor generalization capabilities to new geometries. To overcome these shortcomings, we propose to teach a robot how to grasp an object with a simple and short human demonstration. Hence, our approach neither requires many annotated images nor is it restricted to a specific geometry. We first present a small sequence of RGB-D images displaying a human-object interaction. This sequence is then leveraged to build associated hand and object meshes that represent the depicted interaction. Subsequently, we complete missing parts of the reconstructed object shape and estimate the relative transformation between the reconstruction and the visible object in the scene. Finally, we transfer the a-priori knowledge from the relative pose between object and human hand with the estimate of the current object pose in the scene into necessary grasping instructions for the robot. Exhaustive evaluations with Toyota's Human Support Robot (HSR) in real and synthetic environments demonstrate the applicability of our proposed methodology and its advantage in comparison to previous approaches. | ['Benjamin Busam', 'Nassir Navab', 'Sven Meie', 'Lorenzo Garattoni', 'Luca Minciullo', 'Fabian Manhardt', 'Pengyuan Wang'] | 2021-12-06 | null | null | null | null | ['robotic-grasping'] | ['robots'] | [ 1.91504180e-01 2.30627641e-01 1.85710728e-01 -2.47338921e-01
-3.98566306e-01 -7.02721179e-01 4.49068427e-01 8.55817944e-02
-1.49227053e-01 4.70847219e-01 -5.90690255e-01 2.93332078e-02
-4.83782738e-01 -6.26646996e-01 -8.38747978e-01 -6.17782593e-01
-1.36851713e-01 1.09255219e+00 4.17861938e-01 -2.71082252e-01
2.56525278e-01 9.91204202e-01 -1.72697902e+00 6.34158179e-02
7.34982312e-01 9.89515722e-01 8.74166071e-01 4.58134294e-01
5.45509495e-02 3.20164502e-01 -2.81580031e-01 -7.47895893e-03
4.55086708e-01 4.65872586e-02 -7.74947047e-01 5.50170600e-01
-2.81881914e-03 -6.07946634e-01 -1.07687473e-01 7.03657568e-01
1.59109429e-01 1.80050477e-01 6.81035399e-01 -1.30168962e+00
-2.39387285e-02 4.39726979e-01 -3.77365589e-01 -5.16117096e-01
7.01542199e-01 1.79519802e-01 5.62605202e-01 -9.64225292e-01
9.54801917e-01 1.23512912e+00 4.52764034e-01 5.66899419e-01
-9.42365348e-01 -2.04144374e-01 1.44268483e-01 6.04285039e-02
-1.02386475e+00 -2.00066119e-01 8.41594815e-01 -7.17493296e-01
5.82295060e-01 -1.54613452e-02 8.15478742e-01 9.66034055e-01
-1.38788730e-01 6.86526597e-01 9.67384994e-01 -5.59243798e-01
2.93395847e-01 1.21565357e-01 -3.46006379e-02 7.22934723e-01
8.89537036e-02 -3.64814363e-02 -2.09491521e-01 8.16959217e-02
1.03337228e+00 2.51619011e-01 -2.20955193e-01 -1.27095044e+00
-1.38737798e+00 2.83549517e-01 5.28955817e-01 1.82784274e-01
-6.82654619e-01 9.17603150e-02 1.68958917e-01 1.65288411e-02
-1.20005719e-01 2.37352505e-01 -4.79563236e-01 4.05706577e-02
-2.99483180e-01 4.25856858e-01 9.11752820e-01 1.43577635e+00
7.01242685e-01 -3.73324096e-01 3.40387821e-01 4.15862679e-01
2.33597353e-01 3.66351306e-01 -2.38133490e-01 -1.09532070e+00
5.87104261e-01 7.29956031e-01 6.57750368e-01 -9.18447852e-01
-5.35523355e-01 6.86479211e-02 -2.86659032e-01 6.99196696e-01
7.14403629e-01 1.95450261e-01 -7.71960735e-01 1.18409884e+00
7.17240214e-01 -3.76007974e-01 2.36203647e-04 1.02272081e+00
3.81289959e-01 3.70648623e-01 -2.04255894e-01 3.06935813e-02
1.17262316e+00 -8.47637177e-01 -3.46172005e-01 -3.06311339e-01
3.82958859e-01 -6.47659123e-01 9.98401225e-01 6.66241705e-01
-9.77348268e-01 -6.10063493e-01 -8.98556292e-01 -3.28322202e-02
-2.32200965e-01 4.63441730e-01 6.49595797e-01 1.40600830e-01
-5.10592520e-01 8.16119969e-01 -1.03516185e+00 -5.38378000e-01
1.37807980e-01 5.62134385e-01 -4.97711599e-01 -3.52456838e-01
-4.32623833e-01 1.27062035e+00 7.39329457e-01 4.21117783e-01
-9.00519669e-01 -3.19923252e-01 -6.03825748e-01 -8.62267017e-02
7.32225120e-01 -4.72616524e-01 1.21023667e+00 -6.60053074e-01
-1.70496941e+00 7.06576169e-01 3.69315058e-01 -6.70449510e-02
8.95799875e-01 -4.65339243e-01 4.82869595e-01 3.38355333e-01
-1.68970495e-01 6.49487317e-01 8.53154838e-01 -1.83102417e+00
-5.34064591e-01 -5.41455269e-01 5.14319062e-01 3.33234876e-01
1.59396097e-01 -4.23327386e-01 -4.48120117e-01 -3.41275811e-01
5.90398550e-01 -1.01387179e+00 -2.02907920e-01 3.97190988e-01
-3.20229173e-01 -2.52225608e-01 1.13661134e+00 -5.24508178e-01
5.21549247e-02 -2.17575526e+00 6.95758402e-01 2.49488369e-01
-2.48386756e-01 4.07953281e-03 -2.85630506e-02 7.52658129e-01
2.13417724e-01 -5.96447945e-01 -1.38858333e-01 -2.76751220e-01
1.00178421e-01 3.64597768e-01 -2.51167119e-01 5.39901257e-01
3.49305086e-02 5.71789265e-01 -9.80548203e-01 -3.63776207e-01
5.87235570e-01 4.77019101e-01 -4.84133780e-01 5.21609247e-01
-6.11804962e-01 9.40162659e-01 -8.06228817e-01 6.29675150e-01
5.87609231e-01 1.83488041e-01 2.51951039e-01 -3.36423874e-01
-1.65518910e-01 -1.60558373e-01 -1.39931035e+00 1.90987551e+00
-5.44935346e-01 -7.67786130e-02 4.78601158e-01 -9.76531506e-01
1.13401878e+00 3.65879387e-01 6.35565042e-01 5.03504612e-02
2.61763811e-01 5.37830949e-01 -1.35536060e-01 -8.97188127e-01
1.29105836e-01 -4.05300129e-03 1.64430603e-01 3.29946041e-01
1.14537567e-01 -7.77128279e-01 1.78035181e-02 -2.29265258e-01
7.20992148e-01 8.11831653e-01 3.05545181e-01 -8.19924325e-02
4.08707142e-01 1.87967658e-01 -1.30558433e-02 3.81534189e-01
3.01639348e-01 4.74504471e-01 1.98543534e-01 -4.00433779e-01
-1.23712802e+00 -9.85342026e-01 8.73246938e-02 7.43503690e-01
4.99321401e-01 2.05720589e-01 -7.58706689e-01 -5.90954840e-01
1.91313386e-01 6.36591077e-01 -3.81048292e-01 1.71624303e-01
-7.27820337e-01 4.03234623e-02 -1.79936260e-01 5.18384039e-01
2.26637632e-01 -1.31675994e+00 -1.40199673e+00 1.88962638e-01
-1.07535059e-02 -1.28145456e+00 1.42248839e-01 5.48747703e-02
-1.08238077e+00 -1.38307011e+00 -6.63553238e-01 -9.73577559e-01
1.01760519e+00 1.09891199e-01 5.66896260e-01 -2.82918792e-02
-5.43667674e-01 7.50984192e-01 -5.53466141e-01 -2.74258822e-01
-6.00323856e-01 -1.87634975e-01 1.75398991e-01 -1.33703128e-01
-1.83811501e-01 -6.07962310e-01 -5.49166501e-01 3.89455497e-01
-6.13534570e-01 2.03838170e-01 7.60530353e-01 6.41798437e-01
3.95196825e-01 6.58042878e-02 3.16084653e-01 -3.68234813e-01
1.07591189e-01 -1.02207959e-01 -7.02071249e-01 2.91951418e-01
1.47688866e-01 -2.56897032e-01 4.80788320e-01 -6.56169295e-01
-1.04471803e+00 7.71083593e-01 2.07995519e-01 -5.84268808e-01
-4.67969090e-01 2.85196692e-01 -2.81191200e-01 2.25076582e-02
2.71841526e-01 -8.38329494e-02 8.14684480e-02 -6.71186030e-01
4.16389495e-01 5.10350883e-01 6.84343696e-01 -8.69021237e-01
7.41652966e-01 4.61819798e-01 2.93409675e-01 -7.10678816e-01
-3.81344289e-01 -4.03057247e-01 -1.45188498e+00 -4.51457709e-01
7.06103265e-01 -3.87503773e-01 -1.38348413e+00 3.76942724e-01
-1.39056408e+00 -3.88732105e-01 -2.35960916e-01 6.40906394e-01
-1.15266979e+00 3.68939012e-01 -3.56008857e-01 -1.08449543e+00
-6.98020235e-02 -1.41994119e+00 1.33220446e+00 -2.24598452e-01
1.59805529e-02 -4.68532413e-01 -6.04967535e-01 2.80911952e-01
-4.97482531e-02 6.41785264e-01 1.05355728e+00 -3.73936653e-01
-7.91001976e-01 -4.29611176e-01 -5.86542301e-02 1.01322860e-01
2.84275681e-01 -2.86134947e-02 -5.21959484e-01 -5.07256031e-01
7.61256069e-02 -4.41437274e-01 1.35952577e-01 7.35193267e-02
9.76288855e-01 -3.71784605e-02 -5.71645916e-01 -1.25651378e-02
1.40859914e+00 2.91705877e-01 2.89387524e-01 2.75574476e-01
6.11207724e-01 1.16491354e+00 1.29128468e+00 4.81527060e-01
-3.60530280e-02 1.05667078e+00 8.16152930e-01 3.25358719e-01
9.28412601e-02 -1.43517464e-01 1.92198068e-01 5.93451619e-01
-4.72486705e-01 1.64201949e-02 -1.02297413e+00 2.36994088e-01
-1.76818085e+00 -4.57643062e-01 -1.24012344e-01 2.15715766e+00
4.00757194e-01 9.98808146e-02 1.00989118e-01 3.42568696e-01
4.79033142e-01 -6.52792335e-01 -6.92999721e-01 4.00274061e-02
5.54928243e-01 -1.13778010e-01 1.75576210e-01 4.54189301e-01
-6.37800992e-01 8.54810417e-01 5.07006645e+00 1.54272199e-01
-8.68019342e-01 -2.06747070e-01 -3.40164602e-01 2.87366390e-01
3.10340196e-01 6.91246316e-02 -4.32925701e-01 -3.91667709e-02
1.12634584e-01 2.24080324e-01 4.94701475e-01 1.08734071e+00
-2.12570466e-02 -3.73177439e-01 -1.78910089e+00 6.77721083e-01
-7.46834800e-02 -7.00239360e-01 4.69076522e-02 -1.66411400e-01
2.30004460e-01 -4.79453385e-01 -2.03436032e-01 6.56556040e-02
-4.50004339e-02 -7.92559445e-01 1.06444359e+00 6.53546393e-01
5.80830574e-01 -4.85157579e-01 5.52720070e-01 9.24717247e-01
-9.87815440e-01 -3.36932361e-01 -2.09021702e-01 -7.50721991e-02
2.55215764e-01 -3.10104340e-02 -1.33815002e+00 6.52210414e-01
6.27710104e-01 2.23988310e-01 -1.86537638e-01 9.01145995e-01
-1.95095152e-01 -1.60687983e-01 -3.10904443e-01 -6.98148366e-03
9.84827876e-02 -2.57588387e-01 7.00797975e-01 6.66087627e-01
4.02751267e-01 4.60049659e-01 4.34903264e-01 7.59765208e-01
3.96071762e-01 -1.04670674e-02 -6.15758896e-01 1.60686702e-01
3.06857854e-01 1.03005171e+00 -9.89904761e-01 -1.53645590e-01
-8.94828886e-02 1.08741176e+00 2.85961717e-01 2.70705581e-01
-4.56408024e-01 -1.80498376e-01 9.76319537e-02 1.59167275e-01
4.14755315e-01 -5.93920887e-01 -4.15341966e-02 -8.32481563e-01
3.47088367e-01 -7.09907949e-01 -2.17714548e-01 -1.09626925e+00
-8.63578856e-01 4.76301819e-01 3.09133112e-01 -1.40822184e+00
-3.14762682e-01 -1.02505481e+00 -4.31788564e-02 6.83961809e-01
-1.00150931e+00 -1.41577291e+00 -7.03923821e-01 2.86912382e-01
9.14203763e-01 1.14661284e-01 9.30974603e-01 -1.81774333e-01
4.65671793e-02 -1.56781211e-01 -1.87860817e-01 -1.74645841e-01
3.92511904e-01 -1.12103415e+00 8.36490653e-03 2.23367840e-01
-2.60361642e-01 5.17143965e-01 8.83162498e-01 -6.06275260e-01
-1.80887187e+00 -5.63577771e-01 1.00985907e-01 -6.31963849e-01
4.12334770e-01 -5.43062389e-01 -8.85485172e-01 8.45772028e-01
-2.35329717e-01 5.42212054e-02 -2.58746624e-01 -3.13984007e-01
6.53815493e-02 2.68368963e-02 -1.38156366e+00 4.88720149e-01
1.01112890e+00 -6.00395575e-02 -7.76301146e-01 4.58842099e-01
4.77250218e-01 -8.18292379e-01 -9.33263063e-01 6.66653931e-01
8.33971918e-01 -7.71026134e-01 1.00430596e+00 -5.74131668e-01
4.69351768e-01 -3.53351265e-01 -2.39848375e-01 -1.11956394e+00
1.07501514e-01 -3.22795033e-01 -3.74448709e-02 8.99157345e-01
-9.85293463e-02 -3.26166689e-01 8.79155874e-01 5.71061790e-01
-1.49871066e-01 -8.87243032e-01 -8.23542237e-01 -7.01879501e-01
-9.44644883e-02 -2.37422034e-01 4.69150633e-01 6.05342865e-01
1.31556138e-01 -1.12316303e-01 -1.86351046e-01 6.33220196e-01
7.74878502e-01 5.42926013e-01 1.11185491e+00 -1.40036643e+00
-2.69669563e-01 -1.33716717e-01 -4.20200706e-01 -9.76250529e-01
1.91447392e-01 -6.29808009e-01 4.44269270e-01 -1.61043656e+00
1.07602030e-01 -9.24216270e-01 3.91771227e-01 3.83611351e-01
2.29323581e-01 -1.63730770e-01 3.76056373e-01 3.00587088e-01
-1.25695258e-01 3.33096743e-01 1.49322724e+00 1.73532203e-01
-2.88422674e-01 3.45432431e-01 4.38899606e-01 9.13762212e-01
6.09796405e-01 -2.32347295e-01 -2.90444583e-01 -4.18589890e-01
-1.34943739e-01 5.76874673e-01 7.81485796e-01 -9.95793343e-01
1.02272332e-01 -2.84996659e-01 2.94470966e-01 -7.40316391e-01
7.03837633e-01 -1.51776659e+00 3.10976505e-01 6.88834846e-01
-2.50098705e-01 -2.13537127e-01 9.68834087e-02 5.63879013e-01
2.36509860e-01 -6.07029259e-01 5.22301733e-01 -3.99066061e-01
-6.47352159e-01 1.33603528e-01 3.26155946e-02 -6.27891362e-01
1.44759357e+00 -3.80360395e-01 3.01590472e-01 -2.95026228e-02
-1.13296258e+00 1.72395542e-01 7.16314971e-01 4.54288691e-01
7.23295510e-01 -8.73186290e-01 -4.00323540e-01 1.63827136e-01
5.75197898e-02 6.37151361e-01 9.38713178e-02 7.19051123e-01
-7.54270434e-01 2.62712061e-01 -3.95484358e-01 -9.69974101e-01
-1.27426898e+00 8.81404042e-01 2.26345539e-01 3.74264479e-01
-9.66329753e-01 4.36264038e-01 5.64299077e-02 -6.49470448e-01
5.31169593e-01 -5.41467786e-01 8.97309277e-04 -2.84942687e-01
4.94913608e-02 5.27844548e-01 4.26420867e-02 -5.94094753e-01
-1.38769880e-01 8.13587785e-01 2.41410017e-01 -4.51832190e-02
1.55708098e+00 -9.24853906e-02 -1.78985193e-01 6.21492565e-01
9.14524674e-01 -1.35503143e-01 -1.58162189e+00 -7.21083060e-02
1.07480057e-01 -5.65468252e-01 -6.45065904e-01 -6.91515088e-01
-6.62023842e-01 9.82442379e-01 4.55405772e-01 -1.51120983e-02
8.45624983e-01 3.96382362e-01 3.75798196e-01 8.50782096e-01
1.09150493e+00 -7.98399627e-01 3.40512991e-01 2.94233054e-01
1.45433223e+00 -1.08728421e+00 1.19460918e-01 -9.49815273e-01
-4.77172971e-01 1.49831879e+00 6.12675071e-01 -1.90566435e-01
4.02640939e-01 2.29112417e-01 3.17652039e-02 -3.48624408e-01
-2.26910949e-01 1.33585870e-01 2.39432603e-01 6.08533442e-01
-1.30552799e-01 -4.35441621e-02 1.27633080e-01 1.03877515e-01
-2.21607298e-01 1.68001011e-01 4.15733993e-01 1.33019054e+00
-5.14979362e-01 -8.97665918e-01 -4.79076505e-01 2.29097754e-02
4.79214415e-02 7.63133168e-01 -2.69949317e-01 1.04276514e+00
1.81251362e-01 5.89354694e-01 -1.42947242e-01 -6.97011128e-02
8.20879877e-01 -3.17545421e-02 1.10548735e+00 -8.63633692e-01
-2.63880968e-01 -2.28889361e-02 -2.44313881e-01 -5.61683357e-01
-5.73890746e-01 -7.97816932e-01 -1.52003300e+00 3.17903072e-01
-4.39770192e-01 -1.06603436e-01 9.49760616e-01 9.77167487e-01
-1.06080711e-01 3.04290175e-01 3.94185424e-01 -1.72657323e+00
-8.68306756e-01 -9.00097311e-01 -4.56612438e-01 4.64485973e-01
2.77737916e-01 -1.13699865e+00 -4.01716083e-02 9.37666893e-02] | [5.892202854156494, -0.8650883436203003] |
8a9f0107-aba4-4895-949d-27bc20ebeec1 | can-transfer-entropy-infer-causality-in | 1901.07589 | null | https://arxiv.org/abs/1901.07589v2 | https://arxiv.org/pdf/1901.07589v2.pdf | Can Transfer Entropy Infer Information Flow in Neuronal Circuits for Cognitive Processing? | To infer information flow in any network of agents, it is important first and foremost to establish causal temporal relations between the nodes. Practical and automated methods that can infer causality are difficult to find, and the subject of ongoing research. While Shannon information only detects correlation, there are several information-theoretic notions of "directed information" that have successfully detected causality in some systems, in particular in the neuroscience community. However, recent work has shown that some directed information measures can sometimes inadequately estimate the extent of causal relations, or even fail to identify existing cause-effect relations between components of systems, especially if neurons contribute in a cryptographic manner to influence the effector neuron. Here, we test how often cryptographic logic emerges in an evolutionary process that generates artificial neural circuits for two fundamental cognitive tasks: motion detection and sound localization. We also test whether activity time-series recorded from behaving digital brains can infer information flow using the transfer entropy concept, when compared to a ground-truth model of causal influence constructed from connectivity and circuit logic. Our results suggest that transfer entropy will sometimes fail to infer causality when it exists, and sometimes suggest a causal connection when there is none. However, the extent of incorrect inference strongly depends on the cognitive task considered. These results emphasize the importance of understanding the fundamental logic processes that contribute to information flow in cognitive processing, and quantifying their relevance in any given nervous system. | ['Ali Tehrani-Saleh', 'Christoph Adami'] | 2019-01-22 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 6.40227020e-01 1.08128726e-01 1.36255309e-01 8.73584747e-02
3.04952651e-01 -8.98423731e-01 9.51879680e-01 3.83005828e-01
-2.23910108e-01 9.40318346e-01 2.58353680e-01 -5.51633775e-01
-7.69146323e-01 -8.98930967e-01 -5.45842528e-01 -6.55962825e-01
-4.50458884e-01 1.70971170e-01 2.82517046e-01 -1.79176763e-01
6.59900784e-01 3.25203776e-01 -1.34862697e+00 7.86981955e-02
5.98941505e-01 7.18908310e-01 1.92410663e-01 9.52732801e-01
1.21348001e-01 1.55210233e+00 -6.58595383e-01 -2.33809471e-01
-2.08522797e-01 -1.18931067e+00 -9.74183559e-01 -4.80643690e-01
-1.16937764e-01 -1.46018758e-01 -2.88487434e-01 1.00596833e+00
3.72589342e-02 -2.88834989e-01 8.43568504e-01 -1.40964520e+00
-5.83550811e-01 8.60604823e-01 -7.59618506e-02 7.68557072e-01
5.15213013e-01 2.22034782e-01 1.20929337e+00 -1.26788303e-01
6.34606361e-01 1.18583095e+00 4.86477971e-01 2.43635982e-01
-1.51351023e+00 -4.99773204e-01 -2.20548641e-02 4.24125612e-01
-1.01276565e+00 -4.27250475e-01 5.93869328e-01 -7.79865980e-01
7.78918147e-01 2.90105134e-01 1.26334155e+00 9.55001712e-01
9.16446507e-01 2.06268504e-01 1.14231932e+00 -2.34087497e-01
4.58179504e-01 -3.08696508e-01 8.73695314e-02 7.29095101e-01
6.72722220e-01 2.66128808e-01 -1.00465965e+00 -2.52100170e-01
7.83965349e-01 -4.47272331e-01 -4.83427912e-01 2.93725103e-01
-1.47430217e+00 5.44912994e-01 2.61122793e-01 7.62521148e-01
-3.24180096e-01 6.72846079e-01 1.83187723e-01 5.88482261e-01
1.23047248e-01 8.73555481e-01 -3.68589997e-01 -2.84907073e-01
-5.97420514e-01 1.60548314e-01 1.14996386e+00 2.64012009e-01
5.00809789e-01 -3.56790900e-01 3.75192314e-01 -4.02820902e-03
4.64955389e-01 1.88687563e-01 2.64885992e-01 -1.12300384e+00
-1.80667624e-01 7.76907623e-01 -3.12812358e-01 -1.48183715e+00
-3.37114006e-01 -2.26041228e-01 -5.84018350e-01 3.31669807e-01
5.70779622e-01 -2.37312868e-01 2.06839498e-02 1.96966505e+00
-2.19851494e-01 -3.67593318e-02 -6.14883788e-02 8.72377932e-01
2.64356613e-01 3.60251099e-01 -7.79930651e-02 -4.68991011e-01
1.21959245e+00 1.86032996e-01 -8.25796485e-01 -2.53722847e-01
4.57894504e-01 -4.69746649e-01 5.72533667e-01 3.91667932e-01
-1.00522530e+00 2.63592843e-02 -1.27422249e+00 2.39251330e-01
-2.78402150e-01 -7.34567463e-01 9.22562063e-01 5.26234925e-01
-1.21232951e+00 6.87616706e-01 -7.72683024e-01 -3.55584115e-01
3.14725339e-01 2.64089584e-01 -1.02402158e-01 3.32053006e-01
-1.36962402e+00 1.04573143e+00 2.40570948e-01 9.69163105e-02
-9.64171171e-01 -6.19451344e-01 -2.85908967e-01 4.10538092e-02
2.59571373e-01 -8.58202577e-01 9.20665622e-01 -9.69210923e-01
-1.00769091e+00 4.75700259e-01 -5.02944924e-02 -3.40261549e-01
2.57904619e-01 4.04609352e-01 -1.61014274e-01 3.61046821e-01
-5.22917733e-02 4.85812962e-01 2.46054724e-01 -1.08537233e+00
-3.53334993e-01 -5.87784469e-01 2.49982402e-01 1.05090179e-01
-1.33832455e-01 -1.11328550e-01 3.51140290e-01 -2.74008214e-01
3.20579201e-01 -7.86441028e-01 -1.47499954e-02 1.84660271e-01
-3.45259756e-01 -1.29579991e-01 5.51050365e-01 -2.08610684e-01
1.15295339e+00 -1.71549690e+00 1.95465818e-01 2.19643608e-01
6.14630938e-01 -3.92684519e-01 1.86460033e-01 5.89613855e-01
-8.94077122e-02 5.15032649e-01 -4.94476825e-01 6.00222170e-01
-2.62841880e-01 7.04567432e-02 -2.56110311e-01 6.70709431e-01
2.49822214e-01 7.81397879e-01 -1.19362199e+00 -4.63238269e-01
-1.19033344e-01 5.92795968e-01 -5.37641168e-01 -8.55561122e-02
-3.20958376e-01 5.86230755e-01 -4.51706618e-01 1.90377861e-01
-1.72128901e-01 -4.20512527e-01 4.52246100e-01 7.64919519e-02
-4.61116552e-01 7.02612281e-01 -7.42610037e-01 1.23467231e+00
-1.33919537e-01 1.53460801e+00 -2.83280522e-01 -9.98929441e-01
5.79371691e-01 5.48560977e-01 5.41304529e-01 -6.75798476e-01
4.33756918e-01 2.57352330e-02 9.29051876e-01 -5.60529888e-01
1.84607655e-02 -2.03687429e-01 9.24352631e-02 8.54946494e-01
-3.28584798e-02 -2.61196703e-01 3.98553878e-01 4.05309916e-01
1.85826778e+00 -2.08941147e-01 4.78284538e-01 -5.27114511e-01
2.36568347e-01 1.16698287e-01 3.80385756e-01 5.36885440e-01
-2.50011683e-01 9.27709118e-02 1.31838965e+00 -6.19738549e-02
-8.11301947e-01 -1.03255701e+00 -2.51193643e-01 6.85024023e-01
2.91056514e-01 -3.33293110e-01 -8.26561689e-01 -3.85977440e-02
-2.94489503e-01 8.18966985e-01 -8.08422148e-01 -4.39780891e-01
-1.57755509e-01 -8.80999506e-01 7.44594216e-01 7.69430166e-03
3.29200983e-01 -1.11908865e+00 -1.22555673e+00 1.92994118e-01
-2.38282725e-01 -7.70468652e-01 4.94356304e-02 3.50336790e-01
-9.67962682e-01 -1.44879484e+00 4.36405912e-02 -3.11572731e-01
8.93239498e-01 -6.97542652e-02 8.87392461e-01 3.04966331e-01
-4.14231360e-01 3.93390834e-01 -1.07638016e-01 -3.04008543e-01
-5.44718444e-01 -5.29572964e-01 -7.38592744e-02 -3.68646801e-01
1.79489255e-01 -8.06139588e-01 -5.13927281e-01 3.95453900e-01
-9.67918098e-01 -1.11495256e-01 5.06053686e-01 5.33472836e-01
3.85377072e-02 3.96450490e-01 5.21164715e-01 -6.01705551e-01
9.56305981e-01 -5.52921653e-01 -4.26165342e-01 7.26476610e-02
-7.42453396e-01 3.88727188e-01 3.52343291e-01 -2.33117685e-01
-8.64098608e-01 -2.44458511e-01 3.69058341e-01 2.93392420e-01
-1.58227280e-01 8.01365674e-01 -6.27899319e-02 -5.93870692e-02
8.72864604e-01 1.19553037e-01 2.81209759e-02 3.68390024e-01
1.24382325e-01 9.41472948e-02 3.24360073e-01 -4.31129664e-01
3.23300928e-01 5.30604303e-01 4.38340545e-01 -8.72717857e-01
-2.37596169e-01 9.37904269e-02 -7.01469064e-01 -6.55826449e-01
8.59206975e-01 -2.16333508e-01 -1.29306650e+00 1.28030345e-01
-1.39330018e+00 -3.22531313e-01 -3.10677309e-02 5.66538751e-01
-7.52481580e-01 -1.54200047e-02 -5.29523313e-01 -1.06366169e+00
1.82399720e-01 -8.97609890e-01 3.22406352e-01 -5.71630038e-02
-7.41175890e-01 -1.34728932e+00 4.08812970e-01 8.53465647e-02
2.91243464e-01 2.83405125e-01 1.02291322e+00 -3.33283305e-01
-7.32857347e-01 -6.67128041e-02 4.20379303e-02 -3.54120195e-01
1.81725428e-01 3.41146737e-01 -8.67804110e-01 3.53338242e-01
1.74178794e-01 4.65594791e-02 6.49272382e-01 4.88980472e-01
4.73182470e-01 -6.08455837e-01 -4.87064600e-01 -1.67952552e-02
1.27994573e+00 5.80011070e-01 6.61547303e-01 2.75025261e-03
3.17933679e-01 1.22347713e+00 -8.34093019e-02 8.36837217e-02
1.28820062e-01 2.47252926e-01 4.72355843e-01 4.77414191e-01
-3.16639594e-03 -1.95901439e-01 4.30401891e-01 9.79110360e-01
-2.82552719e-01 -4.09557849e-01 -9.34783936e-01 3.69049132e-01
-1.67310929e+00 -1.35861301e+00 -5.04149735e-01 2.20766473e+00
9.95238841e-01 3.94918382e-01 -3.91695760e-02 4.71440822e-01
6.38201118e-01 -2.70802438e-01 -5.55606127e-01 -1.26016736e-01
-1.25442758e-01 -3.00209641e-01 3.91515046e-01 6.35707498e-01
-3.05269212e-01 5.13201058e-01 7.51736927e+00 -1.18864875e-03
-8.32901955e-01 -4.90517216e-03 6.73052192e-01 -1.54261105e-02
-5.78498483e-01 3.83521080e-01 7.60007603e-03 3.61399978e-01
1.14973605e+00 -5.46621978e-01 6.15899265e-01 -8.01749825e-02
4.30475682e-01 -6.07778311e-01 -1.47895813e+00 5.39391696e-01
-1.33862421e-01 -9.94625092e-01 -3.55323911e-01 3.86153966e-01
4.25108135e-01 -3.19473445e-01 -1.77780762e-01 -6.13441825e-01
6.44963026e-01 -1.01946580e+00 7.69219697e-01 8.40014517e-01
2.37245604e-01 -3.59789848e-01 4.31673497e-01 4.23817515e-01
-9.85226750e-01 2.37093084e-02 -6.90038353e-02 -7.73249567e-01
1.96307704e-01 8.83382082e-01 -8.60071540e-01 -2.16884196e-01
3.14721733e-01 4.69951779e-01 -4.21234995e-01 9.08908606e-01
-5.36211550e-01 6.91885412e-01 -1.95919722e-01 -6.19870961e-01
-2.69553035e-01 -6.93627968e-02 7.82274961e-01 7.68326163e-01
1.38925895e-01 3.40744793e-01 -8.58596921e-01 1.50915444e+00
3.54136199e-01 -3.65887195e-01 -1.06730962e+00 -6.24316454e-01
6.20527923e-01 9.59683418e-01 -1.36618245e+00 -7.07368106e-02
-1.55489922e-01 6.55322731e-01 -5.41362204e-02 1.90607429e-01
-5.86365283e-01 -1.42320707e-01 6.51973128e-01 3.78646292e-02
-3.12761039e-01 -5.48895121e-01 -6.19714797e-01 -7.18337893e-01
-3.46291900e-01 -3.45392466e-01 5.88517636e-02 -7.47592628e-01
-1.09192657e+00 2.41599947e-01 -5.60703985e-02 -6.54368103e-01
-4.54032779e-01 -5.12382150e-01 -5.43457270e-01 5.24526119e-01
-6.56485677e-01 -2.10559189e-01 -1.12349994e-01 3.97009850e-01
4.55342904e-02 1.89680368e-01 5.51138878e-01 -1.36554748e-01
-4.36996251e-01 -2.42641553e-01 -3.99975568e-01 1.07560912e-03
2.00034067e-01 -1.19196022e+00 1.65390208e-01 7.70112753e-01
2.90398806e-01 8.45695198e-01 1.02194107e+00 -8.23166907e-01
-1.56002343e+00 -4.04299170e-01 9.51235533e-01 -6.07953846e-01
1.06905901e+00 -3.45209181e-01 -6.21259749e-01 4.60261554e-01
2.94918984e-01 -3.92204136e-01 7.44390011e-01 6.13032468e-02
-3.31268698e-01 1.48840457e-01 -9.07014370e-01 8.56661379e-01
1.30540562e+00 -7.07533479e-01 -7.55041659e-01 6.74594939e-02
5.64108312e-01 5.03987491e-01 -6.46812737e-01 8.64565521e-02
7.98490226e-01 -1.28180456e+00 7.27782428e-01 -1.86935335e-01
7.09621847e-01 -4.15083528e-01 8.79828632e-03 -1.31473720e+00
-3.87634873e-01 -4.19226527e-01 2.14643404e-01 9.80219781e-01
6.59537792e-01 -8.34799588e-01 3.56618583e-01 6.38189614e-01
2.30687648e-01 -2.27973685e-01 -6.96960986e-01 -5.49143970e-01
-3.79725695e-02 -6.27276123e-01 9.39261243e-02 1.24434984e+00
8.47688317e-01 6.13966644e-01 3.14843148e-01 -4.39271927e-02
7.13191986e-01 -3.81573945e-01 7.89408088e-02 -1.54060984e+00
-2.60837734e-01 -1.07196188e+00 -9.05117154e-01 -3.67086649e-01
5.45189939e-02 -7.27499485e-01 2.14210555e-01 -1.55168164e+00
2.39579350e-01 -1.53055921e-01 -6.73532337e-02 2.72564888e-01
1.25840634e-01 5.49943186e-02 -5.75977862e-02 3.59393835e-01
-1.36513039e-01 1.04292341e-01 1.12602687e+00 -1.28271943e-02
-1.08241811e-01 -4.83921856e-01 -8.47053468e-01 9.25493062e-01
7.09924757e-01 -4.75252002e-01 -6.40546620e-01 -1.06498249e-01
1.05327535e+00 2.86247402e-01 8.40870798e-01 -1.07741308e+00
6.70183241e-01 -2.45490476e-01 3.12788010e-01 -4.18757051e-02
9.02296137e-03 -8.13869417e-01 4.09731597e-01 8.93580675e-01
-8.09544861e-01 5.09626865e-02 9.50632617e-03 6.42031431e-01
1.44488178e-02 -1.78527281e-01 3.02831143e-01 -2.20268980e-01
-5.51664293e-01 -1.75627917e-01 -1.03048909e+00 -3.04686110e-02
8.58498096e-01 -7.37182572e-02 -7.56950438e-01 -4.61014241e-01
-3.88388485e-01 -1.84296861e-01 2.07202196e-01 6.76749200e-02
6.52147472e-01 -9.92518902e-01 -4.64042127e-01 -1.95226595e-01
-2.69120455e-01 -6.54599369e-01 -2.27054611e-01 1.20884764e+00
-5.80428660e-01 6.27026558e-01 -3.26560766e-01 -4.32260960e-01
-8.03467751e-01 4.31289971e-01 3.86798382e-01 3.07945132e-01
-2.22076938e-01 7.83111811e-01 4.13645029e-01 2.95926630e-01
-2.29881972e-01 -3.76282930e-01 -1.67590380e-01 3.33079427e-01
5.37548244e-01 4.28351253e-01 -3.21580142e-01 -3.93165737e-01
-5.53848624e-01 3.42261493e-01 4.88596320e-01 -6.34768426e-01
9.80328798e-01 -2.33932272e-01 -6.61247730e-01 1.06608021e+00
8.03768158e-01 -2.17680186e-01 -1.08947134e+00 3.94139320e-01
4.36848365e-02 -2.87970901e-01 2.10188761e-01 -7.95278847e-01
-8.56867313e-01 9.03940856e-01 1.70725137e-01 1.04000485e+00
1.08143687e+00 2.67922819e-01 8.07044879e-02 4.07633722e-01
5.56615829e-01 -6.49682283e-01 1.99322894e-01 3.00006211e-01
8.06939185e-01 -6.73252463e-01 -4.79240678e-02 -3.34019214e-01
-1.13696761e-01 1.19240379e+00 2.69357413e-01 -7.03902319e-02
6.86170578e-01 7.28178501e-01 -4.71863627e-01 -6.47556126e-01
-1.36737561e+00 -2.21759185e-01 -1.48524828e-02 4.84250635e-01
9.84786630e-01 5.57803884e-02 -5.66458464e-01 1.25489831e-01
-4.86175209e-01 3.80105823e-02 7.59903848e-01 7.54554272e-01
-5.55212617e-01 -6.59072399e-01 -4.26758081e-01 3.86689812e-01
-3.28297883e-01 -1.96528599e-01 -1.14895022e+00 5.93934536e-01
7.22248182e-02 1.32833457e+00 3.45578790e-01 -6.64476693e-01
-2.48940036e-01 -9.15364027e-02 7.81747639e-01 -1.17796421e-01
-2.82358944e-01 -2.02193424e-01 4.61406000e-02 -4.29791033e-01
-6.78889811e-01 -9.74575996e-01 -1.45327890e+00 -7.26771891e-01
-2.91886210e-01 3.61109003e-02 5.80092967e-01 1.21504927e+00
8.73236451e-03 9.04922485e-01 1.21809542e-01 -4.14581627e-01
3.52775425e-01 -6.48986220e-01 -5.12764573e-01 -5.55861592e-02
3.82186234e-01 -6.38165355e-01 -7.96460807e-01 4.49650705e-01] | [7.949141979217529, 3.517305850982666] |
a4d1968c-bd9d-433b-b49e-cf24e86f1676 | advanced-customer-activity-prediction-based | 1904.07687 | null | https://arxiv.org/abs/1904.07687v4 | https://arxiv.org/pdf/1904.07687v4.pdf | Advanced Customer Activity Prediction based on Deep Hierarchic Encoder-Decoders | Product recommender systems and customer profiling techniques have always been a priority in online retail. Recent machine learning research advances and also wide availability of massive parallel numerical computing has enabled various approaches and directions of recommender systems advancement. Worth to mention is the fact that in past years multiple traditional "offline" retail business are gearing more and more towards employing inferential and even predictive analytics both to stock-related problems such as predictive replenishment but also to enrich customer interaction experience. One of the most important areas of recommender systems research and development is that of Deep Learning based models which employ representational learning to model consumer behavioral patterns. Current state of the art in Deep Learning based recommender systems uses multiple approaches ranging from already classical methods such as the ones based on learning product representation vector, to recurrent analysis of customer transactional time-series and up to generative models based on adversarial training. Each of these methods has multiple advantages and inherent weaknesses such as inability of understanding the actual user-journey, ability to propose only single product recommendation or top-k product recommendations without prediction of actual next-best-offer. In our work we will present a new and innovative architectural approach of applying state-of-the-art hierarchical multi-module encoder-decoder architecture in order to solve several of current state-of-the-art recommender systems issues. Our approach will also produce by-products such as product need-based segmentation and customer behavioral segmentation - all in an end-to-end trainable approach. Finally, we will present a couple methods that solve known retail & distribution pain-points based on the proposed architecture. | ['Laurentiu Piciu', 'Andrei Damian', 'Sergiu Turlea', 'Nicolae Tapus'] | 2019-04-11 | null | null | null | null | ['activity-prediction', 'product-recommendation', 'activity-prediction'] | ['computer-vision', 'miscellaneous', 'time-series'] | [ 3.70702855e-02 -1.16286930e-02 -1.86983332e-01 -7.17437565e-01
-6.38876557e-01 -4.15835619e-01 3.55457217e-01 1.86738089e-01
1.87491323e-03 1.80362865e-01 1.31393313e-01 -4.93749797e-01
-5.14481425e-01 -8.16616833e-01 -4.64085639e-01 -3.97236735e-01
-1.88744500e-01 9.93156135e-01 -3.86231124e-01 -1.03389633e+00
4.85772043e-01 4.87667173e-01 -1.82025027e+00 6.76680684e-01
4.64462191e-01 1.45400691e+00 7.65382797e-02 6.71694636e-01
-1.67776853e-01 8.11534226e-01 -2.63946831e-01 -9.18820500e-01
6.00434005e-01 -1.86440066e-01 -4.27240700e-01 -1.13456085e-01
5.86220846e-02 -3.02558571e-01 -1.50773093e-01 5.71260333e-01
5.88078022e-01 5.36790490e-01 7.22735047e-01 -9.26535785e-01
-1.13817656e+00 1.10230982e+00 -4.08549219e-01 3.21637899e-01
2.98946381e-01 -8.94986540e-02 1.18733633e+00 -6.13081217e-01
1.57689780e-01 8.84362340e-01 9.55961049e-01 2.93655217e-01
-1.37938118e+00 -4.87593740e-01 2.75235832e-01 2.96890318e-01
-1.11191058e+00 -7.04569072e-02 7.97620952e-01 -5.47122955e-01
1.41308343e+00 2.79287398e-01 6.52623177e-01 1.08456635e+00
3.11353832e-01 9.70673561e-01 8.87319744e-01 -6.74677119e-02
2.10973710e-01 5.65379143e-01 1.98438853e-01 1.71248630e-01
-2.32936174e-01 4.73654658e-01 -9.32073966e-02 9.59206745e-02
6.16571784e-01 5.32685280e-01 5.63507378e-01 -2.74267746e-03
-4.14848566e-01 1.45701814e+00 2.68335521e-01 2.93743700e-01
-7.32924104e-01 -5.04985452e-02 6.00081265e-01 4.68487173e-01
3.57862204e-01 5.08499444e-01 -8.07046771e-01 -2.38941744e-01
-1.28913188e+00 5.11053085e-01 9.56656098e-01 9.21812654e-01
4.45911229e-01 5.76205432e-01 2.39947051e-01 7.59118199e-01
5.78808248e-01 1.22792244e-01 9.76905227e-01 -4.38939959e-01
1.39354274e-01 3.48727375e-01 -2.72942930e-02 -8.44214141e-01
-6.12711072e-01 -8.25256050e-01 -7.75982738e-01 1.79729313e-01
-4.43517976e-02 -9.79506522e-02 -6.98299766e-01 9.49832916e-01
3.94392898e-03 1.75419867e-01 8.02739710e-02 8.44692230e-01
5.82382321e-01 7.80391932e-01 8.69899523e-03 -1.12497494e-01
1.23205972e+00 -8.88578296e-01 -3.36791605e-01 1.10917941e-01
5.47062218e-01 -1.01414967e+00 7.85618424e-01 8.90534043e-01
-1.00036979e+00 -9.66243148e-01 -1.11407447e+00 8.71735811e-02
-6.80515051e-01 1.78517401e-02 1.09504187e+00 1.06793237e+00
-8.03014636e-01 9.79622841e-01 -4.14155096e-01 -1.71774805e-01
1.91307276e-01 9.30109680e-01 1.95685357e-01 2.67636031e-01
-1.19818091e+00 1.02910495e+00 1.38568327e-01 2.78464258e-01
-8.01619411e-01 -7.52029836e-01 -6.00228786e-01 2.61169761e-01
8.96376967e-02 -5.72788358e-01 1.51973724e+00 -1.32121718e+00
-1.95548248e+00 2.89946169e-01 4.83017743e-01 -1.15930176e+00
1.44547746e-01 -4.55888361e-01 -9.85754669e-01 -6.36754692e-01
-3.70830983e-01 2.39649296e-01 7.12946057e-01 -9.45588887e-01
-8.51535201e-01 -4.14202601e-01 -1.19636871e-01 4.20979261e-02
2.09599018e-01 -5.36731891e-02 3.37488711e-01 -6.17805600e-01
-2.41068125e-01 -1.02433646e+00 -7.32640922e-01 -1.08457160e+00
-2.92757988e-01 -5.07959314e-02 4.31484222e-01 -5.81504703e-01
1.15088594e+00 -1.84952033e+00 -2.84817517e-02 3.12181175e-01
-3.19933832e-01 3.59051615e-01 -5.94919100e-02 8.46467197e-01
-2.21185833e-01 -2.09910184e-01 4.36339319e-01 -3.71445477e-01
3.79013598e-01 1.20517612e-01 -7.49464989e-01 3.40666533e-01
-6.00941759e-03 1.00904155e+00 -6.12141788e-01 1.12500668e-01
6.23942852e-01 7.44506836e-01 -6.31551743e-01 1.33606941e-01
-3.05634260e-01 2.76151955e-01 -2.84328282e-01 6.50691867e-01
5.41358173e-01 1.14906475e-01 7.46073723e-02 -2.34464288e-01
-2.33428776e-02 4.21268135e-01 -1.27377689e+00 1.46094060e+00
-8.87302995e-01 2.01060027e-01 -4.60718244e-01 -1.29526210e+00
1.22225010e+00 1.83518559e-01 7.16327429e-01 -9.66078281e-01
4.71890986e-01 2.57809073e-01 1.62871405e-01 -3.47775251e-01
1.20278895e+00 -4.38287497e-01 -1.12129614e-01 3.52895886e-01
2.28514284e-01 4.62788761e-01 -1.62069008e-01 -2.02506155e-01
5.76321244e-01 4.25640732e-01 1.04124524e-01 8.16172957e-02
4.00152296e-01 -2.34629422e-01 1.64656788e-01 4.89441097e-01
2.33770967e-01 5.23253560e-01 -6.31919131e-03 -5.92533410e-01
-1.15148997e+00 -8.91559243e-01 -1.24451064e-01 1.48970628e+00
-1.97072834e-01 -5.11943996e-02 -3.33063275e-01 -3.94931108e-01
1.97670251e-01 1.25319827e+00 -7.15016961e-01 -8.83855894e-02
-5.50288618e-01 -6.95164382e-01 3.19854796e-01 7.34886348e-01
-1.67212129e-01 -1.29646420e+00 -4.68690634e-01 7.38357723e-01
5.51458776e-01 -5.15591443e-01 -2.68218935e-01 5.08583546e-01
-1.17164969e+00 -5.38766444e-01 -5.42798877e-01 -6.33424580e-01
2.28840299e-02 -1.07287124e-01 1.32846069e+00 -4.95743930e-01
-2.37092331e-01 1.06804989e-01 -5.74299634e-01 -3.37454975e-01
-5.07707596e-01 3.54129642e-01 9.45221037e-02 2.66658962e-01
8.34727943e-01 -6.67550147e-01 -7.97316134e-01 3.25783908e-01
-6.54359996e-01 -4.76020008e-01 9.01168048e-01 7.71662056e-01
7.30966389e-01 1.09120727e-01 1.03723240e+00 -1.35098243e+00
8.40169549e-01 -9.85078156e-01 -5.69592893e-01 -1.49852127e-01
-1.35647082e+00 5.53030744e-02 8.42675209e-01 -6.12000525e-01
-9.80320930e-01 8.10515136e-02 -8.49393666e-01 -3.87912601e-01
-2.13110685e-01 7.00285196e-01 2.86978096e-01 3.84445429e-01
6.24313533e-01 3.97329330e-01 -1.04193367e-01 -6.86445832e-01
8.57140362e-01 7.57989287e-01 2.63434738e-01 -1.00325346e-01
2.65237749e-01 -7.65427481e-03 -1.41174987e-01 -3.93264472e-01
-5.82186639e-01 -8.01073253e-01 -5.12251198e-01 -1.19420268e-01
4.82789040e-01 -6.27139807e-01 -1.22930956e+00 -5.33662364e-02
-5.28830767e-01 -3.32502611e-02 -7.78797150e-01 4.06433284e-01
-7.34959602e-01 1.50716156e-02 -9.20868218e-01 -1.19915617e+00
-8.33219707e-01 -1.18310654e+00 8.61250877e-01 2.73818970e-01
-3.30630362e-01 -1.01145232e+00 2.71253496e-01 5.78355610e-01
8.12995315e-01 -9.46923792e-02 8.00685525e-01 -1.30057430e+00
-2.93462753e-01 -5.97705483e-01 3.25401634e-01 5.04177570e-01
-2.97894418e-01 -2.92785168e-01 -7.84354389e-01 -1.34301707e-01
1.25001028e-01 8.76519606e-02 5.20299375e-01 6.44966304e-01
5.67924380e-01 -2.73787647e-01 -1.56292617e-01 4.10654396e-01
1.71222234e+00 6.74936116e-01 9.68978524e-01 2.87662208e-01
4.94932801e-01 6.89161420e-01 8.91566277e-01 4.96452510e-01
3.72257888e-01 8.37052345e-01 5.20838857e-01 1.50494367e-01
3.90287310e-01 -3.39025170e-01 4.75976378e-01 8.44655991e-01
-2.25624487e-01 -1.56948090e-01 -1.12514853e-01 3.22676361e-01
-1.96794355e+00 -1.37925935e+00 -1.49741441e-01 2.36227155e+00
2.07891136e-01 2.25086734e-01 6.02870584e-01 2.00469583e-01
2.27694586e-01 -3.27624083e-01 -6.25221193e-01 -1.39387631e+00
2.33848929e-01 5.07585645e-01 8.25288773e-01 2.15261176e-01
-1.00257576e+00 7.53653884e-01 6.08837652e+00 7.17076242e-01
-1.14709222e+00 3.30495536e-01 6.68503821e-01 -3.15918773e-01
-3.63685638e-01 -2.85659730e-01 -1.09914863e+00 5.10145724e-01
1.66177976e+00 1.43472433e-01 6.68906510e-01 1.38191164e+00
1.39334351e-01 2.84124702e-01 -1.15208912e+00 9.70517635e-01
1.86683998e-01 -1.38586164e+00 -1.23439714e-01 4.41132933e-01
6.73326969e-01 1.96363643e-01 6.95870042e-01 8.88286591e-01
4.40171719e-01 -1.23343563e+00 5.98701298e-01 5.70366621e-01
1.92234159e-01 -1.14904249e+00 1.12410963e+00 1.22297719e-01
-1.14098275e+00 -4.46429193e-01 -3.97163957e-01 -5.18741012e-02
5.64033628e-01 2.70203590e-01 -7.78799653e-01 5.69433212e-01
5.03713965e-01 5.52681923e-01 -2.13710405e-02 9.05796409e-01
4.87748921e-01 5.74507594e-01 -1.89053461e-01 -1.76068857e-01
4.80291754e-01 -5.92595696e-01 -7.42602348e-02 1.32055402e+00
5.39978385e-01 -1.02529265e-02 -1.02389559e-01 6.78488433e-01
2.80871719e-01 4.21248525e-01 -5.15632510e-01 -8.29932243e-02
-1.06701314e-01 1.39927220e+00 -6.30912960e-01 -2.09938660e-02
-6.49709821e-01 9.25473034e-01 -1.40379190e-01 -6.91150129e-02
-8.85153174e-01 -6.37584254e-02 7.95187294e-01 5.36200404e-01
1.00354445e+00 6.83128610e-02 -1.93966433e-01 -7.45578110e-01
-5.90431452e-01 -1.01094472e+00 2.84088612e-01 -3.68484437e-01
-1.57056260e+00 8.24067295e-01 -3.84427547e-01 -1.35081029e+00
-8.31724346e-01 -6.29646659e-01 -2.84366190e-01 8.40905368e-01
-1.20378757e+00 -1.37509108e+00 4.00034904e-01 5.38473845e-01
9.12538707e-01 -6.90625072e-01 9.34018910e-01 7.71937609e-01
-1.93873659e-01 8.38749051e-01 5.08451700e-01 -2.61138767e-01
2.52772957e-01 -1.24927247e+00 4.28650796e-01 3.75205815e-01
4.18969542e-01 6.79711699e-01 9.17011619e-01 -4.11110878e-01
-1.63150191e+00 -8.90217960e-01 6.97508514e-01 -3.93925428e-01
7.17544854e-01 -2.90682971e-01 -6.52341723e-01 8.51040244e-01
2.76809067e-01 -5.77039838e-01 1.26110244e+00 6.89556599e-01
-1.88935310e-01 -4.98562992e-01 -1.18323326e+00 2.60296077e-01
3.53189975e-01 -3.01961094e-01 -3.63873541e-01 2.69792199e-01
4.11684811e-01 -2.35835478e-01 -1.21345460e+00 -1.17517158e-01
7.73339510e-01 -1.26607668e+00 1.03474271e+00 -6.63044453e-01
4.37874705e-01 7.27131367e-02 -3.60487521e-01 -1.25135708e+00
-5.07869363e-01 -8.95520210e-01 -3.00312847e-01 1.26448476e+00
7.04680622e-01 -5.06619751e-01 1.11549866e+00 4.80346113e-01
-4.08079565e-01 -1.19218349e+00 -5.25533319e-01 -3.59643549e-01
-2.13583186e-03 -6.97253406e-01 7.63466954e-01 6.19940579e-01
-2.16350593e-02 7.78622031e-01 -7.93409646e-01 -1.48848027e-01
1.86307400e-01 4.70256329e-01 5.80387890e-01 -1.31462514e+00
-9.69464481e-01 -6.55644536e-01 -5.63462079e-01 -1.08153808e+00
-2.84502447e-01 -8.17687213e-01 -2.85705507e-01 -1.33945727e+00
-3.57360899e-01 -4.72257614e-01 -6.99075401e-01 -1.80503651e-01
6.54874027e-01 2.48106048e-01 9.51169357e-02 1.31402910e-02
-3.25112969e-01 1.98677287e-01 8.74091148e-01 -7.60942549e-02
-4.57957119e-01 7.22009003e-01 -1.03240061e+00 2.51691192e-01
7.17619777e-01 -3.33869517e-01 -6.36106133e-01 2.70605057e-01
7.46287227e-01 2.03253940e-01 -1.08050719e-01 -6.44994080e-01
4.51486893e-02 2.20663741e-01 2.92015135e-01 -8.01231384e-01
5.10151744e-01 -9.03700709e-01 5.63104987e-01 3.47678483e-01
-3.79785717e-01 3.50855589e-01 -1.23268683e-02 6.67555153e-01
-1.91608697e-01 -5.17375886e-01 4.37395662e-01 -1.47232920e-01
-8.31516564e-01 2.83444792e-01 -4.94003773e-01 -5.47342837e-01
1.02468276e+00 -4.01594728e-01 1.47151381e-01 -4.66478050e-01
-1.04614210e+00 -1.56947702e-01 -1.05172113e-01 7.44163096e-01
4.66845989e-01 -1.03031445e+00 -6.78148925e-01 1.62948102e-01
-1.42445132e-01 -7.58086026e-01 5.75057924e-01 6.84360385e-01
-4.98663247e-01 7.79686451e-01 -3.40270907e-01 -4.15922478e-02
-8.00730348e-01 1.18593156e+00 3.38788182e-01 -6.08926177e-01
-5.71916282e-01 8.57652485e-01 -2.58564919e-01 -3.11937511e-01
5.72368726e-02 -3.00549805e-01 -6.80584550e-01 2.91656017e-01
3.13865006e-01 4.81811047e-01 3.83476198e-01 -9.67027068e-01
-2.70993616e-02 3.21885526e-01 -5.85282028e-01 2.15763479e-01
1.44768763e+00 -1.40211254e-01 6.13090873e-01 5.25462747e-01
1.12060618e+00 -3.72191370e-01 -9.13119435e-01 3.59973870e-02
-4.33174148e-02 -1.34481534e-01 3.37777823e-01 -1.07505250e+00
-1.43014419e+00 9.11829114e-01 1.16943800e+00 8.00265491e-01
1.14004242e+00 -2.61604816e-01 1.26899505e+00 1.43093258e-01
4.00708675e-01 -1.25050008e+00 -3.83813947e-01 1.13779962e-01
5.41684091e-01 -1.26272309e+00 -6.66120946e-02 1.05572104e-01
-1.09654307e+00 1.07017922e+00 6.64710626e-02 -5.84607422e-01
1.02285504e+00 3.22675645e-01 1.55145943e-01 -1.63852021e-01
-6.57150686e-01 -3.20067972e-01 2.09700868e-01 7.05652356e-01
7.01669097e-01 3.58184785e-01 -1.82874352e-01 1.29179931e+00
-4.40708458e-01 1.17284201e-01 2.00583458e-01 4.69080508e-01
-3.23600997e-03 -1.43681169e+00 6.98294416e-02 8.93034577e-01
-8.31702590e-01 -2.44531587e-01 3.36156666e-01 5.95181584e-01
1.99773699e-01 9.55249310e-01 1.22840844e-01 -9.14692998e-01
4.79084313e-01 -1.66988477e-01 4.32849795e-01 -5.37883461e-01
-1.51664639e+00 2.64310032e-01 1.75815001e-01 -4.44617569e-01
-7.97295868e-02 -1.01012254e+00 -1.00094461e+00 -5.83785117e-01
-5.79802871e-01 1.37458012e-01 1.06498837e+00 9.34140384e-01
4.94713932e-01 6.37617648e-01 6.49489224e-01 -1.21386206e+00
-9.35155571e-01 -1.06212258e+00 -1.14063656e+00 2.92447835e-01
-1.65287435e-01 -5.25034785e-01 2.11327866e-01 -2.10739486e-02] | [9.924210548400879, 5.8663716316223145] |
383fcdfc-0177-4b83-aa3f-b6c94074c409 | an-end-to-end-review-of-gaze-estimation-and | 2307.00122 | null | https://arxiv.org/abs/2307.00122v1 | https://arxiv.org/pdf/2307.00122v1.pdf | An End-to-End Review of Gaze Estimation and its Interactive Applications on Handheld Mobile Devices | In recent years we have witnessed an increasing number of interactive systems on handheld mobile devices which utilise gaze as a single or complementary interaction modality. This trend is driven by the enhanced computational power of these devices, higher resolution and capacity of their cameras, and improved gaze estimation accuracy obtained from advanced machine learning techniques, especially in deep learning. As the literature is fast progressing, there is a pressing need to review the state of the art, delineate the boundary, and identify the key research challenges and opportunities in gaze estimation and interaction. This paper aims to serve this purpose by presenting an end-to-end holistic view in this area, from gaze capturing sensors, to gaze estimation workflows, to deep learning techniques, and to gaze interactive applications. | ['Juan Ye', 'Mohamed Khamis', 'Shijing He', 'Yaxiong Lei'] | 2023-06-30 | null | null | null | null | ['gaze-estimation'] | ['computer-vision'] | [ 3.26641589e-01 5.92880696e-03 -3.79460216e-01 -2.83484161e-01
-1.72941655e-01 -3.31238538e-01 3.11773121e-01 -3.89271796e-01
-3.89515996e-01 5.51252961e-01 -7.54644349e-02 -2.04770073e-01
-1.41045690e-01 -2.55429116e-03 -1.27317369e-01 -5.48902810e-01
1.70767412e-01 -6.08260408e-02 -2.15098247e-01 6.50534332e-02
7.64344454e-01 1.81121945e-01 -2.22666883e+00 5.77507541e-02
8.20304632e-01 1.16583896e+00 1.76431671e-01 9.16352451e-01
-1.44259378e-01 5.24256170e-01 -3.38981688e-01 -6.43155932e-01
-3.23274672e-01 -3.01395327e-01 -7.76775479e-01 -2.59293050e-01
7.82810569e-01 -5.04675329e-01 3.48644465e-01 7.82034039e-01
7.32733130e-01 2.94878427e-02 2.92017281e-01 -1.67633712e+00
-5.46721697e-01 -2.81032145e-01 -9.32512879e-01 5.11834681e-01
7.88430333e-01 1.86173767e-01 7.35585392e-01 -7.54959583e-01
2.14377165e-01 9.17182565e-01 5.62281013e-01 8.71208370e-01
-7.12067068e-01 -8.69703710e-01 4.06107940e-02 4.13111895e-01
-1.15378571e+00 -7.65094280e-01 6.61721110e-01 -5.97033441e-01
1.12608886e+00 2.62671977e-01 7.08493590e-01 1.11567688e+00
1.36855185e-01 9.06930864e-01 1.19292915e+00 -9.38896239e-01
-9.74446759e-02 4.02472198e-01 9.82343033e-02 6.78658068e-01
4.83037792e-02 -1.06753193e-01 -1.30897784e+00 2.54268229e-01
4.98232096e-01 3.39214146e-01 -3.57359260e-01 -2.87774205e-01
-7.73322880e-01 5.09820163e-01 3.70250821e-01 1.73499957e-01
-3.64535540e-01 -1.23519041e-01 1.32109970e-01 -1.32752195e-01
5.99256575e-01 2.14735419e-01 -3.97261769e-01 -9.86321628e-01
-1.17642605e+00 -3.42550464e-02 7.77525187e-01 7.35328734e-01
6.09575093e-01 -4.65437382e-01 1.85822725e-01 4.56075698e-01
9.21412110e-01 7.09953070e-01 3.23025048e-01 -9.90695357e-01
3.53991777e-01 6.40317261e-01 9.01722535e-02 -7.21628428e-01
-4.02063668e-01 1.02538370e-01 -3.71422857e-01 7.40056694e-01
5.99847138e-01 -3.09124857e-01 -4.67129737e-01 1.47553980e+00
2.90568620e-01 -3.36325988e-02 -5.80497384e-01 7.30181515e-01
7.04284370e-01 1.29375458e-01 2.72606969e-01 -2.24444330e-01
1.42121923e+00 -8.75457466e-01 -1.06840360e+00 -1.09788932e-01
6.12624705e-01 -6.16489112e-01 1.51111364e+00 6.46724164e-01
-1.24600530e+00 -4.62380260e-01 -1.03732073e+00 -4.78539258e-01
-4.10207987e-01 5.70246950e-02 6.16915762e-01 1.41308093e+00
-1.29673505e+00 2.11068839e-01 -8.98994088e-01 -7.03436852e-01
8.49332273e-01 9.73304272e-01 -1.27241269e-01 3.09513390e-01
-6.09806538e-01 9.08711791e-01 -1.22464612e-01 1.30186975e-01
3.71228755e-01 -8.16672325e-01 -7.08345950e-01 2.54361182e-01
3.61896157e-02 -7.07413435e-01 1.45762479e+00 -1.25418150e+00
-1.81531775e+00 1.16364026e+00 -8.20514023e-01 -3.50392535e-02
2.52114087e-01 -7.50293672e-01 -3.19312304e-01 -1.20192811e-01
-4.86311555e-01 7.43508756e-01 8.98431361e-01 -7.80391932e-01
-1.23112357e+00 -7.93049991e-01 1.42421141e-01 3.97305787e-01
-7.05725133e-01 4.29223716e-01 -2.25364476e-01 3.35306793e-01
-4.51124489e-01 -1.02620637e+00 5.94422877e-01 1.06899567e-01
-1.43250469e-02 -4.97596532e-01 1.20917869e+00 -3.45042199e-01
1.68482494e+00 -2.20666909e+00 2.19687223e-02 -1.68146230e-02
8.93272758e-01 4.13623422e-01 6.02840602e-01 5.19224368e-02
-5.68697453e-02 7.15252683e-02 3.44514668e-01 -7.35716045e-01
-4.24633548e-02 -3.67111146e-01 -1.47554800e-02 2.17276767e-01
-1.19487599e-01 1.15942609e+00 -8.24126124e-01 -4.30922657e-01
5.23103297e-01 7.59607971e-01 -2.06853718e-01 7.08385408e-02
3.50720286e-01 5.73956072e-01 -1.48819983e-01 9.02229548e-01
5.01019835e-01 -7.32745051e-01 -1.18149407e-01 4.86214384e-02
-4.49406594e-01 2.12357119e-01 -7.09880114e-01 1.55718505e+00
-4.47453678e-01 1.35062778e+00 -7.97878951e-02 -1.53075173e-01
4.75530356e-01 8.97795558e-02 4.23768401e-01 -8.43063474e-01
3.55974585e-01 7.82869384e-02 5.69811575e-02 -9.08570528e-01
5.13808370e-01 3.39076221e-01 6.02204502e-01 7.88367391e-01
2.84966920e-02 5.70145190e-01 -2.62677938e-01 -3.68885279e-01
5.91872871e-01 4.42702025e-01 4.58245993e-01 -5.73890395e-02
5.99672735e-01 -4.49861735e-01 -4.00706470e-01 3.35859776e-01
-6.32085025e-01 5.50293028e-01 2.52926230e-01 -4.64572996e-01
-6.17145360e-01 -6.74204946e-01 -1.06689222e-01 1.58362305e+00
-3.39990780e-02 -2.81797230e-01 -1.22572279e+00 -5.67682207e-01
-3.06611031e-01 3.05426598e-01 -7.26782978e-01 1.68933898e-01
-4.17196363e-01 -3.90189499e-01 1.68051884e-01 4.49145705e-01
5.58944583e-01 -1.23844087e+00 -1.30200851e+00 -4.89873052e-01
4.04076762e-02 -7.99487352e-01 -2.24948913e-01 -2.46108383e-01
-1.02157760e+00 -1.13701677e+00 -9.25625026e-01 -4.85351354e-01
3.75394672e-01 5.05902708e-01 1.19245195e+00 1.33560807e-01
6.24740161e-02 5.91484010e-01 -2.14671403e-01 -8.60675931e-01
4.21200246e-01 6.68261826e-01 5.48177287e-02 -6.25287071e-02
1.27908456e+00 -3.69325399e-01 -8.07546556e-01 -3.24390233e-02
-4.31964546e-01 5.92764933e-03 4.28105444e-01 4.73862499e-01
-6.33453354e-02 -6.08370543e-01 2.09954381e-01 -8.29088867e-01
7.73096561e-01 -5.56348503e-01 -3.90870512e-01 2.31719017e-01
-1.10671294e+00 -2.38407925e-01 -2.89630473e-01 -2.52480060e-01
-1.27498674e+00 -2.28366271e-01 -9.26877856e-02 -1.51600420e-01
-5.26529670e-01 3.51316452e-01 -4.58746068e-02 -3.24855030e-01
9.24826920e-01 -3.15539271e-01 2.47201279e-01 -2.14963317e-01
-9.54725593e-03 1.37032175e+00 1.16426744e-01 2.06707721e-03
2.74027854e-01 4.38796997e-01 -6.15531094e-02 -9.30741251e-01
-8.08783531e-01 -4.46851909e-01 -9.31097388e-01 -7.15821803e-01
8.03602815e-01 -5.67439198e-01 -1.49317443e+00 8.68859828e-01
-8.49427104e-01 -2.41245851e-01 2.00095683e-01 3.79962593e-01
-3.57167393e-01 2.69568544e-02 -9.05251652e-02 -1.33132279e+00
-6.81857824e-01 -1.12018526e+00 1.16456413e+00 1.01251125e+00
-7.60363162e-01 -1.21337509e+00 1.21130988e-01 5.68818450e-01
6.53108358e-01 2.64567975e-02 4.92623687e-01 -1.71162128e-01
-4.50346529e-01 -2.58304685e-01 -4.09151912e-01 1.06935631e-02
3.77629578e-01 2.09358096e-01 -1.63196516e+00 -8.22898299e-02
-2.32848451e-02 -2.20401332e-01 1.18505351e-01 7.73157001e-01
1.01262724e+00 1.97043419e-01 -6.44676030e-01 7.02159643e-01
1.06519425e+00 2.47415662e-01 7.85860121e-01 5.64537585e-01
9.00846362e-01 7.72882879e-01 4.33029175e-01 1.56911150e-01
6.87425852e-01 5.50024867e-01 4.86838549e-01 2.65796334e-02
5.18998317e-02 4.29327488e-02 6.49992749e-02 3.76633316e-01
-6.56174541e-01 -2.73057610e-01 -1.21344411e+00 1.53448477e-01
-1.69121099e+00 -9.77090180e-01 -3.38724226e-01 2.31841111e+00
4.41473752e-01 3.16395946e-02 5.13555944e-01 2.56706059e-01
6.06924415e-01 -9.37283114e-02 -7.26157427e-01 -5.19720018e-01
4.29114938e-01 3.53868097e-01 -1.28541356e-02 1.59569263e-01
-8.92272949e-01 5.49989700e-01 7.20096350e+00 2.35343382e-01
-1.66014898e+00 1.85834408e-01 5.11388481e-01 -5.41996241e-01
2.59507298e-01 -6.30222738e-01 -9.00947392e-01 7.41625071e-01
1.21848094e+00 1.24406248e-01 5.08316338e-01 8.53864253e-01
1.53328404e-01 -6.06098950e-01 -1.19098926e+00 1.65919352e+00
4.82690275e-01 -9.88900602e-01 -7.30075240e-01 5.39254010e-01
3.88615370e-01 2.11707547e-01 7.06889987e-01 7.95429125e-02
-5.94898462e-01 -1.19920468e+00 2.13959396e-01 7.16975987e-01
1.19931591e+00 -5.77673078e-01 7.05288589e-01 1.47779480e-01
-6.83276355e-01 -1.93088546e-01 4.93875831e-01 -7.24573791e-01
5.55802956e-02 -2.00401843e-01 -5.43534815e-01 -1.57461002e-01
1.28128636e+00 7.64911056e-01 -5.67247212e-01 1.17458820e+00
-5.98805659e-02 4.09246475e-01 -2.41469979e-01 -3.28611255e-01
-1.83694556e-01 -8.49685967e-02 2.39077322e-02 8.74937117e-01
2.81289726e-01 -5.17986752e-02 -8.40929389e-01 5.30619264e-01
-1.30545586e-01 -3.41165751e-01 -5.01228571e-01 1.41412690e-01
4.23893720e-01 1.42498374e+00 -6.04066968e-01 9.38720927e-02
-8.94712329e-01 9.10305202e-01 1.80630565e-01 4.67413187e-01
-6.45357013e-01 -4.09441173e-01 9.98634100e-01 2.72405773e-01
-1.91534713e-01 -2.20395774e-01 -7.61050224e-01 -8.89430583e-01
2.77543254e-02 -8.16790223e-01 1.76759176e-02 -1.19388950e+00
-8.37218761e-01 3.98683459e-01 -9.27148536e-02 -9.88424957e-01
-5.70549726e-01 -8.68711233e-01 -4.31528807e-01 1.23333085e+00
-1.58995330e+00 -1.07637620e+00 -8.81661296e-01 5.51832616e-01
4.77158606e-01 -8.97727460e-02 8.95284057e-01 2.97860771e-01
-5.52433372e-01 8.27578902e-01 7.87280723e-02 -3.24126691e-01
7.65243411e-01 -1.19692993e+00 2.78734326e-01 3.78724366e-01
-9.95750651e-02 9.48841214e-01 4.29655969e-01 -7.57572576e-02
-1.35097396e+00 7.40303099e-03 1.02825642e+00 -1.26627600e+00
3.32730711e-01 -4.14129347e-01 -7.13340282e-01 6.71396017e-01
5.61338186e-01 -3.44020724e-01 1.27178943e+00 8.56479466e-01
6.51671961e-02 -2.00333651e-02 -1.20432651e+00 6.87640429e-01
8.71962011e-01 -8.62350583e-01 -2.62172222e-01 -2.37167761e-01
-9.51880217e-02 -7.18131661e-01 -4.05933350e-01 7.18784332e-02
1.38881218e+00 -1.27655041e+00 6.76061690e-01 -3.05199295e-01
3.91266614e-01 1.38123274e-01 5.25921822e-01 -7.84744143e-01
-7.59662241e-02 -8.20475280e-01 -8.85117173e-01 1.17447340e+00
1.37099594e-01 -6.06632829e-01 1.26583648e+00 1.39267039e+00
3.40288639e-01 -9.28733289e-01 -6.08290315e-01 1.27584338e-01
-3.58013421e-01 -3.12701941e-01 5.04981577e-01 7.73929894e-01
3.54448825e-01 4.56213802e-01 -2.62570828e-01 -3.25549275e-01
3.96819949e-01 -4.03471082e-01 9.28569376e-01 -1.78113484e+00
3.91794026e-01 -7.73359179e-01 -4.01931852e-01 -1.13818395e+00
-9.58122760e-02 8.10349286e-02 -3.91025931e-01 -1.17011881e+00
1.16984650e-01 -1.65481418e-02 -3.53103310e-01 2.36786649e-01
-4.45520908e-01 6.18201077e-01 1.41333058e-01 3.72940302e-01
-7.85054266e-01 -1.03668775e-02 9.22818840e-01 4.31742996e-01
-4.88766253e-01 3.70030403e-01 -9.32730556e-01 8.74401391e-01
7.49170065e-01 -3.11890524e-02 -4.95950401e-01 -5.62951803e-01
9.06222999e-01 -5.40006697e-01 1.70045093e-01 -1.01130712e+00
5.48023880e-01 1.68259099e-01 6.07206762e-01 -4.54503357e-01
4.41515446e-01 -8.90739799e-01 -3.01459521e-01 -4.43912111e-02
-1.67221531e-01 2.88979173e-01 3.22558731e-01 1.80864170e-01
5.38002923e-02 -1.62263915e-01 4.16161358e-01 2.66533315e-01
-7.11973488e-01 -2.59456970e-03 -1.28577247e-01 -1.44301414e-01
8.67154360e-01 -1.05723667e+00 -3.59429866e-01 -3.66091609e-01
-5.62672317e-01 3.05572785e-02 5.93513966e-01 6.66144252e-01
3.43991488e-01 -8.38181019e-01 5.82837351e-02 3.87690008e-01
8.30808729e-02 -1.79018840e-01 1.92167103e-01 1.14686048e+00
-2.99204588e-01 6.83317184e-01 -5.40197432e-01 -8.83923590e-01
-1.75072193e+00 2.64579177e-01 2.49869391e-01 2.81822234e-01
-6.07905500e-02 1.02106261e+00 -5.27966358e-02 8.05360675e-02
5.11464477e-01 -1.80594787e-01 -7.04254866e-01 2.60592997e-01
9.92461145e-01 8.39728475e-01 2.04684347e-01 -6.46188080e-01
-3.50335091e-01 7.12701023e-01 5.09170331e-02 4.79287617e-02
9.77755308e-01 -7.77063668e-01 2.39354447e-02 7.52373815e-01
9.51705813e-01 -1.38269052e-01 -1.40181029e+00 4.28554378e-02
6.35287985e-02 -6.06936812e-01 3.12573612e-01 -8.57985139e-01
-8.13696980e-01 1.27893078e+00 1.23713732e+00 4.80964601e-01
1.30535710e+00 -1.41287893e-01 6.30096734e-01 1.97523251e-01
1.83978707e-01 -9.85023737e-01 -1.23733975e-01 4.99080420e-01
3.75982016e-01 -1.66666019e+00 -8.49493518e-02 -1.28247216e-01
-5.25245488e-01 1.09174776e+00 7.10676670e-01 3.96383673e-01
8.90521348e-01 1.58998340e-01 2.25310802e-01 -3.57084960e-01
-3.08631986e-01 -3.65332156e-01 5.35636485e-01 1.00701320e+00
9.67199683e-01 -4.96856779e-01 5.88063449e-02 1.41906336e-01
-2.64658302e-01 6.58224523e-01 -2.19537746e-02 9.76514280e-01
-2.31525213e-01 -1.03555703e+00 -3.42481852e-01 6.72506571e-01
-8.11314940e-01 -1.92810133e-01 -2.97787696e-01 7.26610780e-01
1.67091951e-01 1.11107433e+00 1.89101920e-01 -3.57838541e-01
1.24226831e-01 3.24762911e-01 6.13645554e-01 -4.19293016e-01
-6.67018950e-01 -3.52312803e-01 -4.35460061e-01 -6.30704045e-01
-8.43937457e-01 -8.46731842e-01 -6.78359151e-01 -6.37459576e-01
-5.97190917e-01 -3.20415407e-01 1.00391340e+00 1.27975690e+00
6.84852242e-01 4.41629171e-01 1.37425125e-01 -1.42531681e+00
2.18555197e-01 -1.03317320e+00 -3.21718276e-01 -1.02515973e-01
7.78066695e-01 -7.77205944e-01 -1.43449664e-01 3.49513084e-01] | [14.10536003112793, 0.11554199457168579] |
004b9e85-9c7d-4830-af9c-8f554eb1bc43 | detecting-histologic-glioblastoma-regions-of | 2302.00669 | null | https://arxiv.org/abs/2302.00669v2 | https://arxiv.org/pdf/2302.00669v2.pdf | Detecting Histologic & Clinical Glioblastoma Patterns of Prognostic Relevance | Glioblastoma is the most common and aggressive malignant adult tumor of the central nervous system, with a grim prognosis and heterogeneous morphologic and molecular profiles. Since adopting the current standard-of-care treatment 18 years ago, no substantial prognostic improvement has been noticed. Accurate prediction of patient overall survival (OS) from histopathology whole slide images (WSI) integrated with clinical data using advanced computational methods could optimize clinical decision-making and patient management. Here, we focus on identifying prognostically relevant glioblastoma characteristics from H&E stained WSI & clinical data relating to OS. The exact approach for WSI capitalizes on the comprehensive curation of apparent artifactual content and an interpretability mechanism via a weakly supervised attention-based multiple-instance learning algorithm that further utilizes clustering to constrain the search space. The automatically placed pat- terns of high diagnostic value classify each WSI as representative of short or long-survivors. Further assessment of the prognostic relevance of the associated clinical patient data is performed both in isolation and in an integrated manner, using XGBoost and SHapley Additive exPlanations (SHAP). Identifying tumor morphological & clinical patterns associated with short and long OS will enable the clinical neuropathologist to provide additional relevant prognostic information to the treating team and suggest avenues of biological investigation for understanding and potentially treating glioblastoma. | ['Sharath Chandra Guntuku', 'Garv Mehdiratta', 'Sunny Rai', 'Spyridon Bakas', 'MacLean P. Nasrallah', 'Shubham Innani', 'Bhakti Baheti'] | 2023-02-01 | null | null | null | null | ['whole-slide-images', 'multiple-instance-learning'] | ['computer-vision', 'methodology'] | [ 5.19603074e-01 1.29112720e-01 -3.36725086e-01 -2.31298149e-01
-1.12993383e+00 -1.58871099e-01 2.96651810e-01 8.99797320e-01
-6.29364729e-01 1.02004898e+00 5.18299162e-01 -5.86693466e-01
-8.14753830e-01 -3.64298254e-01 7.78454728e-03 -1.33554840e+00
-1.50960684e-01 8.47080827e-01 -3.11631739e-01 -2.87573412e-02
3.94891530e-01 6.91636026e-01 -1.19000101e+00 2.21147805e-01
1.04694426e+00 8.08149457e-01 6.55754566e-01 7.67114103e-01
-1.56297460e-02 6.19917691e-01 -3.13637763e-01 -1.80981178e-02
-3.31410408e-01 -2.67623186e-01 -7.82443821e-01 1.85810439e-02
-2.34900638e-01 8.65480900e-02 1.11178653e-02 7.92259455e-01
6.64754689e-01 -2.36547858e-01 9.46111500e-01 -1.22712016e+00
-1.55823603e-01 5.00433743e-01 -3.34818810e-01 5.46556115e-01
8.22824761e-02 3.88847589e-01 9.61911917e-01 -5.74911177e-01
8.51772785e-01 3.14854890e-01 4.38904643e-01 5.11469841e-01
-1.29680061e+00 -3.40262979e-01 1.03776388e-01 6.19861662e-01
-1.32984936e+00 -2.24216163e-01 3.64694566e-01 -5.51115394e-01
1.10925484e+00 6.00074768e-01 9.16950524e-01 8.39125693e-01
5.07013202e-01 6.91838861e-01 1.10595489e+00 -4.56688046e-01
3.93656492e-01 -1.10936530e-01 5.13807118e-01 8.06336880e-01
2.02101052e-01 -5.65913087e-03 -5.63777268e-01 -2.12043270e-01
-4.03966568e-02 3.96906763e-01 -5.91549575e-01 1.31359145e-01
-1.32463825e+00 6.01525366e-01 4.30609077e-01 3.63755524e-01
-3.64020467e-01 2.37586852e-02 5.30582726e-01 -5.24352379e-02
4.35350835e-01 4.57082987e-01 -6.83144927e-01 4.05714586e-02
-9.43122268e-01 1.63920373e-02 8.86918455e-02 5.50738275e-01
2.51125067e-01 -6.13058686e-01 7.71352574e-02 6.08493388e-01
1.81200653e-01 -9.96509790e-02 1.24058247e+00 -3.68973523e-01
-3.64905335e-02 9.49649036e-01 -2.16434985e-01 -3.75937611e-01
-1.25250900e+00 -6.45918369e-01 -7.37134814e-01 2.43394092e-01
4.57813203e-01 2.06821412e-01 -8.20704281e-01 1.43924713e+00
7.19905552e-03 -4.72440422e-02 1.07350059e-01 6.19531631e-01
6.41204059e-01 -2.73038715e-01 4.07305509e-01 -3.06568742e-01
1.62665820e+00 -5.94680727e-01 -4.18343484e-01 -1.31759346e-01
1.29932082e+00 -3.01915944e-01 8.51883650e-01 1.85207844e-01
-4.43903595e-01 3.56550634e-01 -9.46614444e-01 3.70855778e-02
-5.63847661e-01 1.04096107e-01 8.49011242e-01 4.08721000e-01
-1.04663324e+00 5.98152280e-01 -1.15058696e+00 -8.94709527e-01
9.12467062e-01 6.86743975e-01 -7.38395929e-01 5.88847324e-02
-6.69745922e-01 1.19938231e+00 4.93810326e-01 -6.61674291e-02
-5.58803678e-01 -9.24989104e-01 -4.75648284e-01 -8.98761004e-02
7.58265331e-02 -1.21256292e+00 7.20021904e-01 -7.00811148e-01
-1.05634665e+00 1.00660050e+00 -7.18130112e-01 -4.66207117e-01
1.70484126e-01 6.07464492e-01 -2.17506886e-01 2.49167338e-01
2.04375029e-01 3.33956748e-01 3.23661894e-01 -7.39159405e-01
-9.11273479e-01 -1.02631247e+00 -6.95324421e-01 3.56588304e-01
-2.52386749e-01 1.22344531e-02 2.01834843e-01 -5.45483887e-01
3.51410151e-01 -8.34764540e-01 -6.59178853e-01 -1.45860985e-01
-4.25363630e-01 -6.97561651e-02 4.25561756e-01 -7.60798872e-01
8.31865489e-01 -1.87737358e+00 3.65465909e-01 2.04018787e-01
5.31464338e-01 -2.76421100e-01 2.99055099e-01 1.30080029e-01
-2.95229524e-01 2.99899727e-01 -7.83948153e-02 -3.36031407e-01
-3.07773292e-01 -1.91123992e-01 3.35507542e-01 7.95915186e-01
3.50013971e-01 9.81819808e-01 -1.05242753e+00 -3.97617459e-01
1.33097664e-01 4.99889217e-02 -3.00481498e-01 1.35997273e-02
-9.84319113e-03 5.71359634e-01 -2.90599287e-01 1.02468657e+00
3.63869742e-02 -5.48156023e-01 1.28770575e-01 3.74682136e-02
9.28898603e-02 -9.65998396e-02 -2.47954175e-01 1.40623808e+00
-8.31542313e-02 6.70128942e-01 -3.14168841e-01 -9.61772263e-01
3.91379148e-01 2.41832495e-01 5.87245643e-01 -3.83529305e-01
3.91616791e-01 3.06397766e-01 1.99245691e-01 -8.46828163e-01
2.11323574e-02 -4.68825638e-01 2.35461742e-01 2.95951009e-01
-2.34707613e-02 2.18721002e-01 -4.60172892e-02 1.95283979e-01
1.64940512e+00 -3.00729752e-01 8.58249366e-01 -3.97672683e-01
4.37673002e-01 4.39477473e-01 4.44428027e-01 4.52659935e-01
-5.76420724e-01 6.15991354e-01 4.41720873e-01 -2.13668376e-01
-8.11459720e-01 -8.07201326e-01 -5.71483433e-01 7.77754426e-01
-1.17675923e-01 -6.29107580e-02 -2.83205152e-01 -5.17294824e-01
-5.26405312e-02 8.09572220e-01 -1.07522607e+00 -2.41388112e-01
6.36155009e-02 -1.46766639e+00 4.21045184e-01 6.03354096e-01
-2.22552061e-01 -7.68618703e-01 -7.05175996e-01 3.73167902e-01
-9.74028781e-02 -7.27400661e-01 8.76211151e-02 1.06936955e+00
-8.91825616e-01 -1.40564740e+00 -7.63219953e-01 -6.10594988e-01
1.16307068e+00 1.03553414e-01 4.90863442e-01 3.16633195e-01
-8.59367371e-01 7.89503828e-02 -4.04950351e-01 -6.92218304e-01
-2.85214484e-01 -1.17521338e-01 1.04181739e-02 -2.30216771e-01
6.15853310e-01 -4.94032115e-01 -6.87368393e-01 9.24888253e-03
-5.65417230e-01 3.27350765e-01 7.87798464e-01 1.21470928e+00
7.21144974e-01 2.50759143e-02 4.96323794e-01 -7.11485386e-01
3.87313038e-01 -7.54583299e-01 -1.12712555e-01 1.29260764e-01
-8.39413464e-01 -5.74353850e-03 6.27268255e-01 3.58171947e-02
-7.93588042e-01 -1.33788332e-01 -6.59655873e-03 2.59302080e-01
-6.32145226e-01 8.65054071e-01 -4.49134819e-02 -3.37735601e-02
6.79400921e-01 2.46280193e-01 2.24616200e-01 1.69511288e-01
-2.67623067e-01 8.16335976e-01 3.51012081e-01 -9.47995037e-02
9.80703309e-02 8.34567010e-01 4.03274417e-01 -7.16609776e-01
-7.55804539e-01 -9.32197809e-01 -5.47593057e-01 -1.69466287e-01
8.79043996e-01 -4.60749686e-01 -7.98329353e-01 4.06210154e-01
-7.59160101e-01 -3.80361617e-01 -2.61722580e-02 5.35722792e-01
-8.36668432e-01 9.29556191e-02 -4.59013075e-01 -5.32535493e-01
-3.89942527e-01 -1.49184036e+00 1.02630675e+00 1.92880332e-01
-6.62813246e-01 -1.03013277e+00 6.96318671e-02 6.44960403e-01
2.26570919e-01 2.98346311e-01 1.49840820e+00 -1.03513288e+00
-3.60517442e-01 -4.38315898e-01 -3.03078413e-01 -4.36976552e-01
2.17497349e-01 2.59662550e-02 -1.03039837e+00 -7.38089755e-02
-4.40443933e-01 -1.19901583e-01 8.61703992e-01 6.79252684e-01
9.71688449e-01 7.20069408e-02 -8.63391101e-01 7.49579966e-01
1.63246131e+00 4.55497921e-01 2.42026567e-01 9.54867780e-01
3.17925125e-01 7.16144741e-01 1.72051936e-01 2.84533083e-01
4.70409751e-01 4.34720129e-01 6.64024472e-01 7.30750561e-02
6.83582062e-03 2.45696455e-01 -1.50514975e-01 9.66865793e-02
-2.85340518e-01 -7.03094304e-02 -1.33646631e+00 6.50167823e-01
-1.79932392e+00 -1.06325305e+00 -1.10014498e-01 2.00235510e+00
5.37628949e-01 2.02057794e-01 -2.29453802e-01 4.25944060e-01
3.81023467e-01 -5.36655307e-01 -5.80560088e-01 -4.18569781e-02
-4.81069177e-01 -1.88611895e-01 6.82297826e-01 3.20506394e-01
-7.26641715e-01 5.14875770e-01 6.16768742e+00 5.04835129e-01
-1.14109421e+00 6.99849278e-02 1.14652705e+00 -3.00699562e-01
-5.97780421e-02 -8.05952623e-02 -7.04708457e-01 3.60872418e-01
8.42328906e-01 -5.71053624e-01 1.83176503e-01 5.57614565e-01
6.85990810e-01 -5.14243364e-01 -1.01106536e+00 7.42633402e-01
3.67128588e-02 -1.70513523e+00 -3.90995860e-01 4.59374577e-01
4.42470759e-01 2.68351108e-01 -1.14101522e-01 -2.67517537e-01
1.76296264e-01 -1.07680106e+00 5.26681423e-01 9.28297579e-01
7.83423603e-01 -6.60274804e-01 1.32421863e+00 3.00386280e-01
-6.61599278e-01 -3.44190747e-01 6.00407980e-02 5.52019402e-02
-9.28241163e-02 4.06055331e-01 -1.52453065e+00 4.80808884e-01
5.60538828e-01 6.00550354e-01 -7.24505484e-01 1.26882362e+00
-7.35342083e-03 4.20374900e-01 -1.18120015e-01 -2.31913388e-01
5.67204021e-02 3.14931482e-01 5.08842826e-01 1.07841778e+00
2.52612889e-01 4.67606276e-01 -1.59859821e-01 4.13940996e-01
5.84195793e-01 3.49879235e-01 -2.38618001e-01 -9.19681415e-02
3.16065341e-01 1.31799698e+00 -1.33745539e+00 -2.60143101e-01
-2.48637572e-01 7.06658125e-01 6.22440338e-01 2.63138473e-01
-9.99114215e-02 2.88753919e-02 6.23461068e-01 1.87998086e-01
-8.02319124e-02 2.49545693e-01 -1.02205741e+00 -9.31417584e-01
-5.26198208e-01 -3.94041389e-01 6.64001763e-01 -7.33981967e-01
-1.12670875e+00 6.85551226e-01 -4.10221875e-01 -1.03143597e+00
-3.74697559e-02 -6.63656175e-01 -9.20855939e-01 8.73213530e-01
-1.53358126e+00 -1.25183833e+00 -4.49105084e-01 1.96572900e-01
2.95353770e-01 -2.92928785e-01 1.30884409e+00 -4.71454531e-01
-7.06883848e-01 4.63714361e-01 3.02802712e-01 -2.73487806e-01
4.76346016e-01 -1.46450150e+00 -5.72068095e-01 3.51251066e-01
-6.03833377e-01 3.77429396e-01 1.06227827e+00 -5.33543527e-01
-8.80171835e-01 -1.04249954e+00 9.81715381e-01 -2.13917449e-01
1.07345963e+00 2.87514955e-01 -6.06538832e-01 5.52960932e-01
-8.62876624e-02 -2.48680413e-01 1.56068122e+00 1.56105936e-01
2.11244687e-01 2.34985396e-01 -1.15113127e+00 7.78755486e-01
7.92389154e-01 -3.32689852e-01 -2.72608817e-01 5.95716178e-01
2.79602408e-01 4.55764420e-02 -9.27346766e-01 2.99558312e-01
3.30951691e-01 -8.42130601e-01 5.80881953e-01 -8.89674842e-01
4.96596992e-01 -2.60557443e-01 -7.91152939e-02 -1.43828201e+00
-6.83095276e-01 -1.25510186e-01 5.04277050e-01 5.69188535e-01
7.96889424e-01 -4.84114081e-01 1.19548225e+00 9.30862427e-01
-4.63615298e-01 -1.39381742e+00 -9.36737061e-01 -4.36797976e-01
-1.49585128e-01 -4.92572486e-01 4.78447139e-01 7.24851549e-01
7.69913197e-01 -1.60448819e-01 4.71086830e-01 3.42516810e-01
6.84319317e-01 -6.59801960e-02 2.04680920e-01 -1.14553094e+00
-1.48371235e-01 -1.07546878e+00 -1.13814437e+00 3.20122123e-01
1.96020603e-01 -1.33244300e+00 -1.93804175e-01 -1.67695868e+00
6.54672086e-01 -4.89630491e-01 -5.58365941e-01 4.76893812e-01
-4.77960616e-01 7.43017495e-02 -2.51705170e-01 3.23431432e-01
-3.03145915e-01 3.46157432e-01 6.30342782e-01 -2.70662636e-01
-7.47922882e-02 -9.62896124e-02 -9.68123794e-01 9.45436895e-01
7.77015686e-01 -4.70359504e-01 -1.54227093e-01 7.86832571e-02
9.60766599e-02 2.15713710e-01 4.53513175e-01 -7.83076286e-01
4.76795375e-01 -4.26176667e-01 4.77735549e-01 -5.74805975e-01
1.06000833e-01 -6.93132579e-01 1.25880018e-01 6.28894150e-01
-3.68785411e-01 -1.58318609e-01 -6.44216537e-02 6.39124811e-01
-1.14077918e-01 -3.97543401e-01 6.89016879e-01 -1.23688892e-01
-6.73627377e-01 3.66562635e-01 -8.61818492e-01 -4.65473324e-01
1.39896405e+00 -5.94356954e-01 -5.22446930e-01 -6.07588924e-02
-1.35186124e+00 3.65934102e-03 5.91350675e-01 -2.71972716e-01
5.22456229e-01 -9.43669438e-01 -7.57426023e-01 -6.68735728e-02
6.74466848e-01 -7.92918727e-02 5.14639497e-01 1.41014707e+00
-4.92785990e-01 6.08551741e-01 -2.13365078e-01 -3.68122160e-01
-1.49567330e+00 3.58365983e-01 3.96226555e-01 -5.40879726e-01
-5.32035828e-01 1.17450631e+00 -4.94355112e-02 7.65068382e-02
2.08044782e-01 -5.82038425e-02 -3.93097907e-01 2.03030035e-01
6.21109784e-01 3.69354427e-01 4.79270369e-01 -5.71352422e-01
-4.66706872e-01 -3.01339962e-02 -4.11834389e-01 1.51017904e-02
1.61064255e+00 -1.11177415e-01 -2.97577173e-01 3.53492796e-01
1.15904176e+00 -3.69042039e-01 -8.34215462e-01 7.83320963e-02
4.07128006e-01 -2.89255112e-01 3.26025456e-01 -1.12162578e+00
-7.10329592e-01 3.53245527e-01 5.23684680e-01 -1.78611279e-01
1.16839075e+00 2.34944791e-01 2.45153323e-01 7.81783760e-02
3.21926713e-01 -8.04675400e-01 -3.24754626e-01 4.77799773e-02
7.05241084e-01 -1.38483846e+00 -4.97047184e-03 -2.06266508e-01
-4.07243073e-01 1.25693107e+00 2.69408524e-01 1.57775804e-01
6.16495192e-01 4.53556895e-01 1.97647333e-01 -3.10271055e-01
-1.18594873e+00 -2.18163133e-01 -9.32907164e-02 6.60333812e-01
4.14736867e-01 4.87337351e-01 -5.24574518e-01 1.08021080e+00
-2.90139049e-01 1.12884119e-01 5.95757365e-01 9.99643505e-01
-5.87715626e-01 -9.77438569e-01 -2.03899115e-01 1.18632770e+00
-3.00161988e-01 -3.00760150e-01 -3.31384152e-01 5.52265048e-01
1.25140220e-01 5.67638695e-01 -1.68131627e-02 -1.64288417e-01
-1.26621947e-01 2.94884413e-01 3.67101520e-01 -3.66687238e-01
-4.61166203e-01 1.58172414e-01 6.25267029e-02 -1.18920162e-01
-3.56052458e-01 -1.09537733e+00 -1.62930226e+00 9.70266461e-02
-4.89964992e-01 -8.00285637e-02 8.04045856e-01 1.34666157e+00
3.09051275e-01 7.87102699e-01 1.89486489e-01 -6.50196970e-01
-1.13606386e-01 -7.47427881e-01 -8.12107384e-01 8.70149732e-02
3.36863428e-01 -5.25401950e-01 -5.70615709e-01 1.89015388e-01] | [14.796714782714844, -2.5916895866394043] |
82fda4e4-ff1c-420b-aa83-a2d7ee5a9b87 | few-shot-3d-point-cloud-semantic-segmentation-1 | 2303.15654 | null | https://arxiv.org/abs/2303.15654v1 | https://arxiv.org/pdf/2303.15654v1.pdf | Few-Shot 3D Point Cloud Semantic Segmentation via Stratified Class-Specific Attention Based Transformer Network | 3D point cloud semantic segmentation aims to group all points into different semantic categories, which benefits important applications such as point cloud scene reconstruction and understanding. Existing supervised point cloud semantic segmentation methods usually require large-scale annotated point clouds for training and cannot handle new categories. While a few-shot learning method was proposed recently to address these two problems, it suffers from high computational complexity caused by graph construction and inability to learn fine-grained relationships among points due to the use of pooling operations. In this paper, we further address these problems by developing a new multi-layer transformer network for few-shot point cloud semantic segmentation. In the proposed network, the query point cloud features are aggregated based on the class-specific support features in different scales. Without using pooling operations, our method makes full use of all pixel-level features from the support samples. By better leveraging the support features for few-shot learning, the proposed method achieves the new state-of-the-art performance, with 15\% less inference time, over existing few-shot 3D point cloud segmentation models on the S3DIS dataset and the ScanNet dataset. | ['Song Wang', 'Ziyu Zhao', 'Xinyi Wu', 'Zhenyao Wu', 'Canyu Zhang'] | 2023-03-28 | null | null | null | null | ['point-cloud-segmentation', 'graph-construction'] | ['computer-vision', 'graphs'] | [ 1.22103058e-01 -4.56792749e-02 -3.90979171e-01 -6.41311467e-01
-7.92597651e-01 -1.16567016e-01 3.77211571e-01 4.14984465e-01
-3.04580688e-01 6.86073527e-02 -3.88925761e-01 -8.21141340e-03
-2.78354347e-01 -1.15050340e+00 -8.41525793e-01 -5.28071821e-01
-9.24989432e-02 7.25923181e-01 9.32634175e-01 4.29589823e-02
3.81535679e-01 6.80359662e-01 -1.85866535e+00 4.26639318e-02
9.46149051e-01 1.32209814e+00 4.65170562e-01 5.42740375e-02
-9.84335124e-01 2.94793963e-01 -2.82672226e-01 -2.14690603e-02
4.75225687e-01 1.59273624e-01 -7.16064692e-01 3.15799326e-01
5.60454667e-01 -3.53359699e-01 -7.63055608e-02 1.30514526e+00
3.49342257e-01 4.55687612e-01 4.78553385e-01 -1.41073179e+00
-3.37332278e-01 3.13361913e-01 -7.51815856e-01 1.54646352e-01
-8.95304605e-02 5.10976948e-02 1.09641194e+00 -1.01765239e+00
5.50298691e-01 1.14853215e+00 6.67315304e-01 3.78313422e-01
-8.45774591e-01 -8.19581926e-01 2.62595326e-01 4.32190925e-01
-1.36434329e+00 1.39364311e-02 1.05521286e+00 -3.37768823e-01
1.11297166e+00 -9.34512094e-02 7.79608369e-01 4.17089105e-01
-3.71983558e-01 8.52790356e-01 7.15195119e-01 -6.75580129e-02
6.78104818e-01 -9.25505087e-02 6.22478068e-01 5.97004354e-01
2.02797279e-01 -2.39995241e-01 -3.71982932e-01 -9.94460005e-03
7.53615916e-01 6.49002671e-01 1.58619657e-01 -8.13100278e-01
-8.93685102e-01 7.83124566e-01 8.66642892e-01 3.89153600e-01
-3.25618505e-01 2.67319590e-01 3.05918753e-01 7.87736177e-02
8.04205179e-01 1.77356049e-01 -4.81755197e-01 7.16054216e-02
-1.15576434e+00 8.65052417e-02 5.48264384e-01 1.42417574e+00
1.38418078e+00 -6.87034428e-02 1.84092410e-02 9.43900347e-01
3.08332503e-01 3.54507655e-01 1.59244642e-01 -9.11308706e-01
4.33683455e-01 9.92076099e-01 -1.77111909e-01 -7.57345557e-01
-4.16993976e-01 -3.09903532e-01 -6.68047965e-01 1.00096136e-01
2.36522593e-02 2.55375683e-01 -1.56848776e+00 1.12739003e+00
5.21443605e-01 8.54304194e-01 -3.27110559e-01 8.72028530e-01
1.11132264e+00 7.20556617e-01 1.52076438e-01 3.35666165e-02
1.20043075e+00 -8.30528438e-01 -2.31130317e-01 -3.00138444e-01
5.01501381e-01 -2.85369545e-01 1.08223808e+00 -4.97261174e-02
-7.71190286e-01 -5.94760239e-01 -1.16945994e+00 -1.42572924e-01
-6.69506729e-01 -5.48385203e-01 9.32286263e-01 5.54407358e-01
-8.17132235e-01 7.10152149e-01 -9.54722345e-01 -5.24451137e-01
9.99713957e-01 4.24984545e-01 5.26883528e-02 -3.76664668e-01
-9.00147319e-01 4.85503763e-01 4.12413955e-01 -1.92629978e-01
-5.96604168e-01 -9.79824781e-01 -9.14901495e-01 3.22585881e-01
6.91073716e-01 -6.05817437e-01 1.09104979e+00 -3.02694112e-01
-1.23188794e+00 9.10716474e-01 -2.16971889e-01 -3.53346497e-01
1.91776887e-01 -1.38908669e-01 -1.89601518e-02 3.59273344e-01
4.39541787e-01 8.97053301e-01 7.43240774e-01 -1.26216030e+00
-8.81880283e-01 -6.82238638e-01 8.17801729e-02 1.61540240e-01
-1.06840946e-01 -2.90052921e-01 -7.98768759e-01 -2.84103483e-01
6.60936415e-01 -7.18421459e-01 -4.66601104e-01 1.84764236e-01
-2.05550656e-01 -5.84540784e-01 1.03478193e+00 -4.87542860e-02
5.30479848e-01 -2.29037619e+00 -3.00249029e-02 1.89812332e-01
2.75757372e-01 2.00359344e-01 5.46788909e-02 7.25788176e-02
1.58471346e-01 9.85846221e-02 -5.08903325e-01 -4.00687575e-01
5.84723614e-02 4.31128204e-01 -2.35254735e-01 3.34759712e-01
2.50069529e-01 8.94604206e-01 -9.71867681e-01 -7.74796605e-01
7.35374451e-01 1.59984455e-01 -5.13650179e-01 -1.06376678e-01
-4.68648344e-01 1.27947524e-01 -6.97941542e-01 9.85678792e-01
9.61815894e-01 -3.89996648e-01 -5.09926379e-01 -6.71302974e-02
5.20279491e-03 -8.58205091e-03 -1.09868598e+00 2.28918958e+00
-3.29688430e-01 2.04805225e-01 -2.51291692e-01 -1.15840614e+00
1.05988860e+00 7.30525032e-02 9.46831226e-01 -5.49105227e-01
1.86243147e-01 3.54001522e-01 -4.35651571e-01 -3.21138531e-01
3.25808823e-01 -4.06611025e-01 -1.84044868e-01 1.57077208e-01
3.68383050e-01 -7.53017366e-01 -9.19704586e-02 1.04963563e-01
9.18569028e-01 -5.56852901e-03 -8.08387995e-02 2.79670265e-02
2.03586757e-01 3.61163378e-01 6.61151946e-01 8.48763764e-01
-2.14105055e-01 6.54540241e-01 1.49756595e-01 -4.12010849e-01
-9.76330578e-01 -1.12847638e+00 -1.79969355e-01 8.52893889e-01
7.99183667e-01 -1.59011155e-01 -5.10182977e-01 -5.51536024e-01
2.45708302e-01 9.49723363e-01 -2.02542052e-01 -1.35259122e-01
-4.56012100e-01 -5.86083293e-01 1.25848025e-01 5.65984547e-01
6.23833537e-01 -7.54704595e-01 -6.65289819e-01 1.99340031e-01
6.54064566e-02 -1.25875318e+00 3.13724130e-02 2.25723401e-01
-1.40201175e+00 -1.06423700e+00 -7.08238006e-01 -8.82689178e-01
5.28642416e-01 8.10007513e-01 9.37510490e-01 -6.60088798e-03
-3.22965473e-01 2.90542811e-01 -4.11682963e-01 -4.92549300e-01
2.63869047e-01 1.54265612e-01 -3.04292828e-01 -2.16939911e-01
9.24590707e-01 -7.71636546e-01 -2.95492619e-01 2.83120483e-01
-7.13303387e-01 1.04387030e-02 4.35497105e-01 6.23501837e-01
1.02600992e+00 3.18817526e-01 4.14753377e-01 -8.97357762e-01
9.03061256e-02 -4.04100180e-01 -6.11109972e-01 3.65593694e-02
-4.81993854e-01 -1.85802415e-01 1.91832662e-01 -7.63750076e-02
-8.77809763e-01 1.27745140e-02 -1.67243838e-01 -1.02832973e+00
-3.05464566e-01 2.63230085e-01 -1.53657600e-01 -2.81567663e-01
2.86650479e-01 8.74808952e-02 -2.00258017e-01 -6.80997491e-01
6.14472866e-01 6.01395190e-01 3.14532846e-01 -4.52778637e-01
9.23480451e-01 7.25917459e-01 7.75070712e-02 -1.05811632e+00
-9.77852285e-01 -1.18088853e+00 -9.60955441e-01 -1.78594783e-01
1.12187266e+00 -9.85226691e-01 -3.17129105e-01 5.36895633e-01
-1.13836777e+00 8.61350372e-02 -6.30719781e-01 3.56118172e-01
-6.81804776e-01 3.76181871e-01 -4.66518998e-01 -5.49018919e-01
-3.27208400e-01 -1.06310260e+00 1.35712767e+00 1.57598466e-01
3.73825490e-01 -5.75166404e-01 -2.90352941e-01 3.34245175e-01
3.72990333e-02 2.35090703e-01 1.20377994e+00 -6.93919301e-01
-1.11070633e+00 -3.21616650e-01 -5.29952705e-01 2.73051858e-01
6.25053793e-02 -3.10263991e-01 -7.99535990e-01 -8.01385269e-02
2.49168843e-01 -1.93325579e-01 1.02945960e+00 4.63738948e-01
1.47542500e+00 3.26052725e-01 -4.80402946e-01 8.48961473e-01
1.58630633e+00 9.43918973e-02 3.65532577e-01 1.26906916e-01
1.03304470e+00 4.28988844e-01 8.21143568e-01 3.71818006e-01
2.93989360e-01 3.55327636e-01 6.19053185e-01 -3.07501666e-02
-2.89711840e-02 -1.92458808e-01 -4.90094453e-01 7.82854140e-01
-6.00787774e-02 6.64701909e-02 -1.14433491e+00 6.80856109e-01
-1.96489120e+00 -7.95619071e-01 -1.02262035e-01 1.92620707e+00
3.89663637e-01 4.76337194e-01 -2.11015716e-01 5.93610145e-02
7.79938757e-01 4.11258072e-01 -8.94256473e-01 2.13565230e-01
2.26775825e-01 4.22524691e-01 7.76929796e-01 4.85111997e-02
-1.11495280e+00 1.17565179e+00 5.27267313e+00 1.01998413e+00
-9.69850481e-01 3.44405264e-01 2.89099783e-01 -1.56941593e-01
-1.75081924e-01 1.96395904e-01 -8.51245999e-01 3.01953077e-01
4.06524897e-01 -6.85047805e-02 8.94136280e-02 1.09817302e+00
-1.04069561e-01 -5.17327450e-02 -1.01195586e+00 1.19707417e+00
6.60613701e-02 -1.51027131e+00 1.07497655e-01 -1.69383749e-01
5.90127409e-01 6.22293770e-01 -3.45700651e-01 3.51275057e-01
8.89849812e-02 -6.57538891e-01 5.57275176e-01 3.57938796e-01
6.22341454e-01 -8.19891334e-01 6.48097217e-01 5.92925489e-01
-1.29665768e+00 -4.93311659e-02 -8.00494671e-01 -6.41008615e-02
3.00913543e-01 7.96721697e-01 -8.21390986e-01 5.96767545e-01
9.57751095e-01 1.14724684e+00 -4.02805597e-01 1.33490145e+00
-7.47301355e-02 3.92813355e-01 -5.19984901e-01 -4.19027209e-02
5.23485780e-01 -2.45016411e-01 6.02294683e-01 6.25679433e-01
4.60379511e-01 2.93308020e-01 4.21612442e-01 1.02099717e+00
-5.22967093e-02 -1.04624882e-01 -5.03438354e-01 -9.94719844e-03
6.03977919e-01 1.01330757e+00 -1.13329339e+00 -6.26672924e-01
-6.25455439e-01 6.02775753e-01 2.00477362e-01 1.32435098e-01
-6.50921524e-01 -5.90148926e-01 6.16442323e-01 1.23906046e-01
6.06407642e-01 -5.54097831e-01 -5.70602417e-01 -1.06829941e+00
-1.49639830e-01 -5.48923239e-02 1.72016740e-01 -7.32325017e-01
-1.43153131e+00 1.51827455e-01 2.86859602e-01 -1.45865595e+00
1.24048024e-01 -4.73572493e-01 -7.12806523e-01 5.71223497e-01
-1.71383822e+00 -1.24842906e+00 -5.30688941e-01 5.74784517e-01
9.69745934e-01 2.03514416e-02 3.49774241e-01 4.39511389e-01
-2.29336917e-01 3.07262992e-04 -1.45843849e-01 -4.66899723e-02
2.90201604e-01 -1.13067782e+00 7.26870835e-01 5.74599385e-01
1.46933183e-01 2.58595884e-01 3.63787979e-01 -7.60326087e-01
-1.14742458e+00 -1.29582310e+00 6.13880634e-01 -1.99111104e-01
4.68255311e-01 -4.48319137e-01 -1.24347878e+00 3.33555877e-01
-4.52173740e-01 3.12129557e-01 5.80982089e-01 1.62862569e-01
-1.71996206e-01 -2.33286664e-01 -1.17861044e+00 1.82578981e-01
1.42768884e+00 -4.04588342e-01 -9.57048118e-01 4.84951377e-01
1.25512969e+00 -2.76362449e-01 -8.18103671e-01 6.59649193e-01
-2.16283761e-02 -8.08094203e-01 1.18740630e+00 -4.09548074e-01
1.98789209e-01 -3.52135628e-01 -2.26736113e-01 -9.77841794e-01
-4.06923681e-01 1.09681219e-01 8.46478343e-02 1.01929045e+00
6.07515350e-02 -4.94237959e-01 1.05126035e+00 4.63887155e-01
-5.99716961e-01 -6.38403177e-01 -1.26442909e+00 -8.79142165e-01
-1.18376903e-01 -7.98420787e-01 9.15603220e-01 9.57104206e-01
-4.55108970e-01 2.80074149e-01 1.31007126e-02 2.89203852e-01
1.05712759e+00 4.90599990e-01 7.47868836e-01 -1.65599108e+00
1.22485355e-01 -4.07494426e-01 -8.29514027e-01 -1.05585063e+00
8.39422792e-02 -1.08378887e+00 2.01876253e-01 -1.86847353e+00
-5.00977859e-02 -8.66163015e-01 -4.03927922e-01 5.75523913e-01
6.45712987e-02 1.81135997e-01 3.60403508e-01 3.88282865e-01
-6.68438196e-01 6.33920193e-01 1.17898798e+00 -4.78831172e-01
-2.82122791e-01 1.54580593e-01 -2.27960169e-01 8.44176769e-01
5.93265295e-01 -5.58824480e-01 -5.75106621e-01 -5.91525793e-01
9.56452423e-05 -4.46431376e-02 5.11383414e-01 -1.34138000e+00
4.80750501e-01 -2.75295198e-01 1.36445910e-01 -1.32768595e+00
5.38358986e-01 -9.71815705e-01 -4.57182042e-02 2.62515783e-01
1.38154939e-01 -6.62203848e-01 6.64347783e-02 8.41635048e-01
-1.93223655e-01 -3.49095672e-01 8.35751593e-01 -4.98471767e-01
-1.41463125e+00 8.17335725e-01 2.04941988e-01 -1.27088562e-01
1.19066107e+00 -5.93255401e-01 -7.64589608e-02 2.56282628e-01
-5.98883986e-01 5.76774478e-01 6.62615836e-01 5.86006343e-01
8.36419523e-01 -1.13012898e+00 -1.98203400e-01 2.50540882e-01
3.06778461e-01 9.03116405e-01 5.27017593e-01 5.65572858e-01
-3.98620337e-01 2.39115372e-01 -1.91886470e-01 -1.18037200e+00
-8.22830558e-01 6.25083089e-01 1.41569406e-01 2.84619957e-01
-1.05763841e+00 1.02126467e+00 8.80432799e-02 -5.13194799e-01
1.71697795e-01 -5.66870391e-01 -9.08423811e-02 1.03495009e-01
5.08247362e-03 3.93753439e-01 1.06761523e-01 -3.83335531e-01
-4.37533677e-01 1.02898967e+00 1.20302159e-02 3.01846594e-01
1.61995649e+00 3.33731025e-02 -1.16808765e-01 8.53100479e-01
1.17441428e+00 -6.47371888e-01 -1.22177255e+00 -5.38613081e-01
1.49348751e-01 -7.01990485e-01 3.83577138e-01 -3.04570675e-01
-1.10200858e+00 1.10614908e+00 6.11447513e-01 5.33467755e-02
8.39447439e-01 4.64763969e-01 1.12411106e+00 5.45971334e-01
8.90076458e-01 -1.22791719e+00 -7.37439394e-02 5.08388221e-01
1.39542952e-01 -1.45182443e+00 1.14473954e-01 -8.46253276e-01
-2.16688558e-01 8.33664060e-01 6.09268725e-01 -3.74019146e-01
9.53590274e-01 -2.68575996e-01 -2.40123555e-01 -6.29782379e-01
-3.30875665e-01 -4.67413008e-01 1.58461526e-01 5.92881680e-01
-2.90721714e-01 3.25854197e-02 1.12356827e-01 4.73293066e-01
-9.10723768e-03 1.59591854e-01 1.18690655e-01 1.09233880e+00
-1.00445199e+00 -8.43802035e-01 2.48786062e-03 9.33485329e-01
2.06138715e-01 1.33972809e-01 9.75481048e-02 6.24712825e-01
3.67476463e-01 6.63973451e-01 5.66537797e-01 -3.09710562e-01
4.01365608e-01 2.62406357e-02 3.05985093e-01 -1.11390793e+00
-5.17937727e-02 -1.47824828e-02 -3.95806551e-01 -5.50184608e-01
-6.62307024e-01 -5.56919336e-01 -1.64243829e+00 2.15302110e-02
-5.40238798e-01 6.04144968e-02 8.87665272e-01 1.03775692e+00
4.38194096e-01 5.29031456e-01 5.77371061e-01 -1.14150155e+00
-3.36272091e-01 -7.49933779e-01 -8.29204500e-01 4.05443758e-01
7.44381770e-02 -9.55817640e-01 -3.05451602e-01 -3.53300273e-01] | [8.008505821228027, -3.1560308933258057] |
71d8b7dd-8c9b-488d-9e0f-0f54919c2442 | deep-insights-of-learning-based-micro | 2210.04935 | null | https://arxiv.org/abs/2210.04935v1 | https://arxiv.org/pdf/2210.04935v1.pdf | Deep Insights of Learning based Micro Expression Recognition: A Perspective on Promises, Challenges and Research Needs | Micro expression recognition (MER) is a very challenging area of research due to its intrinsic nature and fine-grained changes. In the literature, the problem of MER has been solved through handcrafted/descriptor-based techniques. However, in recent times, deep learning (DL) based techniques have been adopted to gain higher performance for MER. Also, rich survey articles on MER are available by summarizing the datasets, experimental settings, conventional and deep learning methods. In contrast, these studies lack the ability to convey the impact of network design paradigms and experimental setting strategies for DL-based MER. Therefore, this paper aims to provide a deep insight into the DL-based MER frameworks with a perspective on promises in network model designing, experimental strategies, challenges, and research needs. Also, the detailed categorization of available MER frameworks is presented in various aspects of model design and technical characteristics. Moreover, an empirical analysis of the experimental and validation protocols adopted by MER methods is presented. The challenges mentioned earlier and network design strategies may assist the affective computing research community in forging ahead in MER research. Finally, we point out the future directions, research needs, and draw our conclusions. | ['Girdhari Singh', 'Santosh Kumar Vipparthi', 'Monu Verma'] | 2022-10-10 | null | null | null | null | ['micro-expression-recognition'] | ['computer-vision'] | [-1.25815555e-01 -9.75702628e-02 -4.98672813e-01 -5.65164626e-01
7.32416511e-02 -3.15520585e-01 2.06059843e-01 -1.40765170e-02
-2.66556680e-01 5.44868946e-01 -1.23633511e-01 2.67981917e-01
-2.80614078e-01 -7.13706672e-01 2.68605072e-02 -8.50619435e-01
-2.25499392e-01 1.05772614e-01 -5.31996310e-01 -5.38860977e-01
-6.73563639e-03 9.68131125e-01 -1.49264121e+00 1.02030344e-01
1.33840799e-01 1.44813335e+00 -2.72806436e-01 2.55770892e-01
-1.47252709e-01 9.85712230e-01 -6.68611944e-01 -8.18619251e-01
-7.91443214e-02 -3.17285120e-01 -5.88553548e-01 1.38197653e-02
-1.42677307e-01 -2.59777248e-01 -5.33268929e-01 8.19640875e-01
1.00494099e+00 1.94434926e-01 5.74699879e-01 -1.73314202e+00
-5.15349686e-01 5.57535946e-01 -3.62301022e-01 1.19074360e-01
2.18063161e-01 -9.49619785e-02 1.07610536e+00 -9.22743618e-01
5.03564894e-01 1.10113633e+00 7.25851595e-01 6.61399961e-01
-1.08552468e+00 -7.73864985e-01 1.39042497e-01 5.13016403e-01
-1.50707972e+00 -7.59363353e-01 1.09370959e+00 -2.88640648e-01
1.07000840e+00 1.83931798e-01 9.49699521e-01 1.53814757e+00
1.28241390e-01 8.54821801e-01 1.04999220e+00 -4.59076583e-01
4.49756056e-01 4.95534182e-01 2.31969312e-01 5.24711788e-01
-9.72501561e-02 -1.67984262e-01 -3.79036576e-01 -1.38938755e-01
5.96379340e-01 -1.89166948e-01 2.58835465e-01 -3.18013281e-01
-4.04181927e-01 1.05274653e+00 2.35949382e-01 6.86955512e-01
-5.58409333e-01 3.37605812e-02 9.38061953e-01 4.21216935e-01
4.83355314e-01 5.79207063e-01 -2.84558654e-01 -4.57554072e-01
-7.95487583e-01 8.16133544e-02 9.96557593e-01 8.39733005e-01
7.18357146e-01 5.69584906e-01 -1.77859724e-01 1.28528070e+00
2.15687826e-01 -1.27014384e-01 3.49259228e-01 -1.19743443e+00
-3.50385308e-01 4.20291930e-01 -3.44408065e-01 -1.69742692e+00
-7.15031147e-01 -4.67213959e-01 -1.01431596e+00 6.71568932e-03
-2.13935599e-01 -5.68429768e-01 -8.70960802e-02 1.69360697e+00
4.87904474e-02 -1.23745181e-01 -1.27716273e-01 8.34166288e-01
1.03475928e+00 3.55396092e-01 2.39365399e-01 -3.33995879e-01
1.27450967e+00 -9.91142511e-01 -1.15298367e+00 -1.67164858e-02
6.84030473e-01 -5.38393974e-01 8.63931715e-01 3.52534711e-01
-9.12070990e-01 -2.77820915e-01 -9.50586796e-01 1.54506713e-01
-6.11792207e-01 2.37832159e-01 1.14629400e+00 1.12602985e+00
-1.13742971e+00 3.95371675e-01 -5.85031867e-01 -9.78381336e-01
4.04319197e-01 4.79408741e-01 -2.86122590e-01 2.07537025e-01
-1.37813854e+00 9.52841103e-01 3.28996256e-02 3.84401709e-01
-6.87903464e-01 -3.24917674e-01 -5.90172052e-01 -8.44798703e-03
9.64317471e-02 -5.42067111e-01 1.20186222e+00 -1.24199069e+00
-2.08684731e+00 9.55424786e-01 -9.83644463e-03 -4.70660329e-01
1.18003942e-01 1.75534457e-01 -7.57276833e-01 2.61188716e-01
-4.83907878e-01 5.10402262e-01 5.77170432e-01 -9.99426901e-01
-1.28677264e-01 -2.82959551e-01 4.24173564e-01 1.45636901e-01
-7.58313596e-01 4.33683753e-01 -3.25811028e-01 -4.78435308e-01
-5.57429969e-01 -8.78137171e-01 -2.61225343e-01 2.39175782e-01
-1.34754777e-01 -2.64552146e-01 8.17384362e-01 1.00132816e-01
1.57819521e+00 -2.20673895e+00 -1.02800563e-01 3.46189469e-01
5.31073868e-01 4.21810538e-01 -2.12882385e-01 9.61682677e-01
-1.86592460e-01 1.43501624e-01 5.13269484e-01 -4.21641737e-01
3.36540967e-01 1.88589334e-01 1.19385093e-01 4.92493898e-01
-5.82750365e-02 9.82584357e-01 -6.37200356e-01 -3.61501366e-01
5.28461397e-01 7.16891408e-01 -2.43372098e-01 2.44247913e-01
1.64850026e-01 -5.85157797e-02 -4.23954874e-01 9.96956706e-01
4.37037766e-01 -1.15845032e-01 4.15473133e-01 -6.64479852e-01
-1.66738685e-02 -3.27889115e-01 -8.33481133e-01 1.11055541e+00
-6.37461066e-01 8.88086736e-01 4.41074908e-01 -1.27333963e+00
1.33470321e+00 3.97666186e-01 9.12228227e-01 -9.11006451e-01
5.66109121e-01 4.92429808e-02 -3.48395966e-02 -6.79498851e-01
4.93285209e-01 -2.72239260e-02 3.48779671e-02 3.21280181e-01
1.28592521e-01 4.04078901e-01 -1.62067711e-02 -3.48359421e-02
8.81778121e-01 -5.22405207e-01 2.97559381e-01 -1.22848023e-02
4.86546636e-01 -4.54367757e-01 5.71815372e-01 4.79617178e-01
-8.16624880e-01 2.38460824e-02 7.39207685e-01 -3.72249186e-01
-8.47744703e-01 -4.86838728e-01 -1.44030228e-01 1.39835155e+00
-4.15691845e-02 -7.49339283e-01 -9.02682424e-01 -4.04817343e-01
-2.91962415e-01 5.47351658e-01 -8.59459579e-01 -2.63296813e-01
-9.82942730e-02 -1.02578413e+00 9.74547565e-01 5.71708381e-01
6.38743162e-01 -9.52610672e-01 -5.02292216e-01 3.11094075e-01
-2.03482509e-01 -1.28969049e+00 2.37460971e-01 4.72833849e-02
-6.64362609e-01 -9.35154796e-01 -4.16252077e-01 -5.56477010e-01
2.35159233e-01 1.34359822e-01 9.68537569e-01 -9.53959227e-02
-1.33521974e-01 6.40584588e-01 -5.44497728e-01 -4.57626462e-01
-4.91952337e-02 2.45696038e-01 2.30946675e-01 2.18567610e-01
9.04342353e-01 -8.40174913e-01 -7.15492129e-01 3.95264328e-01
-6.40023053e-01 -3.40133876e-01 8.09667587e-01 6.90786481e-01
4.67352092e-01 1.15596585e-01 9.26107168e-01 -6.98607981e-01
1.20142102e+00 -6.69716060e-01 5.59290387e-02 1.22461684e-01
-9.96170104e-01 -5.16469479e-01 3.49756360e-01 -1.62230730e-01
-8.41428995e-01 -4.21179622e-01 -5.14280081e-01 -4.33980733e-01
-2.85973787e-01 6.62629485e-01 -1.55253693e-01 -4.60302204e-01
6.37420833e-01 1.42552825e-02 3.26592356e-01 -2.99837142e-01
8.87084231e-02 9.54809546e-01 -2.20237747e-01 -6.05868459e-01
-1.16189852e-01 3.89007777e-01 -5.63654155e-02 -1.05563045e+00
-7.52180815e-01 -2.43605345e-01 -4.00258183e-01 -6.37078345e-01
5.79371691e-01 -8.18951964e-01 -8.03577363e-01 6.47984564e-01
-7.70264864e-01 -2.61273474e-01 -1.77531227e-01 1.61522686e-01
-5.66680610e-01 1.76964954e-01 -9.40120697e-01 -9.86619174e-01
-7.55372763e-01 -1.17402232e+00 7.97209978e-01 2.76593834e-01
-6.82229400e-01 -1.30373418e+00 5.79460384e-03 4.24033105e-01
7.96136737e-01 4.93721962e-01 8.73984098e-01 -6.25951767e-01
2.33258247e-01 -3.77991706e-01 -2.22387254e-01 4.84690398e-01
-7.80736879e-02 4.20721829e-01 -1.25554752e+00 -2.85130352e-01
-1.10612519e-01 -5.49892604e-01 1.14330009e-01 2.65130162e-01
1.41846621e+00 -1.74061283e-01 -2.39564255e-01 5.55262387e-01
1.21806693e+00 3.46462101e-01 6.16731107e-01 4.22641248e-01
3.02404135e-01 7.08314002e-01 6.01048648e-01 8.41427684e-01
3.83873135e-01 8.22449625e-01 4.04585749e-01 -1.72072351e-01
9.14148986e-02 9.00372788e-02 4.70493168e-01 1.01064301e+00
-1.30507231e-01 -3.57655168e-01 -6.34689152e-01 1.67525172e-01
-1.80026865e+00 -9.00237978e-01 1.58217505e-01 1.46062970e+00
4.99736130e-01 -2.79369742e-01 3.07775527e-01 2.63678938e-01
6.07429802e-01 3.08687001e-01 -6.38591707e-01 -8.99078071e-01
-4.14480209e-01 7.86088854e-02 4.57014851e-02 6.36834130e-02
-9.42877710e-01 9.32190061e-01 7.28083324e+00 1.05781102e+00
-1.56121850e+00 -3.99381444e-02 6.49585724e-01 -1.51205473e-02
1.66049004e-01 -4.66158479e-01 -7.41714418e-01 2.08410725e-01
1.20478177e+00 -3.04339558e-01 3.72321069e-01 1.12540531e+00
5.50238788e-01 6.36796132e-02 -1.07710683e+00 1.61608863e+00
9.01029184e-02 -1.31802547e+00 -2.45812222e-01 -3.64101343e-02
4.25458610e-01 1.19493559e-01 8.93822834e-02 4.64003533e-01
-1.48972243e-01 -1.02209246e+00 3.87907326e-01 5.56400836e-01
7.96041131e-01 -8.96825671e-01 1.00769293e+00 -1.70359641e-01
-1.02400696e+00 -1.84035242e-01 -3.98492157e-01 -5.39303362e-01
-1.83405012e-01 7.33671784e-01 -4.90195215e-01 4.24704045e-01
8.53106558e-01 9.74869013e-01 -3.76631558e-01 7.33127832e-01
4.37320471e-02 6.56592190e-01 -1.38732344e-01 -5.82391202e-01
1.86363399e-01 -2.86518425e-01 2.11327866e-01 1.57946229e+00
-2.09611971e-02 -4.27020863e-02 -7.80469626e-02 7.87357152e-01
-4.86941040e-02 3.56879979e-01 -6.62021220e-01 -4.69696820e-01
6.14965677e-01 1.97896576e+00 -4.28500980e-01 5.02982177e-02
-3.96193802e-01 6.55029953e-01 3.86503577e-01 4.38684285e-01
-7.48471320e-01 -5.91440618e-01 1.08325577e+00 -6.06537890e-03
-3.23308170e-01 -1.86464578e-01 -2.77882963e-01 -9.53568041e-01
-4.65231717e-01 -9.98138428e-01 2.68253297e-01 -7.52118886e-01
-1.53540123e+00 7.40497530e-01 -4.66088951e-02 -8.11031282e-01
-5.19421324e-02 -6.69663727e-01 -5.86205482e-01 4.85175729e-01
-1.20469856e+00 -9.65438902e-01 -4.92692649e-01 5.78230858e-01
3.99056584e-01 -4.99211788e-01 1.31515563e+00 6.73527598e-01
-1.22753596e+00 1.02695417e+00 2.76341766e-01 2.01826289e-01
6.48106337e-01 -6.74258292e-01 -3.99545133e-01 1.55743524e-01
-2.51533598e-01 5.89198828e-01 6.56944513e-01 1.62439182e-01
-1.46668494e+00 -9.16549861e-01 6.21214807e-01 3.04973386e-02
8.02518427e-01 -5.95632613e-01 -3.99770677e-01 6.41166747e-01
4.32206422e-01 -1.56024009e-01 1.40273225e+00 3.08692425e-01
9.98433027e-03 -5.08669078e-01 -1.43209302e+00 8.89227450e-01
8.07288170e-01 -5.69844842e-01 1.90609813e-01 1.61200926e-01
4.58208732e-02 5.37544563e-02 -1.34759665e+00 3.18287432e-01
8.22068512e-01 -1.18732274e+00 6.27949834e-01 -4.23931688e-01
1.46845847e-01 2.77394265e-01 -3.06397736e-01 -1.20046139e+00
-5.78158617e-01 -6.47397041e-01 -3.67008626e-01 1.42463303e+00
6.05246238e-03 -4.48746502e-01 8.84969831e-01 7.38092005e-01
1.28729910e-01 -1.32035017e+00 -5.75524867e-01 -5.52660942e-01
-2.99739912e-02 -6.93889678e-01 4.59732682e-01 1.11162901e+00
3.19760323e-01 6.86562538e-01 -3.42871279e-01 -4.82986510e-01
3.41045767e-01 -1.95117950e-01 7.57338107e-01 -1.07960844e+00
1.58508137e-01 -7.37029731e-01 -5.74746490e-01 -6.77143633e-01
4.72845316e-01 -6.89749777e-01 -7.34515965e-01 -1.28188455e+00
6.98809177e-02 -5.45255244e-01 -4.52744484e-01 4.59755570e-01
6.30030990e-01 2.04664662e-01 1.41475827e-01 1.97478868e-02
-7.51726925e-01 8.61973286e-01 9.35535431e-01 -4.07217555e-02
3.49424332e-02 -2.61816680e-01 -9.63330626e-01 6.49269223e-01
1.21449411e+00 -7.13670254e-02 -6.73168063e-01 -1.97075501e-01
4.40497100e-01 -2.61025518e-01 1.49732381e-01 -7.84334719e-01
3.33729655e-01 -7.99772963e-02 3.08432788e-01 -2.32495829e-01
6.57080233e-01 -9.97637272e-01 1.74123824e-01 -1.54041111e-01
-3.73605281e-01 8.09000283e-02 1.74569502e-01 2.41471231e-01
-3.58032107e-01 -1.63156569e-01 9.70824182e-01 1.01426519e-01
-1.05010617e+00 5.05872250e-01 -8.70594263e-01 -1.83614194e-01
1.30992866e+00 -5.03008008e-01 1.22212015e-01 -6.58262074e-01
-9.32689011e-01 2.08604231e-01 6.12718202e-02 4.19312894e-01
5.88904083e-01 -1.39969563e+00 -1.93112820e-01 1.36283943e-02
2.14703277e-01 -6.92661107e-01 4.58725482e-01 1.19714665e+00
-2.86376864e-01 3.73294204e-01 -6.14282012e-01 -2.95823455e-01
-1.34282637e+00 3.56616199e-01 7.21360087e-01 -8.17142427e-02
-5.30313961e-02 6.50923848e-01 -1.29867077e-01 -5.32312334e-01
4.28253680e-01 4.25805479e-01 -5.30837893e-01 4.43537861e-01
3.11672777e-01 7.03276277e-01 1.87338695e-01 -5.86339951e-01
-4.15363640e-01 3.85920614e-01 -9.69282165e-02 2.32092097e-01
1.42875195e+00 -6.22735977e-01 -3.26469570e-01 6.68993115e-01
1.40597451e+00 -3.40700328e-01 -6.00578249e-01 -1.56557769e-01
-9.07267854e-02 -5.03848642e-02 2.54587024e-01 -8.02594185e-01
-1.51696932e+00 9.78596210e-01 6.35136008e-01 1.26527086e-01
1.30995238e+00 -2.53246158e-01 5.47187567e-01 5.38423121e-01
4.45449054e-01 -1.64934301e+00 1.57846853e-01 6.29725277e-01
7.91358590e-01 -1.11933088e+00 -3.41051131e-01 -2.55417317e-01
-6.69690132e-01 1.45783281e+00 7.65838146e-01 7.44967461e-02
1.08958137e+00 4.92895424e-01 1.41514257e-01 -4.67900157e-01
-8.68654311e-01 1.12811983e-01 -3.94554585e-01 9.29591298e-01
8.73740494e-01 -2.64824387e-02 -3.83450240e-01 9.79302347e-01
-1.62448391e-01 1.86679170e-01 3.11882257e-01 8.24079275e-01
-1.66278690e-01 -1.24204588e+00 1.18892930e-01 5.45751691e-01
-6.92819655e-01 1.44451633e-01 -7.30353355e-01 6.67098284e-01
1.01338392e-02 1.12272429e+00 -1.09772660e-01 -8.39966416e-01
5.64908922e-01 -5.00152521e-02 1.88570067e-01 -2.42857024e-01
-6.11155808e-01 -1.74941689e-01 4.69603002e-01 -6.39347017e-01
-6.61640882e-01 -5.52151322e-01 -9.25068438e-01 -9.42498386e-01
-2.08082601e-01 2.34775037e-01 5.40163338e-01 7.18135595e-01
7.20264077e-01 5.28654397e-01 7.99941599e-01 -7.44850278e-01
-1.43765762e-01 -7.62227416e-01 -8.52351308e-01 1.82640478e-01
-2.01526895e-01 -8.22182655e-01 -2.01335117e-01 -4.27882403e-01] | [13.593334197998047, 1.8767642974853516] |
b39d5b8d-4888-43f9-9156-918403bb4be3 | time-and-cost-efficient-bathymetric-mapping | 2210.10263 | null | https://arxiv.org/abs/2210.10263v1 | https://arxiv.org/pdf/2210.10263v1.pdf | Time and Cost-Efficient Bathymetric Mapping System using Sparse Point Cloud Generation and Automatic Object Detection | Generating 3D point cloud (PC) data from noisy sonar measurements is a problem that has potential applications for bathymetry mapping, artificial object inspection, mapping of aquatic plants and fauna as well as underwater navigation and localization of vehicles such as submarines. Side-scan sonar sensors are available in inexpensive cost ranges, especially in fish-finders, where the transducers are usually mounted to the bottom of a boat and can approach shallower depths than the ones attached to an Uncrewed Underwater Vehicle (UUV) can. However, extracting 3D information from side-scan sonar imagery is a difficult task because of its low signal-to-noise ratio and missing angle and depth information in the imagery. Since most algorithms that generate a 3D point cloud from side-scan sonar imagery use Shape from Shading (SFS) techniques, extracting 3D information is especially difficult when the seafloor is smooth, is slowly changing in depth, or does not have identifiable objects that make acoustic shadows. This paper introduces an efficient algorithm that generates a sparse 3D point cloud from side-scan sonar images. This computation is done in a computationally efficient manner by leveraging the geometry of the first sonar return combined with known positions provided by GPS and down-scan sonar depth measurement at each data point. Additionally, this paper implements another algorithm that uses a Convolutional Neural Network (CNN) using transfer learning to perform object detection on side-scan sonar images collected in real life and generated with a simulation. The algorithm was tested on both real and synthetic images to show reasonably accurate anomaly detection and classification. | ['Jaejeong Shin', 'Peter Ifju', 'Andrew Ortega', 'Antonio Diaz', 'Ruoyao Qin', 'Andres Pulido'] | 2022-10-19 | null | null | null | null | ['point-cloud-generation'] | ['computer-vision'] | [ 2.09506318e-01 -1.32926971e-01 9.48726296e-01 -3.98872018e-01
-3.12127322e-01 -6.10678136e-01 2.90374070e-01 2.94528324e-02
-7.74388313e-01 3.93694282e-01 -3.03473592e-01 -2.56305993e-01
-1.12647846e-01 -9.66327965e-01 -7.39991724e-01 -7.43876576e-01
-6.72046006e-01 5.59819579e-01 4.12284434e-01 -7.51263499e-01
2.53422707e-01 7.61938751e-01 -1.76252687e+00 -4.17409092e-01
6.91602826e-01 9.15805697e-01 6.15224838e-01 9.04186368e-01
-5.98975755e-02 -1.62632659e-01 -4.32731569e-01 1.43844321e-01
6.99719489e-01 -2.13647515e-01 9.93570238e-02 -1.40242800e-01
7.45919168e-01 -6.67432845e-01 1.08015470e-01 1.07771432e+00
4.93155688e-01 1.36852011e-01 5.99176645e-01 -6.81020617e-01
3.03475380e-01 1.49362579e-01 -4.92154568e-01 1.03794284e-01
1.13302976e-01 -2.04162017e-01 4.55589890e-01 -9.42975461e-01
3.48014496e-02 9.99372363e-01 1.13664412e+00 2.76334375e-01
-6.42830014e-01 -5.69370389e-01 -5.23349106e-01 -2.07155094e-01
-1.25342524e+00 -1.62711427e-01 6.84935987e-01 -3.78966987e-01
7.31134355e-01 7.89605081e-02 1.06444633e+00 1.47844076e-01
3.11330110e-01 2.92596996e-01 8.98504019e-01 -3.35099846e-01
4.27019596e-01 -4.53880131e-01 -4.75645304e-01 6.40526652e-01
5.88913500e-01 1.13165505e-01 -4.36525911e-01 -2.17465937e-01
7.80469656e-01 3.05850446e-01 -6.13414466e-01 -3.86293143e-01
-7.03028917e-01 7.97366560e-01 4.37601417e-01 -8.61729961e-03
-3.49288791e-01 3.59899789e-01 7.10398480e-02 5.12959540e-01
4.39196855e-01 4.57422316e-01 -5.89038134e-01 -2.53958017e-01
-9.86562014e-01 3.16938877e-01 1.06022048e+00 7.26205289e-01
9.73071575e-01 6.67435825e-01 1.25215137e+00 6.49790704e-01
7.19833136e-01 1.51354086e+00 4.58944649e-01 -6.70014858e-01
1.52520895e-01 2.95805216e-01 3.19680065e-01 -1.07350051e+00
-5.48155725e-01 -3.39073777e-01 -5.52114606e-01 6.68008029e-01
1.83590904e-01 -5.69615722e-01 -1.06176877e+00 9.78840232e-01
4.18883711e-01 4.30886418e-01 6.41902745e-01 1.27432847e+00
9.47588503e-01 7.72851825e-01 -6.80446506e-01 9.69197974e-03
1.15626287e+00 -1.50379315e-01 -4.53381062e-01 -4.18616831e-01
7.30668128e-01 -5.12530148e-01 3.90597463e-01 4.21771079e-01
-6.43120944e-01 -1.90157935e-01 -1.26019156e+00 3.83067161e-01
-4.02337134e-01 -2.62743592e-01 3.75577897e-01 4.84039247e-01
-1.07922518e+00 4.41628426e-01 -1.29411125e+00 -1.79112598e-01
-6.09594546e-02 3.09390545e-01 -5.70662141e-01 -5.01177497e-02
-9.58654940e-01 8.18590701e-01 -2.19621450e-01 6.58630371e-01
-1.06907523e+00 -6.62484646e-01 -1.43533564e+00 -1.58311799e-01
-8.50953609e-02 -2.02843532e-01 1.03899944e+00 -7.88351297e-01
-1.47230172e+00 3.17988425e-01 1.48327291e-01 -6.01041079e-01
2.40658686e-01 -4.80245650e-01 -9.87984389e-02 2.95311719e-01
-3.08563355e-02 3.29357326e-01 8.45237076e-01 -1.26949072e+00
-7.78524458e-01 -4.34945494e-01 -1.85701549e-01 4.96487588e-01
2.28487998e-01 -4.83391494e-01 3.15496355e-01 -8.47467333e-02
9.73792493e-01 -7.93936312e-01 -4.49153304e-01 1.72230050e-01
1.79451153e-01 3.93389910e-01 1.21990776e+00 -3.36407602e-01
1.28407151e-01 -2.08986902e+00 -3.34286600e-01 3.33098024e-01
-4.38182652e-01 2.61439532e-01 -6.55181259e-02 7.40292072e-01
3.54020059e-01 -1.76143885e-01 -4.61383343e-01 -2.89419681e-01
-4.38230872e-01 6.98858976e-01 -2.41447493e-01 8.21156561e-01
-7.73943961e-02 1.45232737e-01 -9.89264905e-01 -7.95178711e-02
3.88789207e-01 4.95704979e-01 -4.95881230e-01 2.50645995e-01
2.72780925e-01 3.00594598e-01 -2.85078555e-01 7.04717219e-01
1.10125065e+00 7.59956062e-01 -3.94467324e-01 2.15094104e-01
-7.92994440e-01 -2.48711053e-02 -1.45081162e+00 1.35205114e+00
-6.92719460e-01 8.89847279e-01 8.17067206e-01 -8.20944250e-01
1.52532721e+00 2.65615076e-01 1.39265299e-01 -4.36528802e-01
-2.47438014e-01 6.54300153e-01 1.82248726e-02 -8.45836997e-01
7.00929582e-01 -4.49896663e-01 1.25382915e-01 -6.50924025e-03
-3.03175896e-01 -1.03681350e+00 -4.39235687e-01 -1.93616375e-01
9.75101352e-01 2.69475281e-02 2.08596084e-02 -3.96913528e-01
3.05819631e-01 1.63397044e-01 5.89430094e-01 6.87557578e-01
3.59328240e-01 9.26108122e-01 -3.69362831e-02 -8.54310572e-01
-9.63172734e-01 -7.64256477e-01 -2.90357590e-01 4.17522788e-01
4.40346450e-01 4.04234439e-01 -2.81353593e-01 -2.70857811e-02
1.61268190e-01 1.15457162e-01 -4.14835125e-01 3.23006988e-01
-5.56889236e-01 -3.85962367e-01 4.35342193e-01 2.60598660e-01
6.70529902e-01 -7.19757617e-01 -1.42495656e+00 3.35387588e-01
3.77791524e-01 -8.40376675e-01 2.99906224e-01 2.30310857e-01
-1.32917511e+00 -1.02829790e+00 -6.28162563e-01 -8.05063665e-01
8.57474744e-01 6.92943394e-01 7.68646777e-01 2.94532686e-01
-1.45704359e-01 4.41207230e-01 -8.80930066e-01 -9.07506347e-01
-1.46569133e-01 -6.57640398e-01 2.25197017e-01 -1.93886593e-01
2.57100046e-01 -7.60107577e-01 -7.10691750e-01 4.96095061e-01
-8.72476816e-01 -2.97358990e-01 4.85061318e-01 8.98515284e-01
3.47283959e-01 1.74980059e-01 1.74719110e-01 -3.59654337e-01
-1.07198425e-01 -4.23310548e-01 -1.19374561e+00 -6.55705452e-01
9.70520079e-02 -4.37261939e-01 3.28041136e-01 6.98595420e-02
-6.98978901e-01 3.98090184e-01 -4.36713994e-01 -1.77204549e-01
-3.90966564e-01 7.38827825e-01 2.92286158e-01 -4.64468002e-01
6.86802208e-01 5.72101295e-01 3.76017421e-01 -4.98723000e-01
-4.00946677e-01 8.59072924e-01 3.95237118e-01 2.03063130e-01
1.13484156e+00 1.06461143e+00 4.46801126e-01 -1.82391226e+00
-3.71553928e-01 -7.21966922e-01 -4.64355886e-01 -1.68756306e-01
3.99230152e-01 -1.01384997e+00 -4.32202399e-01 7.42864370e-01
-8.75586808e-01 -4.77207094e-01 6.06243312e-02 8.81200194e-01
-7.10380450e-02 4.90559101e-01 -1.74759939e-01 -1.16751027e+00
-4.90514070e-01 -9.93096113e-01 1.17276788e+00 4.80988473e-01
3.14477146e-01 -1.11310935e+00 2.12260574e-01 -4.53657098e-02
4.88249511e-01 4.20911103e-01 9.98873264e-02 -4.71467823e-01
-4.07297552e-01 -5.09199440e-01 2.40980968e-01 3.15052122e-01
1.71604097e-01 -6.42580837e-02 -7.38538027e-01 -3.81544650e-01
2.17430085e-01 -1.89827755e-01 7.17283010e-01 7.49179542e-01
1.00315511e-01 -2.09085375e-01 -1.06077343e-01 9.76398349e-01
1.78898132e+00 2.06356332e-01 3.82243216e-01 3.33174229e-01
3.99523467e-01 5.81628561e-01 9.66954112e-01 6.09292030e-01
2.56702751e-01 2.12259755e-01 1.26740921e+00 -1.55568615e-01
2.75412172e-01 4.81497981e-02 2.72778243e-01 5.75337112e-01
-3.55700284e-01 -8.12367052e-02 -9.26373541e-01 9.06499147e-01
-1.34468412e+00 -6.47370398e-01 -7.16961384e-01 2.26902676e+00
9.70159471e-02 -3.33293617e-01 -7.06829786e-01 2.58007973e-01
2.31009364e-01 -2.19028443e-01 -1.18476242e-01 -5.00022590e-01
-1.86679456e-02 3.58299255e-01 1.02698851e+00 8.67826521e-01
-8.36261272e-01 6.87609971e-01 4.80638409e+00 -7.18044415e-02
-1.38940120e+00 -2.88231641e-01 -4.43321228e-01 4.78304714e-01
-3.13678265e-01 -7.38540292e-02 -8.70802343e-01 2.39018306e-01
5.18014193e-01 6.96105480e-01 -1.23993337e-01 1.02781975e+00
5.50006032e-01 -7.32326508e-01 -6.87276542e-01 8.95602047e-01
1.57921717e-01 -1.11860108e+00 -2.21222416e-01 1.27695231e-02
7.48357713e-01 5.34057140e-01 -2.73563117e-01 -3.04685503e-01
1.37156308e-01 -5.75000405e-01 5.85071206e-01 3.49996477e-01
5.63333511e-01 -5.89584708e-01 1.55437529e+00 5.50471961e-01
-9.96422112e-01 2.25109030e-02 -7.71696866e-01 -7.69301653e-01
2.59160340e-01 5.56843996e-01 -1.40461206e+00 1.97000548e-01
1.02526355e+00 5.81041873e-01 1.59961596e-01 1.52456033e+00
-3.74816984e-01 6.22134447e-01 -1.04021084e+00 -4.36504364e-01
7.39791214e-01 -5.14187694e-01 8.99365664e-01 9.75516915e-01
1.15913916e+00 5.56913435e-01 -9.99773517e-02 2.40942508e-01
4.13824022e-01 5.40836854e-03 -1.05078495e+00 5.68294644e-01
4.60448474e-01 1.24062645e+00 -5.71149111e-01 1.02252206e-02
-2.38209561e-01 3.83782119e-01 -7.86435544e-01 1.69081613e-01
-4.56245020e-02 -8.66081595e-01 6.78867042e-01 3.74809057e-01
5.37328780e-01 -7.79686332e-01 5.28473146e-02 -5.94149113e-01
-3.87919575e-01 -1.92236230e-01 -3.72657254e-02 -9.53268290e-01
-6.93626404e-01 3.35723549e-01 -6.84254691e-02 -1.83052921e+00
-2.51678467e-01 -7.19125211e-01 -7.75537610e-01 8.45994174e-01
-1.98228562e+00 -8.75908613e-01 -8.40129972e-01 1.50578082e-01
6.12445951e-01 -3.07915173e-02 8.93367648e-01 -1.12183340e-01
4.00182396e-01 -1.64749637e-01 4.79967922e-01 2.16873258e-01
2.13213474e-01 -1.37835634e+00 1.69183865e-01 8.23251486e-01
-1.41831152e-02 9.48508605e-02 1.22456670e+00 -7.18525052e-01
-1.95558059e+00 -8.18270028e-01 4.65187222e-01 1.00808688e-01
5.37197828e-01 -1.79536462e-01 -9.61106062e-01 3.44020098e-01
-1.52288318e-01 3.32299948e-01 5.73350251e-01 -4.46540892e-01
4.20425594e-01 -4.02884126e-01 -1.23714352e+00 1.46020442e-01
4.61459547e-01 3.39057446e-01 -7.30471969e-01 1.21508762e-01
8.83920491e-02 -9.55673397e-01 -4.63004827e-01 5.28719246e-01
6.46428525e-01 -1.06877863e+00 5.82946122e-01 2.16998637e-01
2.32834116e-01 -6.42501533e-01 -2.46357217e-01 -1.78295577e+00
3.81896049e-01 -5.59408844e-01 6.25850141e-01 5.22630095e-01
2.89396226e-01 -9.80149984e-01 9.95304286e-01 -1.37139246e-01
-8.09045076e-01 -2.41826639e-01 -1.21745098e+00 -5.57620883e-01
-2.12103873e-01 -4.31126565e-01 2.78766364e-01 5.48961818e-01
-3.74571651e-01 -6.75694197e-02 -1.93313673e-01 1.25241733e+00
8.42291415e-01 1.40134200e-01 1.09871209e+00 -1.58326721e+00
1.98947534e-01 4.18379813e-01 -8.84047627e-01 -1.11390150e+00
-2.53405362e-01 -1.77047729e-01 8.02097142e-01 -1.66159260e+00
-7.66704142e-01 -7.39647329e-01 4.89533275e-01 4.38452750e-01
6.07641101e-01 6.67238951e-01 -3.26243013e-01 2.03656688e-01
1.98190972e-01 4.94555026e-01 1.08892179e+00 2.28301749e-01
-4.37683940e-01 3.71051073e-01 2.73823470e-01 1.14750528e+00
5.09568572e-01 -5.75312972e-01 -2.17469424e-01 -8.36529076e-01
6.72719836e-01 3.35807204e-01 2.61723250e-01 -1.27591491e+00
4.45481420e-01 5.78456149e-02 2.03691021e-01 -9.29696262e-01
7.33024299e-01 -1.27244711e+00 3.70237678e-02 8.45187128e-01
5.42296886e-01 -2.01394215e-01 8.69567841e-02 6.17973328e-01
-5.65857589e-01 -8.85551810e-01 7.53193736e-01 -4.56937194e-01
-9.70966339e-01 9.40528512e-02 -6.62418365e-01 -4.10091072e-01
6.15269423e-01 -7.07036734e-01 -5.20729199e-02 -7.02379346e-01
-1.50893569e-01 3.37042749e-01 5.29415905e-01 -2.08120376e-01
1.38693964e+00 -4.89813834e-01 -9.98117149e-01 7.41535842e-01
-6.54423535e-02 8.87662768e-01 2.47960463e-01 5.92470467e-01
-1.61360729e+00 -1.09320335e-01 -7.67879263e-02 -1.09544349e+00
-1.25271606e+00 -4.22410548e-01 6.74804151e-01 6.82044327e-01
-9.62549627e-01 1.04214013e+00 -1.35128528e-01 -4.18987751e-01
-3.49095225e-01 -6.03199601e-01 -3.34301800e-01 -1.13949236e-02
5.63258469e-01 2.07383454e-01 1.25132322e-01 -6.80303097e-01
-1.60622418e-01 1.16789782e+00 5.79785407e-01 -1.38262466e-01
1.75421882e+00 -1.07344210e-01 -1.05228601e-02 2.33415052e-01
9.26912069e-01 2.33581990e-01 -1.55062544e+00 -6.15829080e-02
-3.82268965e-01 -7.68479288e-01 4.11591917e-01 -1.25978708e-01
-1.01258624e+00 1.13269103e+00 6.08784676e-01 4.11289126e-01
8.89673948e-01 -1.50957480e-01 7.47859836e-01 9.24331486e-01
4.84340012e-01 -8.14143419e-01 -7.27169141e-02 9.01341677e-01
8.54786098e-01 -1.27140892e+00 5.80575056e-02 -1.67546654e-03
-2.36388177e-01 1.51381361e+00 2.66422629e-01 -5.53152978e-01
8.11689258e-01 6.01138592e-01 6.26055300e-01 -1.82346210e-01
-1.27906606e-01 -1.30951181e-01 -4.14246827e-01 7.94159114e-01
-2.80942857e-01 -1.10938936e-01 -1.17334742e-02 -4.72750552e-02
-5.41641533e-01 -5.09187579e-01 1.29093671e+00 1.27945340e+00
-1.10535920e+00 -4.10665423e-01 -6.87386811e-01 2.00514019e-01
-3.70401740e-01 -1.08510643e-01 3.53345603e-01 8.42216313e-01
1.42531604e-01 6.43238068e-01 6.65623248e-01 -2.34474968e-02
3.35143954e-01 -4.29781377e-01 -2.09769495e-02 -7.44715273e-01
-1.01072848e-01 2.94044912e-01 9.87441167e-02 2.57485211e-02
-5.29651523e-01 -7.38594174e-01 -1.67793143e+00 3.64634633e-01
-4.29079860e-01 7.10059166e-01 1.50890708e+00 8.35270643e-01
-1.46428287e-01 -1.79244250e-01 9.52088535e-01 -1.45297217e+00
-2.99328536e-01 -1.19018793e+00 -1.04360366e+00 -2.48111114e-01
7.54637301e-01 -7.35450327e-01 -9.48508084e-01 -6.74251094e-02] | [7.4679741859436035, -1.7527503967285156] |
a5c3abe4-9c13-4d94-8f51-d171012f39ad | physics-informed-machine-learning-with | 2206.10718 | null | https://arxiv.org/abs/2206.10718v1 | https://arxiv.org/pdf/2206.10718v1.pdf | Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management | Avoiding over-pressurization in subsurface reservoirs is critical for applications like CO2 sequestration and wastewater injection. Managing the pressures by controlling injection/extraction are challenging because of complex heterogeneity in the subsurface. The heterogeneity typically requires high-fidelity physics-based models to make predictions on CO$_2$ fate. Furthermore, characterizing the heterogeneity accurately is fraught with parametric uncertainty. Accounting for both, heterogeneity and uncertainty, makes this a computationally-intensive problem challenging for current reservoir simulators. To tackle this, we use differentiable programming with a full-physics model and machine learning to determine the fluid extraction rates that prevent over-pressurization at critical reservoir locations. We use DPFEHM framework, which has trustworthy physics based on the standard two-point flux finite volume discretization and is also automatically differentiable like machine learning models. Our physics-informed machine learning framework uses convolutional neural networks to learn an appropriate extraction rate based on the permeability field. We also perform a hyperparameter search to improve the model's accuracy. Training and testing scenarios are executed to evaluate the feasibility of using physics-informed machine learning to manage reservoir pressures. We constructed and tested a sufficiently accurate simulator that is 400000 times faster than the underlying physics-based simulator, allowing for near real-time analysis and robust uncertainty quantification. | ['Hari Viswanathan', 'Dylan Robert Harp', "Daniel O'Malley", 'Aleksandra Pachalieva'] | 2022-06-21 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-3.33608925e-01 -6.38901070e-02 9.71905217e-02 3.22630256e-02
-7.43888438e-01 -4.51350808e-01 5.84974885e-01 4.98139203e-01
-1.15445167e-01 1.07807553e+00 -1.15783051e-01 -7.78327942e-01
-2.31211215e-01 -1.21286714e+00 -1.09170318e+00 -6.49950981e-01
-7.58050621e-01 8.07509422e-01 1.14871249e-01 -2.45323941e-01
5.39796293e-01 8.12812090e-01 -1.42893755e+00 6.05029017e-02
1.20879936e+00 1.01615989e+00 1.94898978e-01 6.86267376e-01
-2.37487659e-01 4.93756115e-01 -1.59229085e-01 5.12361646e-01
3.00205588e-01 1.22168921e-02 -6.06190264e-01 -6.25665069e-01
-1.50504634e-01 -5.08922100e-01 -1.01347543e-01 7.87372828e-01
3.49357337e-01 1.21046841e-01 1.08897483e+00 -7.47910380e-01
-1.27052665e-01 5.24505138e-01 -4.23971593e-01 1.14119500e-01
-2.45992020e-01 5.14982700e-01 5.13405204e-01 -9.71650839e-01
-1.07990004e-01 1.04173887e+00 8.81295621e-01 5.43241240e-02
-1.31728911e+00 -5.94935834e-01 -8.80979598e-02 -4.37629491e-01
-1.32004428e+00 -3.89749229e-01 3.08524311e-01 -1.02105892e+00
1.25768065e+00 1.37689086e-02 9.58602965e-01 2.98384607e-01
6.17286325e-01 5.69128525e-03 1.22142005e+00 -1.27298132e-01
7.38336027e-01 7.72894844e-02 -4.24192041e-01 3.13563704e-01
4.29334998e-01 5.24778187e-01 -9.07072574e-02 -4.13513839e-01
9.87081647e-01 -2.45148256e-01 -4.27699029e-01 -2.12447811e-03
-4.90175754e-01 8.75789702e-01 5.12840331e-01 -1.35545477e-01
-3.58352363e-01 5.61290860e-01 2.58873075e-01 -1.77409314e-02
3.31961840e-01 9.25824225e-01 -6.76884592e-01 -1.84160024e-01
-1.08798182e+00 7.12289572e-01 1.11756372e+00 8.11485350e-01
8.74537051e-01 3.86148930e-01 -6.09032549e-02 4.83651936e-01
6.87528849e-01 1.05002582e+00 1.35055995e-02 -1.14395607e+00
2.34673351e-01 2.68838614e-01 6.09828711e-01 -5.92268646e-01
-2.17753366e-01 3.50483917e-02 -7.41320312e-01 5.12573957e-01
3.87417346e-01 -4.19204712e-01 -1.07600939e+00 1.15041518e+00
8.40082541e-02 1.81844875e-01 1.31212533e-01 6.83758497e-01
2.69425720e-01 9.51747835e-01 4.72594082e-01 -3.39019448e-02
1.15600216e+00 -3.07827741e-01 -5.07690728e-01 -9.08560306e-02
5.69730878e-01 -1.64790496e-01 9.68483567e-01 -2.86397133e-02
-1.12471163e+00 2.36685798e-01 -1.09014618e+00 2.18721628e-01
-5.51685929e-01 -3.07371199e-01 7.28717566e-01 3.90012234e-01
-6.42884135e-01 1.27164757e+00 -1.20892406e+00 1.55449435e-01
4.24437910e-01 2.95297027e-01 7.07700104e-02 3.84003580e-01
-1.56345141e+00 1.36466897e+00 1.89625174e-01 5.35282671e-01
-1.15491104e+00 -1.41100824e+00 -9.97858465e-01 2.69570261e-01
-2.07788736e-01 -3.98939878e-01 1.36044681e+00 2.74457899e-03
-1.92349494e+00 2.71011386e-02 -2.26213573e-03 -4.23047870e-01
8.58310819e-01 -7.99696371e-02 7.19699338e-02 2.30659153e-02
-1.07402988e-01 1.33531407e-01 4.53080475e-01 -1.38085985e+00
-1.60707191e-01 2.28027012e-02 -1.97471350e-01 1.77352712e-01
1.11555390e-01 -3.43530029e-01 1.27563253e-02 -4.35197949e-02
-2.45283283e-02 -6.29410923e-01 -4.52071816e-01 1.05079792e-01
-1.00108676e-01 3.26533914e-01 7.07720637e-01 -8.59481990e-01
7.05723822e-01 -1.48433983e+00 -1.48230478e-01 4.69677657e-01
-1.84310064e-01 3.39221247e-02 3.62263680e-01 6.38120890e-01
2.52276868e-01 5.67819357e-01 -9.30484474e-01 7.38852937e-03
2.90510897e-02 2.08147675e-01 -3.33418816e-01 6.47286832e-01
7.24637628e-01 6.94769561e-01 -8.99868429e-01 -2.25691840e-01
6.18497491e-01 5.25831997e-01 -6.51143193e-01 3.52396607e-01
-5.76632202e-01 6.96533144e-01 -5.26906729e-01 5.77643275e-01
1.02530932e+00 -1.56432956e-01 7.92100057e-02 -2.01160554e-02
-6.57729447e-01 4.51385789e-02 -1.16050601e+00 1.08039725e+00
-1.10291946e+00 2.70867497e-01 5.73374748e-01 -8.05451810e-01
9.67935026e-01 1.62482802e-02 3.82951111e-01 -7.35170543e-01
1.37727216e-01 7.58920670e-01 -1.99820325e-01 -6.87256217e-01
6.68092132e-01 -7.79883683e-01 2.02037483e-01 7.78427124e-02
-4.27733779e-01 -1.08758926e+00 -4.13953096e-01 -1.77746862e-01
7.79070735e-01 2.08887294e-01 -2.79041290e-01 -1.18522263e+00
3.16633373e-01 2.60752916e-01 3.23291719e-01 6.21851206e-01
1.37637511e-01 4.06397820e-01 4.52321827e-01 -1.76266521e-01
-1.34228802e+00 -7.76336551e-01 -8.43186080e-01 2.57331461e-01
4.28441256e-01 2.50213534e-01 -3.22436243e-01 2.06255227e-01
7.69126952e-01 7.73196340e-01 -4.99167353e-01 -3.34064849e-02
-5.65139472e-01 -1.02849734e+00 4.05742317e-01 7.04669595e-01
4.97152954e-01 -5.94357848e-01 -5.17467797e-01 5.91917813e-01
4.49757844e-01 -7.41325557e-01 3.53021920e-01 3.93558502e-01
-9.79652345e-01 -1.09191442e+00 -5.63887775e-01 -1.15060009e-01
4.44248021e-01 -5.21679282e-01 1.04522955e+00 2.76506901e-01
-8.36608931e-02 -1.76071897e-01 1.26737460e-01 -5.64197421e-01
-7.12361097e-01 -1.68509379e-01 -1.27653137e-01 -7.60790527e-01
-1.97196007e-01 -4.31645393e-01 -8.18687737e-01 2.41749212e-01
-6.83133602e-01 -1.61708042e-01 3.12682092e-01 6.37620032e-01
5.96221089e-01 3.82625729e-01 5.74460566e-01 -7.64312148e-01
5.94244540e-01 -9.30553079e-01 -1.15315044e+00 1.47351280e-01
-6.07456505e-01 2.66396493e-01 7.05697894e-01 -1.47467449e-01
-1.12109280e+00 -1.89617932e-01 -1.49751902e-01 -3.35474610e-01
2.75211126e-01 9.12233472e-01 1.14479689e-02 -2.79001981e-01
5.46797454e-01 -7.28067532e-02 1.48894545e-03 -3.78004104e-01
-1.09491341e-01 7.64958501e-01 2.31572345e-01 -1.28162336e+00
6.20161355e-01 3.32659751e-01 2.76897132e-01 -1.16178226e+00
-1.31818786e-01 1.06468923e-01 -2.52678663e-01 -4.26204205e-02
3.76259297e-01 -1.21663809e+00 -9.57041264e-01 6.42620325e-01
-9.27459598e-01 -1.03712952e+00 -1.68317229e-01 4.48763996e-01
-4.68052328e-01 -6.19903170e-02 -5.21545649e-01 -1.36269331e+00
-5.28501630e-01 -1.38371634e+00 1.06442368e+00 3.50062877e-01
1.44505471e-01 -1.10997403e+00 -3.24680693e-02 -3.39419127e-01
1.07426822e+00 5.96468449e-01 9.57609892e-01 -1.36161357e-01
-6.87971294e-01 5.11276834e-02 -2.79547364e-01 1.03120971e-02
-9.16281063e-03 4.31866378e-01 -1.08802819e+00 -2.04830453e-01
-8.47630494e-04 -1.98078409e-01 8.82428467e-01 6.93451762e-01
1.27893281e+00 -2.65705764e-01 -4.34891641e-01 6.35602236e-01
1.65600967e+00 1.08319312e-01 6.46151662e-01 4.43928242e-01
4.38750029e-01 4.91973460e-01 2.87107885e-01 7.51130223e-01
4.42746311e-01 1.42718196e-01 5.03968596e-01 1.26570076e-01
3.61331791e-01 -2.46081874e-01 -1.50682434e-01 3.70530874e-01
-6.75097406e-02 -4.69227582e-02 -1.36488140e+00 5.57342768e-01
-1.40286124e+00 -6.57433212e-01 -1.71094716e-01 2.40382361e+00
1.13023376e+00 1.38077751e-01 -4.67006773e-01 -2.44260773e-01
6.14199162e-01 -2.37978354e-01 -7.48906016e-01 -5.40910661e-01
1.63860977e-01 2.23214611e-01 1.20049620e+00 8.13670635e-01
-7.88282335e-01 5.85691094e-01 6.53478670e+00 1.69145852e-01
-1.48648667e+00 -1.68681324e-01 6.70144796e-01 3.88234437e-01
-6.78462327e-01 2.51387715e-01 -9.01562929e-01 5.84913194e-01
1.36624181e+00 -1.89421520e-01 5.15472293e-01 5.16788840e-01
7.59755313e-01 -6.28315270e-01 -9.97697413e-01 2.66203076e-01
-8.36715639e-01 -1.84499359e+00 -1.77241743e-01 5.03901094e-02
6.28167570e-01 2.73400605e-01 -4.09435540e-01 4.64925885e-01
6.26955092e-01 -1.24357808e+00 8.82191479e-01 8.82960975e-01
1.10306430e+00 -4.94507670e-01 7.12904572e-01 2.70461529e-01
-1.17914069e+00 -1.13600038e-01 -2.59053349e-01 -2.33484581e-01
2.49042317e-01 8.87527823e-01 -8.25389266e-01 2.53140032e-01
7.72477090e-01 3.62995505e-01 1.85177565e-01 1.18357790e+00
1.07109308e-01 4.65825081e-01 -1.07830369e+00 -1.53893903e-01
2.60253120e-02 -2.47428924e-01 1.16286896e-01 1.01098573e+00
6.85477078e-01 3.11417937e-01 4.77118827e-02 1.42863464e+00
4.92131077e-02 -2.47651026e-01 -5.65203071e-01 -1.80433959e-01
9.85517621e-01 7.34699845e-01 -3.58456284e-01 -1.24316141e-01
1.73845425e-01 1.56904012e-01 1.26424944e-02 2.67945081e-01
-7.72195399e-01 -4.01426524e-01 8.22284162e-01 4.95759457e-01
6.69768676e-02 -4.57750112e-01 -5.05137205e-01 -7.53002584e-01
-2.27651477e-01 -1.78045243e-01 -1.37699515e-01 -5.92074156e-01
-1.28367174e+00 1.02440447e-01 3.61583471e-01 -8.59921575e-01
-6.33498048e-03 -7.66849935e-01 -9.79109108e-01 1.53126252e+00
-2.43424320e+00 -7.00323999e-01 -4.06763673e-01 -9.80609879e-02
9.62410271e-02 3.51101786e-01 8.24545860e-01 1.35648981e-01
-5.24179161e-01 1.36975825e-01 4.36845511e-01 -1.91461489e-01
1.35066584e-01 -1.38496995e+00 2.62788832e-01 3.61831844e-01
-1.43458593e+00 4.27570939e-01 1.01343858e+00 -1.12584996e+00
-2.17016506e+00 -1.31358123e+00 1.59722820e-01 -2.82297842e-02
1.05047178e+00 -3.63146164e-03 -1.38937891e+00 2.56365657e-01
-1.69205800e-01 3.93960595e-01 1.11845128e-01 -3.70495379e-01
2.60827482e-01 1.14501402e-01 -1.53030288e+00 1.86122730e-01
4.98669356e-01 -3.75463963e-01 -2.05732733e-01 3.93885702e-01
3.90147120e-01 -8.60574841e-01 -1.41339815e+00 8.76637459e-01
5.17119050e-01 -3.43010128e-01 6.16984904e-01 -4.65818495e-01
6.95021570e-01 -2.29137853e-01 -1.92488536e-01 -1.36888099e+00
1.22962512e-01 -4.42245871e-01 -2.51991481e-01 9.62744832e-01
6.23715460e-01 -7.70450890e-01 6.42507374e-01 1.27386951e+00
-1.13297462e-01 -8.21921587e-01 -9.65117693e-01 -6.63021266e-01
1.02421331e+00 -3.26008081e-01 9.91639376e-01 6.89703763e-01
-3.59053090e-02 -5.05086780e-01 6.40495941e-02 9.02260542e-01
7.22584963e-01 1.34141296e-01 3.72463703e-01 -1.21262503e+00
-9.22763944e-02 -4.39307749e-01 1.33957580e-01 -4.63270634e-01
1.22541443e-01 -5.36904991e-01 5.17127514e-01 -1.60069907e+00
-3.01616013e-01 -1.23524594e+00 6.17788173e-02 3.64939570e-01
7.74015933e-02 -3.56041998e-01 -3.36982220e-01 3.19158465e-01
5.68659484e-01 1.13499510e+00 1.27798140e+00 -2.38843068e-01
-6.34567082e-01 -3.61497998e-01 -1.26364723e-01 4.43625420e-01
9.20524955e-01 -2.88442791e-01 6.65526688e-02 -5.04731774e-01
1.79777980e-01 5.65067828e-01 3.55133235e-01 -9.84657407e-01
1.02343552e-01 -6.37106001e-01 4.40145165e-01 -3.85880142e-01
1.44354180e-01 -8.31637919e-01 3.01406920e-01 6.55148983e-01
3.20267603e-02 -3.84071678e-01 9.58647072e-01 5.12256324e-01
3.88383605e-02 -1.72560945e-01 1.04462576e+00 -4.81669158e-01
-5.55856705e-01 5.00576138e-01 -4.59642977e-01 -2.04908345e-02
9.46574450e-01 -5.98679809e-03 -4.79827464e-01 1.49179876e-01
-3.71044636e-01 7.77983963e-01 6.28168643e-01 -1.22468196e-01
4.49594885e-01 -8.21453929e-01 -6.70731425e-01 1.81145251e-01
-1.33351415e-01 7.46364653e-01 2.90975451e-01 3.83332342e-01
-1.33887827e+00 3.00822780e-02 2.00749040e-02 -6.80663228e-01
6.50019571e-02 -3.38890031e-02 1.30226707e+00 -9.71156359e-02
-6.24056995e-01 5.48229814e-01 -3.08407634e-01 -6.82593644e-01
-3.92826974e-01 -5.82747459e-01 4.55943532e-02 -1.95122048e-01
3.52188975e-01 3.52098823e-01 3.35634917e-01 -1.25971898e-01
-2.56947339e-01 3.91629785e-01 4.08466458e-01 6.88857064e-02
1.65526712e+00 1.95857957e-01 -1.43461451e-01 3.89806688e-01
1.03683448e+00 -2.74742097e-01 -1.72064388e+00 1.82337120e-01
-5.48320599e-02 -4.93999213e-01 6.26441121e-01 -8.39879334e-01
-1.00840914e+00 8.18414509e-01 3.43490899e-01 8.02289695e-02
4.66327786e-01 -4.53715891e-01 4.52470124e-01 4.48336214e-01
2.15073928e-01 -1.10117674e+00 -4.78033096e-01 7.06743658e-01
1.11545765e+00 -1.20914674e+00 2.20394120e-01 -2.06477538e-01
-1.60500154e-01 1.04996634e+00 7.87611187e-01 -1.92719743e-01
1.24157429e+00 9.83813763e-01 -2.58678589e-02 -3.59645724e-01
-2.75966734e-01 6.32764101e-01 -4.87182438e-01 8.70513991e-02
6.76935688e-02 1.64217308e-01 4.99086119e-02 3.27746958e-01
3.92889008e-02 1.57934189e-01 5.72612643e-01 1.25278926e+00
-5.81679285e-01 -5.86977661e-01 -4.32778239e-01 7.27175593e-01
-1.60722792e-01 -1.08482487e-01 5.17789602e-01 6.29404664e-01
-2.44179904e-01 6.98337615e-01 2.81731993e-01 1.31822512e-01
4.70417321e-01 -1.69567153e-01 8.28658119e-02 -5.70063293e-01
-3.23354065e-01 -2.39418775e-01 -2.00746972e-02 -2.23638386e-01
-1.56419966e-02 -5.03030300e-01 -1.85071933e+00 -6.00808978e-01
-4.36877757e-01 5.41922629e-01 1.14015198e+00 9.01467085e-01
3.29695284e-01 3.66449326e-01 6.80574179e-01 -1.52228296e+00
-1.03048408e+00 -1.03596473e+00 -1.00529993e+00 -1.44686267e-01
4.96854931e-01 -1.19673896e+00 -7.40500093e-01 -5.20922244e-01] | [6.438438415527344, 3.246082067489624] |
2153ea79-4f44-4581-b49b-9b101586756d | deep-mds-framework-for-recovering-the-3d | 2210.15200 | null | https://arxiv.org/abs/2210.15200v1 | https://arxiv.org/pdf/2210.15200v1.pdf | Deep-MDS Framework for Recovering the 3D Shape of 2D Landmarks from a Single Image | In this paper, a low parameter deep learning framework utilizing the Non-metric Multi-Dimensional scaling (NMDS) method, is proposed to recover the 3D shape of 2D landmarks on a human face, in a single input image. Hence, NMDS approach is used for the first time to establish a mapping from a 2D landmark space to the corresponding 3D shape space. A deep neural network learns the pairwise dissimilarity among 2D landmarks, used by NMDS approach, whose objective is to learn the pairwise 3D Euclidean distance of the corresponding 2D landmarks on the input image. This scheme results in a symmetric dissimilarity matrix, with the rank larger than 2, leading the NMDS approach toward appropriately recovering the 3D shape of corresponding 2D landmarks. In the case of posed images and complex image formation processes like perspective projection which causes occlusion in the input image, we consider an autoencoder component in the proposed framework, as an occlusion removal part, which turns different input views of the human face into a profile view. The results of a performance evaluation using different synthetic and real-world human face datasets, including Besel Face Model (BFM), CelebA, CoMA - FLAME, and CASIA-3D, indicates the comparable performance of the proposed framework, despite its small number of training parameters, with the related state-of-the-art and powerful 3D reconstruction methods from the literature, in terms of efficiency and accuracy. | ['Zohreh Azimifar', 'Shima Kamyab'] | 2022-10-27 | null | null | null | null | ['face-model'] | ['computer-vision'] | [-1.90087646e-01 3.48086916e-02 4.19597656e-01 -4.01706815e-01
-1.54099271e-01 -1.51384294e-01 7.69252717e-01 -4.04924065e-01
-3.40417236e-01 3.16841573e-01 1.69963434e-01 1.75786287e-01
-3.75911891e-01 -7.75947988e-01 -6.14078224e-01 -8.63421500e-01
-7.92325195e-03 6.95851088e-01 -4.75071728e-01 -8.72000530e-02
1.79497048e-01 1.16644645e+00 -1.81183231e+00 -2.80931979e-01
1.96483552e-01 1.09612787e+00 -1.19249091e-01 7.72134438e-02
-6.51628003e-02 2.74002422e-02 -3.69757831e-01 -5.64602256e-01
8.12802136e-01 -1.17528126e-01 -2.98529238e-01 2.47338310e-01
7.40777254e-01 -5.17835259e-01 -3.67977947e-01 9.42454040e-01
7.74175823e-01 1.06943980e-01 1.04230130e+00 -1.18602300e+00
-8.92087877e-01 -1.30402729e-01 -9.41330314e-01 -1.30775496e-01
5.01235187e-01 -3.83226901e-01 4.89504039e-01 -1.43884254e+00
7.02295840e-01 1.60006535e+00 7.28045225e-01 4.91263330e-01
-1.01508784e+00 -5.59781194e-01 -3.86328846e-01 1.99506417e-01
-1.72066975e+00 -4.75351274e-01 1.07421947e+00 -5.97752810e-01
4.77891207e-01 -2.34551039e-02 4.75244403e-01 8.42028379e-01
1.32798105e-01 2.50374496e-01 1.21819758e+00 -5.51357567e-01
1.15400687e-01 7.13979676e-02 -2.86399603e-01 8.29998434e-01
1.82185590e-01 1.87920868e-01 -4.07603621e-01 -1.40345737e-01
9.82770145e-01 2.27714464e-01 -7.20179379e-02 -6.09531522e-01
-7.59218276e-01 7.89597809e-01 4.14278746e-01 4.58748251e-01
-5.76632679e-01 -5.18508673e-01 8.29488561e-02 7.64171854e-02
4.88581777e-01 -1.09182797e-01 -1.56171158e-01 3.97203684e-01
-7.74187028e-01 1.08117767e-01 7.06843436e-01 7.82715261e-01
8.18176925e-01 1.43693343e-01 1.66907877e-01 6.75764561e-01
7.00965762e-01 5.38565755e-01 5.38685441e-01 -9.11770225e-01
1.57158226e-01 8.29827726e-01 -8.18777010e-02 -1.63292897e+00
-4.70932901e-01 -2.43209168e-01 -1.10597587e+00 6.36624277e-01
1.31836772e-01 1.19304344e-01 -7.37244368e-01 1.89955878e+00
7.87521362e-01 2.62598336e-01 3.27915549e-01 9.73486245e-01
1.01447463e+00 4.65642005e-01 -4.19280440e-01 -3.41121316e-01
1.24283826e+00 -4.96800393e-01 -6.34107828e-01 1.72082871e-01
-5.56644723e-02 -9.00419176e-01 7.67322361e-01 2.42475227e-01
-9.98673797e-01 -8.19635689e-01 -1.12350202e+00 -1.37393057e-01
-4.65133101e-01 3.61956149e-01 1.77040592e-01 4.73314941e-01
-1.18891525e+00 4.33560610e-01 -4.77977246e-01 -4.34568822e-01
4.07447994e-01 4.36238110e-01 -1.00671494e+00 -1.51953280e-01
-8.16398203e-01 8.94835055e-01 1.55515283e-01 4.11664724e-01
-8.28596592e-01 -4.83703464e-01 -7.25700319e-01 6.05671518e-02
-1.90710220e-02 -4.87777859e-01 4.46667522e-01 -7.58933246e-01
-1.55092752e+00 1.18949246e+00 -4.00800481e-02 -8.09443300e-04
5.77261031e-01 -7.46396780e-02 -3.55914593e-01 1.68358073e-01
-4.97970320e-02 6.15941763e-01 1.21888936e+00 -1.46010959e+00
-9.48241279e-02 -1.02617455e+00 7.16712326e-02 3.88974249e-01
-3.29314291e-01 -5.43376617e-02 -3.96216094e-01 -4.19466943e-01
5.42972207e-01 -8.31847668e-01 1.87838301e-01 2.26674929e-01
-1.88244060e-01 -2.82959610e-01 9.57858264e-01 -7.76622534e-01
6.34330750e-01 -2.26619172e+00 5.79559624e-01 2.56700575e-01
1.01756923e-01 2.65292913e-01 -2.30190605e-01 2.94807374e-01
-3.00147474e-01 -2.10283741e-01 -3.33580852e-01 -7.32280135e-01
-2.19542682e-02 1.26450956e-01 1.71026707e-01 8.96011829e-01
2.57756691e-02 3.81977767e-01 -3.85525972e-01 -3.61037791e-01
3.57236534e-01 1.02162027e+00 -3.72274935e-01 4.12682176e-01
4.96340901e-01 5.04543185e-01 -1.76397666e-01 5.56585610e-01
1.33964992e+00 1.35055885e-01 -9.37908962e-02 -5.35557270e-01
-9.35032368e-02 -5.75117409e-01 -1.57436371e+00 1.69866383e+00
-4.39731896e-01 3.67802799e-01 2.20945030e-01 -9.77621853e-01
1.24481738e+00 5.23468614e-01 6.81229651e-01 -6.27627015e-01
3.03820521e-01 2.49472350e-01 -2.80235171e-01 -5.66455364e-01
-1.20305307e-01 -1.03558399e-01 5.83122671e-01 4.34794605e-01
2.76224077e-01 8.99162143e-02 -3.93062569e-02 -2.29554683e-01
4.25965041e-01 1.02454051e-01 4.25509065e-01 -2.97063291e-01
1.02433348e+00 -8.81571770e-01 3.82511199e-01 -1.10515468e-02
-5.49737066e-02 8.15253437e-01 2.32150972e-01 -7.41188943e-01
-1.29398978e+00 -9.77327108e-01 -4.56943661e-01 3.48380744e-01
4.15108576e-02 1.37070000e-01 -9.91444886e-01 -5.17190039e-01
2.12774381e-01 1.37659371e-01 -7.62361407e-01 3.88889830e-03
-7.49244153e-01 -5.67582667e-01 3.89986068e-01 1.36193484e-02
7.80364573e-01 -9.89559889e-01 -5.32789588e-01 -6.29850253e-02
1.97311863e-01 -1.12889957e+00 -2.06852078e-01 -3.74330282e-01
-7.53861189e-01 -1.30850255e+00 -8.60107183e-01 -9.06014919e-01
1.01149285e+00 1.96229637e-01 6.42423868e-01 -3.15704532e-02
-3.56028795e-01 3.97857457e-01 -2.60158032e-02 -1.87070742e-01
-1.77136242e-01 -6.64335966e-01 6.27045035e-01 6.23116672e-01
5.10012805e-01 -9.26341474e-01 -6.65044308e-01 3.68593246e-01
-8.96982253e-01 -1.58402547e-01 6.60867929e-01 7.52260566e-01
6.07795596e-01 4.97129150e-02 4.10564601e-01 -3.74897808e-01
3.93199563e-01 -4.27229077e-01 -6.60811603e-01 1.25472501e-01
-5.12706578e-01 -1.69013515e-01 6.79114163e-01 -2.36677825e-01
-9.11216140e-01 1.56139418e-01 -2.06714228e-01 -7.33034015e-01
-3.90329272e-01 3.37611943e-01 -5.14554262e-01 -3.11233461e-01
6.19917750e-01 3.75750005e-01 3.47861052e-01 -7.91671813e-01
2.70817041e-01 6.16725206e-01 3.28523219e-01 -3.52864742e-01
9.53972936e-01 7.94494212e-01 5.74454248e-01 -7.49980390e-01
-4.40231115e-01 -1.82415053e-01 -1.08383477e+00 -1.74038619e-01
8.03717077e-01 -9.98318613e-01 -8.64266276e-01 7.21256375e-01
-1.40577877e+00 4.14884955e-01 3.20561752e-02 5.92276216e-01
-5.68484128e-01 4.96574223e-01 -2.47370616e-01 -6.29625142e-01
-3.78645122e-01 -1.30940163e+00 1.14983380e+00 3.42086732e-01
3.31037849e-01 -8.22400868e-01 3.10146865e-02 3.46445441e-01
1.75990835e-01 4.72630739e-01 1.00199068e+00 -6.25416577e-01
-3.95733923e-01 -3.25824112e-01 -2.07858726e-01 6.67598367e-01
2.48335704e-01 -1.04992889e-01 -1.02296484e+00 -4.80352610e-01
5.15828490e-01 -7.04452619e-02 1.29468292e-01 2.87160009e-01
8.80356193e-01 -3.32552254e-01 1.37587935e-01 8.35291088e-01
1.54711354e+00 2.97246158e-01 3.13926190e-01 1.41474590e-01
6.40145659e-01 7.48615146e-01 3.41730744e-01 6.32490575e-01
2.58217633e-01 6.96311057e-01 6.03263021e-01 -1.68477103e-01
-2.88467169e-01 -2.14472443e-01 4.43540439e-02 9.33761835e-01
-2.55978048e-01 9.19030160e-02 -6.14117801e-01 1.67859912e-01
-1.40352917e+00 -6.73032820e-01 4.11813021e-01 2.31233716e+00
4.26193744e-01 -3.79973739e-01 -1.71613231e-01 3.58152509e-01
8.14240992e-01 9.53830183e-02 -5.73169827e-01 -2.05149576e-01
-2.37370670e-01 1.57896876e-01 9.09380056e-03 6.20433629e-01
-7.81672239e-01 6.38136625e-01 5.42717600e+00 7.00114310e-01
-1.28930330e+00 1.23953871e-01 3.95687103e-01 1.88293457e-01
-6.76247198e-03 -3.40955049e-01 -7.04493940e-01 3.13827455e-01
5.08216500e-01 1.83598790e-02 6.33562624e-01 7.76936352e-01
3.07465643e-01 2.22234294e-01 -1.22208858e+00 1.52320611e+00
5.84980309e-01 -1.03635097e+00 4.70251620e-01 3.35102499e-01
7.09483385e-01 -4.51426268e-01 2.43738890e-01 -1.04049139e-01
-3.56584817e-01 -1.06053376e+00 5.09198368e-01 6.90633833e-01
7.32807815e-01 -8.95968080e-01 7.03754544e-01 2.59223431e-01
-9.78759050e-01 -7.16833845e-02 -6.23706818e-01 9.79178324e-02
-5.60498536e-02 5.47433197e-01 -7.57108808e-01 7.07505167e-01
6.81524813e-01 6.30106330e-01 -4.75331545e-01 8.03918600e-01
2.53149439e-02 -1.38923794e-01 -3.47828269e-01 3.88411909e-01
4.85139415e-02 -7.45724857e-01 6.84294820e-01 6.05405033e-01
6.82840526e-01 2.82913268e-01 -2.34540761e-01 8.53291094e-01
-3.26090157e-01 4.88979310e-01 -9.99891877e-01 4.51360822e-01
5.42519450e-01 1.43818831e+00 -4.30791974e-01 -1.26387671e-01
-3.79001439e-01 8.00332546e-01 1.63631082e-01 3.93168777e-01
-6.48123682e-01 -1.01565406e-01 7.11333930e-01 1.46963932e-02
1.52067229e-01 -2.74128824e-01 -3.07171494e-02 -8.53020549e-01
3.29877108e-01 -8.95864725e-01 1.15925603e-01 -7.04740763e-01
-1.17183685e+00 9.03095245e-01 5.33784851e-02 -1.26827216e+00
-1.66630417e-01 -8.78786027e-01 -3.92831326e-01 9.95133519e-01
-1.37384951e+00 -1.22338641e+00 -4.48489249e-01 8.98930669e-01
3.62589747e-01 -7.20919847e-01 8.97983193e-01 6.41351163e-01
-4.22188133e-01 5.71878791e-01 2.52206028e-01 1.09470345e-01
5.95640421e-01 -8.54456842e-01 1.59697086e-01 6.22329772e-01
1.75830990e-01 7.46478975e-01 3.86874139e-01 -3.26757908e-01
-1.70689142e+00 -7.37747848e-01 6.29840016e-01 -2.24189103e-01
-1.61869992e-02 -2.23518655e-01 -6.63079679e-01 5.06302774e-01
4.58914898e-02 2.47976512e-01 6.14556730e-01 -2.38797754e-01
-3.80280703e-01 -3.56969982e-01 -1.61411095e+00 4.36807722e-01
1.05982006e+00 -5.79834819e-01 -6.35593116e-01 2.62172133e-01
3.03479910e-01 -3.18841130e-01 -1.13487411e+00 5.23710608e-01
8.29825461e-01 -1.32239640e+00 1.20837367e+00 -3.88034642e-01
1.99531510e-01 -4.81621951e-01 -4.50172186e-01 -1.20304477e+00
-2.60128230e-01 -2.07974553e-01 -1.19262017e-01 1.19110441e+00
-6.05469905e-02 -5.78175485e-01 7.21016526e-01 3.28430533e-01
7.60225430e-02 -7.94236124e-01 -1.39409685e+00 -4.08666074e-01
-2.14911953e-01 1.65633067e-01 9.25248265e-01 9.33827579e-01
-7.47455716e-01 1.10617392e-01 -4.14253980e-01 3.19706529e-01
8.54595602e-01 -2.86289714e-02 8.62081468e-01 -1.52849722e+00
7.57043734e-02 -2.47726545e-01 -8.50210667e-01 -7.90375590e-01
4.41193461e-01 -7.79717147e-01 -4.59170699e-01 -1.00733674e+00
-3.67301852e-02 -2.48728260e-01 -1.68653175e-01 1.25524968e-01
3.28271538e-01 4.23028201e-01 2.80171454e-01 2.13405460e-01
1.19141445e-01 8.41772079e-01 1.35355604e+00 -1.09406328e-02
-9.55637321e-02 -7.06917793e-02 -4.69431639e-01 8.97419930e-01
2.76352584e-01 -2.58485854e-01 -4.47510391e-01 -6.78739607e-01
-1.24408685e-01 6.37224466e-02 2.76743412e-01 -1.13293684e+00
1.93240866e-01 1.24597318e-01 7.32588112e-01 -6.98554158e-01
7.31971622e-01 -1.24704385e+00 3.74821037e-01 1.12542957e-01
-7.34811276e-02 4.03204650e-01 5.16046435e-02 3.54859620e-01
-3.27329546e-01 -3.35949928e-01 1.02727354e+00 -1.78903267e-01
-4.81874734e-01 7.60290563e-01 3.21257293e-01 -3.12883556e-01
1.05296183e+00 -4.68467951e-01 2.09487617e-01 -1.20136380e-01
-7.12209463e-01 -5.00740647e-01 3.77264857e-01 4.67148304e-01
9.31767285e-01 -1.77460361e+00 -9.44929600e-01 7.96836138e-01
-1.78551868e-01 1.64607987e-01 3.77612114e-01 6.26595974e-01
-6.80242419e-01 1.44357190e-01 -7.67617762e-01 -6.49419785e-01
-1.29747093e+00 5.82387030e-01 4.69121933e-01 3.10917348e-01
-6.12008154e-01 6.40099645e-01 2.82446474e-01 -6.63538635e-01
3.73802036e-01 1.77801549e-01 -5.04064381e-01 2.42173180e-01
5.40411294e-01 4.83412564e-01 2.55642772e-01 -1.35690963e+00
-3.41577053e-01 1.29642093e+00 2.37637937e-01 -1.56540960e-01
1.46048999e+00 -1.52895942e-01 -5.73983908e-01 5.50089106e-02
1.77500582e+00 2.13288814e-02 -1.09935403e+00 -3.12483281e-01
-2.47168630e-01 -7.68026173e-01 9.08746570e-03 -3.32374662e-01
-1.30363727e+00 9.20475900e-01 1.17137921e+00 -1.49051890e-01
1.16682434e+00 -2.99242407e-01 5.26988387e-01 3.18996847e-01
3.10706705e-01 -7.57594347e-01 1.96554974e-01 2.94673562e-01
1.37794578e+00 -1.20120144e+00 5.97483031e-02 -8.58748406e-02
-2.16864869e-01 1.36240613e+00 6.11027658e-01 -2.86419362e-01
1.04705226e+00 -1.54039070e-01 1.11954100e-01 -2.79255539e-01
-6.53039142e-02 1.77169561e-01 2.23764420e-01 6.22482896e-01
1.93453699e-01 -1.39938578e-01 -2.96523184e-01 1.00164860e-01
-2.20092461e-01 -9.48853344e-02 2.62691468e-01 4.31752175e-01
-1.03145234e-01 -9.59478676e-01 -6.28356636e-01 -6.57028034e-02
-2.27218777e-01 1.87167853e-01 -9.96024832e-02 9.24527645e-01
4.06024307e-01 6.29399776e-01 1.69123858e-01 -2.83033758e-01
4.89923269e-01 -3.57129145e-03 6.52659416e-01 -2.41698697e-01
-4.23573032e-02 -4.08877619e-02 -6.07463121e-01 -3.26605678e-01
-6.50921643e-01 -6.39829755e-01 -9.37418163e-01 -3.33406776e-01
5.20379916e-02 -4.14918065e-02 1.05675900e+00 7.45077729e-01
5.06197095e-01 -1.04907289e-01 1.13226926e+00 -1.19440627e+00
-6.75337732e-01 -9.93298829e-01 -6.87966824e-01 7.15907395e-01
4.41615850e-01 -1.14363539e+00 -4.85864371e-01 -7.83509687e-02] | [13.211397171020508, 0.3007567226886749] |
5d4af61e-f502-48be-9c1c-44aa282397aa | exbrainable-an-open-source-gui-for-cnn-based | 2201.04065 | null | https://arxiv.org/abs/2201.04065v1 | https://arxiv.org/pdf/2201.04065v1.pdf | ExBrainable: An Open-Source GUI for CNN-based EEG Decoding and Model Interpretation | We have developed a graphic user interface (GUI), ExBrainable, dedicated to convolutional neural networks (CNN) model training and visualization in electroencephalography (EEG) decoding. Available functions include model training, evaluation, and parameter visualization in terms of temporal and spatial representations. We demonstrate these functions using a well-studied public dataset of motor-imagery EEG and compare the results with existing knowledge of neuroscience. The primary objective of ExBrainable is to provide a fast, simplified, and user-friendly solution of EEG decoding for investigators across disciplines to leverage cutting-edge methods in brain/neuroscience research. | ['Chun-Shu Wei', 'Jian-Xue Huang', 'Chia-Ying Hsieh', 'Ya-Lin Huang'] | 2022-01-10 | null | null | null | null | ['eeg-decoding', 'eeg-decoding'] | ['medical', 'time-series'] | [-2.11562306e-01 -5.49043775e-01 4.08607692e-01 -5.15649438e-01
-1.46269664e-01 -3.76638025e-01 3.15487385e-01 -2.67089218e-01
-5.89179337e-01 8.00136626e-01 7.52555206e-02 -7.47267425e-01
-2.34478787e-01 -1.45750031e-01 -7.01154709e-01 -3.92752528e-01
-6.53739512e-01 2.05769405e-01 -3.48321766e-01 -3.24172266e-02
3.33992422e-01 7.64804006e-01 -1.30796123e+00 4.61669803e-01
6.16150379e-01 1.04525495e+00 4.99208868e-01 6.19481623e-01
4.67678726e-01 2.65037477e-01 -8.49868417e-01 1.95440844e-01
-2.80991793e-01 -2.85264194e-01 -5.95881164e-01 -6.79388404e-01
-9.76603329e-02 -3.00725460e-01 -3.92525911e-01 9.37881112e-01
8.22610021e-01 1.83574967e-02 4.77252245e-01 -1.14982259e+00
-6.38005972e-01 2.61635333e-01 -1.13144979e-01 9.71635878e-01
1.47227511e-01 5.02652466e-01 4.24857959e-02 -6.19480014e-01
5.17431617e-01 7.53838301e-01 5.79159558e-01 6.85405254e-01
-1.34134531e+00 -1.25492072e+00 -2.16697887e-01 5.69467962e-01
-1.40053010e+00 -4.29365486e-01 1.86468840e-01 -7.67072201e-01
1.60265684e+00 3.06275070e-01 1.38706791e+00 1.65361869e+00
7.67859578e-01 2.90772140e-01 1.19299138e+00 6.72108606e-02
4.28123921e-01 4.65996563e-02 5.88515222e-01 2.32544258e-01
-2.06339657e-02 3.34319949e-01 -1.07106900e+00 -8.98093805e-02
1.16456115e+00 -1.97499871e-01 -5.64652801e-01 4.30435836e-01
-1.16972768e+00 2.74920136e-01 3.87649089e-01 3.51583868e-01
-6.01720273e-01 5.21673501e-01 4.06142861e-01 2.11924389e-01
5.01428127e-01 6.92579389e-01 -6.42012537e-01 -8.35764289e-01
-1.16751957e+00 1.35596856e-01 5.39098918e-01 8.35918009e-01
1.99179992e-01 3.13641816e-01 -1.41326487e-01 6.73923075e-01
1.11449704e-01 1.38377279e-01 6.51545167e-01 -6.68762326e-01
1.72844648e-01 2.27307767e-01 -9.82332006e-02 -5.99212170e-01
-1.08899832e+00 -6.44826829e-01 -8.03489923e-01 4.05909449e-01
-8.09369385e-02 -3.74486178e-01 -8.42036724e-01 1.40625560e+00
-5.05207181e-01 3.72676253e-01 -4.81451511e-01 9.17763710e-01
1.14966166e+00 2.99291313e-01 4.28842753e-01 4.22615677e-01
1.48562312e+00 -4.02769893e-01 -6.87832773e-01 -1.71542019e-01
7.30141282e-01 3.48894671e-02 9.53423202e-01 6.98553860e-01
-1.09397614e+00 -2.57533640e-01 -1.17082500e+00 -3.58664617e-02
-7.65013158e-01 4.91557688e-01 7.61498332e-01 5.55958629e-01
-1.40944481e+00 8.55933309e-01 -1.40914035e+00 -3.19392949e-01
9.30609524e-01 8.29915583e-01 -7.22391129e-01 4.68650877e-01
-1.08243012e+00 1.32835352e+00 5.11132956e-01 5.23918234e-02
-1.07188475e+00 -1.03270423e+00 -4.46049631e-01 3.71863842e-01
-5.72673678e-01 -7.75520325e-01 1.08684742e+00 -6.95919216e-01
-1.24548697e+00 5.99985898e-01 -2.03279674e-01 -5.30081511e-01
7.94497058e-02 -1.99035197e-01 -5.41392446e-01 9.10169408e-02
-3.81398529e-01 9.22324181e-01 2.74187177e-01 -6.39629126e-01
-1.74921185e-01 -4.08098459e-01 -3.80265146e-01 6.21450432e-02
-2.97578961e-01 5.28136790e-01 -4.70432580e-01 -5.50676882e-01
-3.75085890e-01 -4.34723288e-01 1.90295592e-01 3.99357490e-02
-3.49214584e-01 1.41153082e-01 6.11925900e-01 -1.10713279e+00
1.13923049e+00 -2.20095158e+00 5.56143075e-02 1.84727252e-01
4.65944141e-01 8.07692483e-02 -3.67576554e-02 2.51649916e-01
-8.99009645e-01 1.68934718e-01 -2.96025537e-02 -2.72697896e-01
4.30286229e-02 -1.59158915e-01 -3.95681597e-02 5.83591700e-01
7.79861733e-02 1.17087185e+00 -5.28005481e-01 2.12972835e-01
4.08883393e-01 7.53604352e-01 -5.40922821e-01 1.05547391e-01
2.82801628e-01 8.16590011e-01 1.36087924e-01 5.84802032e-01
6.23753071e-01 -2.86676764e-01 2.38907058e-02 -7.77737871e-02
-3.56839657e-01 5.53605676e-01 -5.77721238e-01 1.82078528e+00
-3.25483829e-01 1.53890705e+00 -3.01985182e-02 -6.69662952e-01
3.21864456e-01 5.65143406e-01 2.73589611e-01 -8.32697213e-01
5.80187023e-01 -9.87126529e-02 2.19659165e-01 -6.40235007e-01
-2.21304834e-01 4.17522520e-01 7.68612564e-01 6.58174813e-01
5.99401057e-01 2.36098230e-01 1.01130307e-01 -6.49542809e-02
1.18772519e+00 2.25581393e-01 -7.57974852e-03 -6.30062580e-01
-2.92677075e-01 -3.01285923e-01 -2.41587102e-01 5.61406910e-01
9.14000645e-02 5.03349006e-01 6.97497904e-01 -4.75495279e-01
-6.94083333e-01 -7.26686895e-01 -5.58887601e-01 1.00862479e+00
-4.59126234e-01 -6.70909643e-01 -1.17324173e+00 1.27046943e-01
-3.33065897e-01 8.15464377e-01 -1.04031193e+00 -2.90281951e-01
-3.49137545e-01 -1.00566757e+00 5.79538465e-01 7.56059945e-01
3.05064410e-01 -1.37418258e+00 -1.14953279e+00 2.66727477e-01
2.48877823e-01 -8.74097586e-01 1.39197484e-01 7.94308126e-01
-9.05406713e-01 -1.16086257e+00 -6.42640531e-01 -5.74455440e-01
5.88282943e-01 -2.77101815e-01 8.92279148e-01 -3.66992019e-02
-8.09708834e-01 2.08760321e-01 1.21295964e-02 -6.84474289e-01
4.25870627e-01 -2.58938298e-02 6.42458647e-02 -5.86012244e-01
6.13850296e-01 -9.65740740e-01 -8.48367155e-01 4.32406254e-02
-7.43682206e-01 5.44219315e-01 4.02264774e-01 7.50363827e-01
2.21256137e-01 -4.69614446e-01 5.05615175e-01 -3.54962170e-01
1.15941417e+00 -6.76403940e-01 -7.27061868e-01 1.80722222e-01
-5.34160793e-01 -3.01873744e-01 5.07510900e-01 -4.38682765e-01
-5.09370446e-01 -2.69365042e-01 -5.17063797e-01 -2.98439205e-01
-4.99895334e-01 6.77334547e-01 2.15873525e-01 -3.73619527e-01
7.29809999e-01 4.83952612e-01 -3.45455915e-01 -5.57565868e-01
-1.12953819e-01 5.29888034e-01 7.92266369e-01 -2.75994360e-01
-1.94600984e-01 1.93665370e-01 -2.82757789e-01 -7.08380222e-01
2.25165248e-01 -1.21911831e-01 -6.13824129e-01 -4.24916983e-01
1.00683343e+00 -7.81808317e-01 -1.21422637e+00 3.69746387e-01
-1.32991874e+00 -6.86570466e-01 3.34001392e-01 8.10253739e-01
-5.71256399e-01 -2.80445129e-01 -5.37619114e-01 -4.82635498e-01
-6.60354018e-01 -1.40076292e+00 7.99368620e-01 1.00751467e-01
-6.00474834e-01 -1.01740336e+00 6.43747114e-03 -4.51365650e-01
7.45149374e-01 3.69674414e-02 1.10752904e+00 -6.49328411e-01
-3.08091819e-01 -9.33471695e-02 -3.98007661e-01 1.99098945e-01
-3.00495028e-01 -9.63745043e-02 -1.23635626e+00 -1.77825302e-01
-5.84019348e-02 -2.75569975e-01 4.19162750e-01 7.52024889e-01
1.98875952e+00 -2.30988469e-02 -5.66368520e-01 1.32994199e+00
1.02719021e+00 6.82194948e-01 8.11137021e-01 5.42622864e-01
1.69152007e-01 1.65825054e-01 -5.01526713e-01 5.26188076e-01
-1.09630913e-01 5.00073910e-01 3.35919619e-01 -3.22042227e-01
2.63249725e-01 2.34702632e-01 9.82309505e-03 3.78601313e-01
-3.26715827e-01 1.42692681e-02 -1.17790198e+00 1.09048329e-01
-1.43349135e+00 -6.72776759e-01 5.88897429e-02 1.75583136e+00
5.48419356e-01 -3.62925343e-02 -1.21322364e-01 -2.18418181e-01
3.30040187e-01 -5.55365622e-01 -6.63401604e-01 -2.56904244e-01
7.66539872e-02 8.73223424e-01 3.54347855e-01 -2.06029221e-01
-7.62843192e-01 5.51636934e-01 8.22474957e+00 6.48090661e-01
-1.47464788e+00 5.06818891e-01 4.92048651e-01 -7.58203745e-01
2.91591287e-01 -6.17541909e-01 -2.92482346e-01 7.04818308e-01
1.67233002e+00 -3.56044352e-01 1.05116355e+00 6.64149106e-01
7.40372479e-01 -1.92037597e-01 -1.25464869e+00 1.61289346e+00
-2.74940103e-01 -1.96430552e+00 -4.35559034e-01 -1.52473487e-02
7.59822875e-02 7.03425109e-01 2.83345014e-01 1.34394184e-01
-1.04984000e-01 -1.63091815e+00 7.62072384e-01 6.83521211e-01
1.44282293e+00 -7.80776083e-01 5.82557976e-01 1.55615494e-01
-7.04392970e-01 -4.04331051e-02 -1.43894538e-01 6.87451065e-02
2.47419570e-02 -8.21252093e-02 -5.36478281e-01 4.48778234e-02
1.17535257e+00 8.76751304e-01 -7.56914735e-01 1.54565167e+00
1.19580999e-01 7.53080487e-01 -1.45666227e-01 -9.94996503e-02
2.13882867e-02 6.31906912e-02 1.52584881e-01 1.62703252e+00
4.71195549e-01 2.37230077e-01 -8.97823095e-01 1.43331158e+00
-5.91217354e-02 -2.06822574e-01 -3.78908575e-01 -3.93298119e-01
3.43506724e-01 1.17957497e+00 -7.86387444e-01 -1.88440964e-01
-1.75328285e-01 7.95654416e-01 3.69454384e-01 6.19029045e-01
-8.80361557e-01 -5.48987567e-01 7.91504622e-01 -1.08955622e-01
-3.54086995e-01 -4.19041604e-01 -8.12497497e-01 -8.27634037e-01
-2.56496012e-01 -7.86584258e-01 -8.11812356e-02 -1.50661385e+00
-7.25642204e-01 9.66641068e-01 3.84230644e-01 -7.38104224e-01
-1.42160371e-01 -1.21221924e+00 -7.72027910e-01 1.37950349e+00
-1.01176810e+00 -6.19717658e-01 -3.69227588e-01 8.25617254e-01
3.63328785e-01 -1.85001180e-01 1.29774022e+00 5.32652438e-01
-8.84242952e-01 3.11832726e-01 2.75530159e-01 -2.66764984e-02
3.50386739e-01 -1.10936463e+00 7.59379506e-01 3.97073597e-01
-1.65674135e-01 1.17818737e+00 5.70307553e-01 -4.11009908e-01
-1.24710953e+00 -7.97644556e-01 2.90832371e-01 -3.58774036e-01
6.25393569e-01 -9.04622018e-01 -7.30046928e-01 1.02463984e+00
5.00032306e-01 -3.59372109e-01 1.03723073e+00 4.85204197e-02
9.78739634e-02 3.48104030e-01 -9.34866011e-01 7.31004298e-01
8.45649540e-01 -5.75241208e-01 -3.57621759e-01 5.18736362e-01
-1.58919729e-02 -5.32852173e-01 -6.82585835e-01 -7.23091215e-02
7.87914932e-01 -9.18340087e-01 6.95467293e-01 -6.99504018e-01
1.45464808e-01 2.54731625e-01 4.49754208e-01 -1.81020129e+00
-3.24438125e-01 -7.50756562e-01 -1.28121018e-01 2.16233969e-01
4.85786259e-01 -8.61998320e-01 3.47533584e-01 8.64393234e-01
-7.03527510e-01 -9.86679196e-01 -1.09854877e+00 -5.82693875e-01
-3.19251493e-02 -1.00311613e+00 7.15835989e-01 3.94356400e-01
7.23479867e-01 -1.56232983e-01 -1.19732752e-01 -1.06429882e-01
-3.56539413e-02 -7.04252362e-01 2.17480838e-01 -1.06724429e+00
8.91859978e-02 -7.45399415e-01 -5.34109473e-01 -8.35324764e-01
3.52168493e-02 -9.98911202e-01 -2.77152002e-01 -1.74385786e+00
2.41202727e-01 -1.78110693e-03 -4.45675224e-01 8.44853938e-01
4.81974095e-01 4.11653131e-01 -9.90603417e-02 9.30151194e-02
-2.17269138e-01 2.49088198e-01 9.36646640e-01 9.72419158e-02
-9.97073948e-02 -4.17949766e-01 -5.51826298e-01 3.72917920e-01
9.99358475e-01 -6.38845265e-01 -3.63834620e-01 -6.31055534e-01
3.53710167e-02 -8.48017782e-02 6.63506389e-01 -1.41672766e+00
3.90419573e-01 3.75914514e-01 1.14783895e+00 -5.38036644e-01
5.47771275e-01 -5.37105799e-01 5.02407253e-01 3.42278987e-01
-2.72930622e-01 5.50646305e-01 9.28658366e-01 5.13724163e-02
1.61614254e-01 2.05142409e-01 6.78964615e-01 -1.14702858e-01
-5.02660394e-01 3.02555948e-01 -9.25073266e-01 -3.57169420e-01
8.40297461e-01 -3.53393614e-01 -7.13416874e-01 -2.45711014e-01
-1.08644247e+00 1.23045426e-02 7.95775950e-02 3.20609987e-01
8.48779619e-01 -1.05402315e+00 -4.30074871e-01 8.68191957e-01
5.62239960e-02 -8.58659148e-01 4.22902346e-01 1.24615967e+00
-8.33096683e-01 9.69788730e-01 -1.08145535e+00 -4.98213261e-01
-1.02931547e+00 4.13682386e-02 7.75930107e-01 5.06067693e-01
-1.04720008e+00 8.65095794e-01 2.14266136e-01 7.93468207e-02
6.05071008e-01 -5.50882876e-01 -4.06336516e-01 -2.68123239e-01
1.02892220e+00 1.97108015e-01 7.17923522e-01 -1.96271792e-01
-4.32031751e-01 -2.39049897e-01 2.27703974e-01 -3.11303198e-01
1.77649021e+00 3.79993796e-01 -1.36410743e-01 3.92623723e-01
1.08001256e+00 -9.25128758e-01 -1.12773168e+00 7.30495512e-01
-2.71476179e-01 -1.01048894e-01 2.90542364e-01 -1.38014627e+00
-1.05824292e+00 1.27484417e+00 1.15540338e+00 -4.08101045e-02
1.00632119e+00 -2.98736691e-01 4.34948266e-01 3.84747475e-01
3.87903810e-01 -1.00431025e+00 -3.09938639e-01 4.95347619e-01
1.19239938e+00 -5.15315950e-01 -1.74128667e-01 4.99661833e-01
-4.00083959e-01 1.48385894e+00 6.41346812e-01 -1.05152145e-01
9.88089442e-01 5.60270071e-01 -6.23892173e-02 -7.56941259e-01
-8.82528007e-01 2.14476764e-01 5.85558057e-01 8.96619022e-01
5.04326582e-01 7.23137781e-02 -1.47925287e-01 1.24757755e+00
-5.11697769e-01 4.92783666e-01 1.15217872e-01 9.49210644e-01
-2.71664131e-02 -5.29901385e-01 -1.71127409e-01 9.16009486e-01
-5.87529659e-01 -5.74742556e-01 -8.34212825e-02 9.77359295e-01
4.99846116e-02 6.66592836e-01 2.35664740e-01 -5.20369053e-01
1.55699953e-01 2.71667123e-01 5.67291021e-01 -7.22716630e-01
-9.51782227e-01 -7.63012245e-02 -2.22463086e-02 -7.19018996e-01
2.68401712e-01 -3.09038639e-01 -1.14771676e+00 -3.67896467e-01
-1.05394855e-01 -1.33559331e-01 1.29010451e+00 9.36829209e-01
8.17208409e-01 1.03367424e+00 -4.07758474e-01 -1.44038582e+00
1.25415847e-01 -1.38477170e+00 -8.93806159e-01 -1.92897409e-01
2.53878117e-01 -7.98717022e-01 -3.16852272e-01 -1.08653512e-02] | [13.133481979370117, 3.449796438217163] |
5ea28d69-95ae-4023-ba8e-90d74981c68f | nanoflow-scalable-normalizing-flows-with | 2006.06280 | null | https://arxiv.org/abs/2006.06280v4 | https://arxiv.org/pdf/2006.06280v4.pdf | NanoFlow: Scalable Normalizing Flows with Sublinear Parameter Complexity | Normalizing flows (NFs) have become a prominent method for deep generative models that allow for an analytic probability density estimation and efficient synthesis. However, a flow-based network is considered to be inefficient in parameter complexity because of reduced expressiveness of bijective mapping, which renders the models unfeasibly expensive in terms of parameters. We present an alternative parameterization scheme called NanoFlow, which uses a single neural density estimator to model multiple transformation stages. Hence, we propose an efficient parameter decomposition method and the concept of flow indication embedding, which are key missing components that enable density estimation from a single neural network. Experiments performed on audio and image models confirm that our method provides a new parameter-efficient solution for scalable NFs with significant sublinear parameter complexity. | ['Sang-gil Lee', 'Sungwon Kim', 'Sungroh Yoon'] | 2020-06-11 | null | http://proceedings.neurips.cc/paper/2020/hash/a1c3ae6c49a89d92aef2d423dadb477f-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/a1c3ae6c49a89d92aef2d423dadb477f-Paper.pdf | neurips-2020-12 | ['normalising-flows'] | ['methodology'] | [-1.88564345e-01 4.17531170e-02 -3.60001802e-01 -1.34652749e-01
-5.15458882e-01 -4.51704741e-01 5.99490404e-01 -3.51274282e-01
-3.11880022e-01 9.46278751e-01 1.20682627e-01 -2.81572819e-01
-1.26156777e-01 -9.79262888e-01 -7.59641588e-01 -8.36161017e-01
1.53995484e-01 3.81231695e-01 -2.47516483e-02 1.53937489e-01
-2.22468320e-02 6.62879169e-01 -1.23612678e+00 -4.49546814e-01
8.71061742e-01 1.11022282e+00 1.53665647e-01 4.31916386e-01
-4.30934995e-01 6.94939137e-01 -5.12308478e-01 -6.83345437e-01
1.11975357e-01 -5.00918865e-01 -3.49013835e-01 -2.54641026e-01
4.60263193e-01 -8.75873387e-01 -4.98746157e-01 1.14491332e+00
3.82537186e-01 -2.93770749e-02 1.10208941e+00 -1.41450620e+00
-4.66595203e-01 7.69102216e-01 -4.33531314e-01 8.12671632e-02
-2.86615014e-01 -9.10780057e-02 1.04575694e+00 -8.16465437e-01
1.32575408e-01 1.33715534e+00 6.04714751e-01 6.53616607e-01
-1.51649058e+00 -8.75245154e-01 -1.88383758e-02 -3.64419648e-05
-1.68863451e+00 -6.63598835e-01 7.42740273e-01 -3.72550726e-01
6.61870182e-01 -4.70805503e-02 7.89025962e-01 1.12202930e+00
2.09054574e-01 6.78245127e-01 5.72480023e-01 -1.20952807e-01
4.89149988e-01 3.77407163e-01 -1.65654004e-01 7.75537789e-01
5.25837004e-01 -1.26408949e-01 -4.54870522e-01 -1.58262953e-01
1.56050611e+00 -1.88574255e-01 -2.16165394e-01 -4.47680384e-01
-8.52985799e-01 1.03302193e+00 3.32453132e-01 6.21021092e-02
-2.12461770e-01 7.47700393e-01 3.00741911e-01 4.55561988e-02
5.04771769e-01 1.56287737e-02 -7.23650958e-03 -5.47995389e-01
-9.79756594e-01 2.25911990e-01 8.78657699e-01 9.30332601e-01
8.87160838e-01 6.19685829e-01 4.24526632e-02 7.92737365e-01
4.25704896e-01 6.84643507e-01 2.83906639e-01 -1.21908462e+00
2.80984968e-01 4.51564223e-01 -6.12386353e-02 -1.09901476e+00
-1.34661451e-01 -6.88623846e-01 -1.33662641e+00 -1.20735392e-01
4.18066412e-01 -9.94120911e-02 -6.16464734e-01 2.11420178e+00
2.71169603e-01 4.15898889e-01 -1.20909289e-01 6.27306461e-01
4.13599312e-01 9.22563136e-01 3.17743123e-02 -1.61460817e-01
1.10847580e+00 -6.70814872e-01 -8.16327155e-01 1.27976790e-01
4.32004035e-01 -4.56718594e-01 1.22509801e+00 3.42134982e-01
-1.22879612e+00 -3.64035606e-01 -1.19577622e+00 -7.66938701e-02
-3.03208455e-02 9.03360918e-03 8.37157905e-01 1.03025508e+00
-1.13997865e+00 5.59918046e-01 -9.63992059e-01 4.03700247e-02
6.07074022e-01 4.62609351e-01 -9.96725932e-02 3.19702059e-01
-9.93217170e-01 5.32067478e-01 2.89672554e-01 6.86129108e-02
-9.71723080e-01 -1.07422554e+00 -8.70725453e-01 6.03686690e-01
5.83395502e-03 -9.27734792e-01 9.80174839e-01 -5.25407732e-01
-2.02457738e+00 6.76381662e-02 -1.96351483e-01 -5.23176670e-01
5.09495318e-01 -1.53838456e-01 -5.52041456e-02 3.22481632e-01
-2.73998946e-01 8.08406055e-01 1.19752920e+00 -9.71219540e-01
-2.08915800e-01 1.00857638e-01 -3.79385916e-03 4.85294266e-03
-9.90122139e-01 -2.56490827e-01 -1.73110291e-01 -6.29397988e-01
-1.03765093e-01 -5.33198595e-01 -4.79025207e-02 2.57761329e-01
-4.30071235e-01 -4.95956875e-02 5.59500933e-01 -4.09919977e-01
1.46747363e+00 -2.10455561e+00 1.89698264e-01 3.21817964e-01
4.65145558e-01 9.01169926e-02 3.10827959e-02 3.07988107e-01
2.41658807e-01 3.17029566e-01 -2.91310281e-01 -6.01366520e-01
2.44495749e-01 1.73427567e-01 -4.78292912e-01 4.68640685e-01
3.32216531e-01 8.23103368e-01 -5.84565699e-01 -5.52318513e-01
3.29697698e-01 9.87082720e-01 -9.13927972e-01 2.83274949e-01
-8.22959468e-02 1.86597124e-01 -2.83587933e-01 4.53983009e-01
8.84746134e-01 -3.56821805e-01 -6.21306263e-02 -3.79409075e-01
1.40249981e-02 1.84085310e-01 -1.46929049e+00 1.48799324e+00
-6.86276495e-01 3.17926526e-01 1.09899394e-01 -8.79300177e-01
9.94311571e-01 2.82252878e-01 3.71757060e-01 -1.96279958e-01
4.74586576e-01 3.42540592e-01 -2.32006118e-01 2.10711777e-01
4.95422304e-01 -4.01288390e-01 -3.77362780e-02 4.43135649e-01
5.14843583e-01 -1.82438880e-01 3.54836524e-01 3.33776861e-01
8.07181180e-01 -1.57428667e-01 2.27515519e-01 -5.38720667e-01
6.52800381e-01 -7.30281711e-01 6.17199421e-01 6.20914042e-01
-4.42136191e-02 5.66351116e-01 8.61154675e-01 -2.60785758e-01
-1.24499547e+00 -1.40340900e+00 -2.27791980e-01 5.24836004e-01
-6.07724898e-02 -6.57172680e-01 -9.41456854e-01 -2.22091869e-01
-2.21139625e-01 4.27847117e-01 -3.38157058e-01 -2.57767200e-01
-6.01381600e-01 -8.72012317e-01 6.87490880e-01 6.71992958e-01
5.93026936e-01 -5.98203659e-01 -1.90663084e-01 2.38908470e-01
-8.07884261e-02 -1.20621908e+00 -6.85355008e-01 -4.61668149e-02
-1.01270580e+00 -4.38395411e-01 -1.03365040e+00 -4.85676557e-01
5.44797897e-01 -2.02753618e-01 7.84708500e-01 -3.50809634e-01
-3.88637111e-02 1.82123542e-01 2.54435480e-01 1.93017093e-03
-6.44214451e-01 4.23268378e-01 2.57228255e-01 2.24312887e-01
-1.07209444e-01 -1.13553107e+00 -6.95677698e-01 2.21908495e-01
-9.62908506e-01 1.20249733e-01 5.21341681e-01 6.90086901e-01
4.42879975e-01 -1.97133735e-01 8.44278038e-01 -6.09211445e-01
6.94215059e-01 -4.51980084e-01 -9.74743307e-01 1.44219115e-01
-6.43412530e-01 2.90953845e-01 9.18978512e-01 -4.92010325e-01
-1.14010525e+00 -9.77485776e-02 -2.30292514e-01 -6.72864974e-01
2.56241679e-01 1.14857875e-01 -2.36307427e-01 -1.78028554e-01
3.17867875e-01 3.69260222e-01 1.42354980e-01 -4.88120526e-01
5.36533237e-01 3.78767669e-01 5.50589979e-01 -7.52750754e-01
8.84436965e-01 3.86184841e-01 4.30310845e-01 -8.77342045e-01
-3.33025038e-01 -8.17113891e-02 -3.29233110e-01 -2.00483680e-01
6.04829967e-01 -1.04518223e+00 -1.05221987e+00 4.74590212e-01
-1.24329138e+00 -2.99046785e-02 -4.77112204e-01 5.97644866e-01
-7.00350106e-01 3.89939755e-01 -8.44142318e-01 -9.48905170e-01
-3.93951088e-01 -1.07525492e+00 8.31742048e-01 4.30043459e-01
-2.59420741e-02 -1.09901512e+00 1.83360070e-01 -1.87994286e-01
8.30120802e-01 -1.37523875e-01 1.09981453e+00 -1.92325160e-01
-8.21451128e-01 6.47977786e-03 -4.61425573e-01 7.82249570e-01
1.05500974e-01 2.64310807e-01 -1.10136390e+00 -1.78333282e-01
-6.90167211e-03 -1.40770495e-01 6.59745038e-01 6.25317276e-01
1.22142088e+00 -4.09935534e-01 -5.63057847e-02 1.09986269e+00
1.54608595e+00 -3.07868663e-02 6.23508990e-01 -2.26843163e-01
8.35665226e-01 7.82132521e-02 -2.50237644e-01 6.97490573e-01
1.46624923e-01 3.48572969e-01 3.22156787e-01 1.39304996e-02
-2.46947870e-01 -5.06027102e-01 5.05379438e-01 1.42779100e+00
-4.02417369e-02 -3.76961648e-01 -4.17462260e-01 3.49898398e-01
-1.56604779e+00 -7.03153908e-01 3.36263895e-01 2.20293427e+00
8.03798676e-01 9.47391838e-02 2.43836507e-01 2.34761193e-01
7.78425217e-01 1.01901457e-01 -4.05670136e-01 -3.04535180e-01
6.47340044e-02 3.41006756e-01 3.84087354e-01 5.65108120e-01
-7.79720902e-01 8.91234159e-01 7.04420805e+00 1.21188796e+00
-9.20498073e-01 1.07544221e-01 5.98941803e-01 -1.47556439e-01
-5.99941075e-01 -1.13864519e-01 -1.06205332e+00 5.95614493e-01
1.24639404e+00 -3.73563826e-01 4.17716205e-01 9.12124634e-01
8.74416307e-02 1.21610440e-01 -9.86802936e-01 1.20431602e+00
-3.15444708e-01 -1.60582256e+00 5.31861246e-01 2.24074662e-01
4.17047739e-01 -3.62544060e-01 2.05506325e-01 7.61519596e-02
1.07950486e-01 -8.49008679e-01 7.62588322e-01 4.63575959e-01
8.96394432e-01 -1.13385570e+00 4.34702069e-01 1.68442518e-01
-1.23792362e+00 1.58602044e-01 -6.86784625e-01 1.00169957e-01
7.01253712e-01 7.12753594e-01 -6.57471478e-01 1.61516935e-01
3.32808286e-01 6.44380391e-01 -8.13480392e-02 9.54270899e-01
5.98801449e-02 7.49067783e-01 -5.50598323e-01 -1.16207980e-01
1.20233156e-01 -5.84542096e-01 4.73785400e-01 1.13174474e+00
6.81410968e-01 -4.48635489e-01 -3.68750244e-01 1.33516955e+00
-3.55895251e-01 1.31267950e-01 -4.36651796e-01 -3.02610606e-01
6.63472116e-01 1.32679105e+00 -8.40899050e-01 -1.68518692e-01
-2.76177347e-01 6.03229225e-01 2.78450102e-01 3.20284784e-01
-1.11754990e+00 -3.37831408e-01 7.98707962e-01 2.40875393e-01
2.50142068e-01 -3.66029829e-01 -1.06491052e-01 -1.29419184e+00
-1.08428903e-01 -3.00499976e-01 -9.54355672e-02 -2.33183607e-01
-1.13057888e+00 5.47092199e-01 1.60582945e-01 -1.11476934e+00
-4.27887768e-01 -5.40060639e-01 -3.16089839e-01 7.90428340e-01
-1.61361611e+00 -9.19666469e-01 -2.16266736e-01 5.82280397e-01
1.41703695e-01 -1.74516246e-01 7.92464137e-01 7.32842267e-01
-8.48793447e-01 9.25544143e-01 8.79837498e-02 -1.91589132e-01
2.94640064e-01 -1.12843180e+00 2.51468241e-01 7.59761035e-01
1.30931437e-01 7.50730515e-01 4.73791927e-01 -3.42162967e-01
-1.27484035e+00 -7.96617627e-01 4.91324574e-01 1.66899249e-01
7.81748414e-01 -6.30126476e-01 -8.04310858e-01 4.95714784e-01
6.06259815e-02 -8.08027387e-02 6.92267597e-01 -2.05670923e-01
-2.96203703e-01 -3.78427804e-01 -1.03099525e+00 5.59784353e-01
9.47201371e-01 -5.17805994e-01 6.33829907e-02 -5.26472591e-02
7.97458172e-01 -1.96286384e-02 -9.25883472e-01 2.21285716e-01
6.47592962e-01 -1.04199374e+00 1.02237177e+00 4.39553186e-02
2.49355122e-01 -1.17761634e-01 -2.22430527e-01 -9.95790660e-01
-2.92569131e-01 -1.02551651e+00 -5.75550199e-01 1.61720085e+00
2.29692012e-01 -8.88182163e-01 8.73350441e-01 4.28228289e-01
4.68205214e-02 -6.48492038e-01 -1.11801553e+00 -9.80747640e-01
1.85773253e-01 -4.34134334e-01 6.84171557e-01 4.96330649e-01
-3.37034523e-01 3.43168348e-01 -5.91403902e-01 -2.40711257e-01
6.96399689e-01 -3.64223182e-01 6.99338019e-01 -1.25958836e+00
-3.30021471e-01 -7.25422382e-01 -4.85053569e-01 -1.45089626e+00
2.63519257e-01 -7.27872729e-01 -1.91906527e-01 -1.16899252e+00
2.42347419e-01 -5.65286994e-01 -1.66731760e-01 -7.63172703e-03
2.79162049e-01 2.77165353e-01 -4.61582392e-02 6.95189461e-02
-8.97936970e-02 1.14789963e+00 1.17766523e+00 1.48570776e-01
-5.49870282e-02 4.90566716e-03 -6.59615338e-01 7.02337742e-01
6.46281064e-01 -4.42553878e-01 -9.77315307e-01 -3.08929205e-01
3.79835308e-01 -1.67067960e-01 2.52042323e-01 -1.11243534e+00
1.73248589e-01 -5.25731035e-02 1.60115466e-01 -3.52237552e-01
6.51007831e-01 -7.34813631e-01 1.98916316e-01 3.16572875e-01
-5.69063984e-02 2.07349770e-02 1.48014799e-01 5.45352936e-01
-2.00117618e-01 -2.96193361e-01 8.59915912e-01 2.15689555e-01
-2.01507673e-01 6.35149479e-01 -3.52503598e-01 1.02221213e-01
5.90619266e-01 -2.40905397e-02 -2.36388057e-01 -5.61698496e-01
-4.16678280e-01 -4.27494556e-01 3.13210279e-01 -9.63480026e-03
5.20856857e-01 -1.56806004e+00 -4.36619312e-01 3.81509274e-01
-4.64564890e-01 1.51165441e-01 3.22480202e-01 6.05727971e-01
-7.58395076e-01 4.01620418e-01 -2.16027260e-01 -6.71086252e-01
-5.31740010e-01 9.39059854e-02 3.35382849e-01 -1.99839368e-01
-6.05185688e-01 8.61905754e-01 6.07156873e-01 -1.26165673e-02
2.49855191e-01 -5.05320430e-01 1.33087203e-01 1.02741122e-01
5.97594857e-01 5.52179337e-01 -1.13637976e-01 -5.34671903e-01
-2.17227682e-01 5.68331718e-01 4.76197898e-02 -3.35160553e-01
1.19530249e+00 -1.66285008e-01 -6.63688406e-02 4.20185089e-01
1.40591633e+00 -2.81659693e-01 -1.63185596e+00 -1.59600481e-01
-5.69019794e-01 -3.98170203e-01 3.18721235e-01 -2.01288342e-01
-1.27867901e+00 1.11633432e+00 3.40453923e-01 1.62303925e-01
9.93090391e-01 -2.17556104e-01 8.82607520e-01 1.55054048e-01
1.88021153e-01 -9.21443343e-01 5.82280718e-02 3.37481856e-01
7.66206443e-01 -6.31811321e-01 -1.67096108e-01 -5.33139884e-01
-5.85643761e-02 1.16397750e+00 5.75309515e-01 -2.33988807e-01
9.45980787e-01 6.64703846e-01 -4.63923097e-01 2.71039158e-01
-5.58428168e-01 2.88284391e-01 1.13330215e-01 4.04107451e-01
2.52682775e-01 -1.26870155e-01 -5.03532961e-02 6.36737168e-01
-3.04650009e-01 -2.35393960e-02 5.88242292e-01 4.36024755e-01
-2.82534510e-01 -1.10103524e+00 9.80315208e-02 3.69922400e-01
-3.08582097e-01 -2.45207280e-01 3.76799434e-01 7.69059539e-01
-3.46646965e-01 4.35220003e-01 3.66650373e-01 -2.59801388e-01
-1.39951706e-04 1.35089904e-01 7.88974941e-01 -1.80325553e-01
-8.80079642e-02 2.63880640e-01 -3.23227346e-01 -5.38364053e-01
-2.34570652e-01 -2.84641355e-01 -1.06706917e+00 -7.48379767e-01
-3.81619155e-01 1.43840592e-02 7.33815670e-01 7.66556144e-01
4.11012918e-01 4.07549173e-01 6.88789070e-01 -8.03978741e-01
-6.87670231e-01 -8.64776194e-01 -8.67744267e-01 -1.05377950e-01
2.59603858e-01 -9.11613584e-01 -6.18943572e-01 -3.89533676e-02] | [7.2060112953186035, 3.82244610786438] |
5e5b1284-5a19-48eb-af7c-c4922a2cf115 | out-of-distribution-detection-with-distance | 2002.03328 | null | https://arxiv.org/abs/2002.03328v5 | https://arxiv.org/pdf/2002.03328v5.pdf | Kullback-Leibler Divergence-Based Out-of-Distribution Detection with Flow-Based Generative Models | Recent research has revealed that deep generative models including flow-based models and Variational Autoencoders may assign higher likelihoods to out-of-distribution (OOD) data than in-distribution (ID) data. However, we cannot sample OOD data from the model. This counterintuitive phenomenon has not been satisfactorily explained and brings obstacles to OOD detection with flow-based models. In this paper, we prove theorems to investigate the Kullback-Leibler divergence in flow-based model and give two explanations for the above phenomenon. Based on our theoretical analysis, we propose a new method \PADmethod\ to leverage KL divergence and local pixel dependence of representations to perform anomaly detection. Experimental results on prevalent benchmarks demonstrate the effectiveness and robustness of our method. For group anomaly detection, our method achieves 98.1\% AUROC on average with a small batch size of 5. On the contrary, the baseline typicality test-based method only achieves 64.6\% AUROC on average due to its failure on challenging problems. Our method also outperforms the state-of-the-art method by 9.1\% AUROC. For point-wise anomaly detection, our method achieves 90.7\% AUROC on average and outperforms the baseline by 5.2\% AUROC. Besides, our method has the least notable failures and is the most robust one. | ['Hongmei Wei', 'Kenli Li', 'Zhiming Liu', 'Ji Wang', 'Zhenbang Chen', 'Wanwei Liu', 'Jialu Pan', 'Yufeng Zhang'] | 2020-02-09 | null | null | null | null | ['group-anomaly-detection'] | ['methodology'] | [-3.60033691e-01 -2.13570625e-01 -8.38241801e-02 2.33691055e-02
-6.62953436e-01 -3.95588070e-01 6.15020216e-01 3.48158143e-02
-1.45393968e-01 5.50745845e-01 -5.62467128e-02 -5.07552922e-01
-1.37412354e-01 -7.88452446e-01 -6.05627477e-01 -8.08933377e-01
-3.42616707e-01 2.43411422e-01 2.59869903e-01 1.21014453e-01
3.06136757e-01 3.84099275e-01 -1.51609290e+00 -2.17127651e-01
1.09588909e+00 1.22448993e+00 -2.76640445e-01 7.45603144e-01
-2.27311581e-01 7.04326749e-01 -8.52261066e-01 -3.46087664e-01
4.46534932e-01 -6.45124972e-01 -3.39517534e-01 2.26829704e-02
8.26617122e-01 -5.54576397e-01 -6.13611519e-01 1.05075312e+00
4.70469803e-01 3.15065056e-01 1.06995988e+00 -1.54986835e+00
-7.31586099e-01 1.93146691e-02 -7.55563498e-01 7.85210550e-01
-2.52237823e-02 3.44981194e-01 1.17375410e+00 -9.53715026e-01
3.29722792e-01 1.08928812e+00 4.43206042e-01 6.58554971e-01
-1.23840892e+00 -6.92390203e-01 1.43881783e-01 4.55003045e-02
-1.38076866e+00 -2.07890585e-01 7.69672513e-01 -4.38012958e-01
7.85519898e-01 1.29030123e-01 4.65355784e-01 1.15153992e+00
3.74642849e-01 9.50794399e-01 8.10148537e-01 -1.07712522e-01
3.00946295e-01 3.95517051e-03 2.67089419e-02 9.05367732e-01
5.48580766e-01 1.03280105e-01 -5.46653569e-01 -3.10352355e-01
8.10824037e-01 -1.73257329e-02 -2.22114369e-01 -1.30229712e-01
-7.62787104e-01 9.11179483e-01 2.19205499e-01 1.89633980e-01
-1.72078565e-01 2.21811369e-01 3.64174545e-01 2.33247504e-01
5.53668082e-01 2.72657990e-01 -5.24543859e-02 -3.11863631e-01
-9.16292429e-01 4.06640708e-01 7.58145094e-01 6.55204833e-01
4.51603472e-01 6.15921021e-01 -1.70770675e-01 8.17652106e-01
4.15262341e-01 6.87681615e-01 4.35905576e-01 -8.67920160e-01
2.80137837e-01 3.73790830e-01 5.14743403e-02 -1.11157632e+00
-1.37247682e-01 -7.76765823e-01 -1.01839316e+00 2.77456462e-01
7.54688978e-01 -9.86583233e-02 -8.99247944e-01 1.60794604e+00
2.36814588e-01 5.13111651e-01 6.52446374e-02 8.43982220e-01
5.78796625e-01 7.91726291e-01 -1.69382706e-01 -1.43552005e-01
9.11827981e-01 -7.24729061e-01 -7.19291508e-01 -8.29795375e-02
7.71748006e-01 -6.85878873e-01 1.06573057e+00 5.47515810e-01
-7.69069731e-01 -4.67866927e-01 -1.02681851e+00 5.39555788e-01
-8.42408016e-02 -1.68960929e-01 5.63542366e-01 8.75482321e-01
-9.08029854e-01 5.94937146e-01 -8.84140313e-01 -2.67228544e-01
6.10660076e-01 -7.74076357e-02 -6.31416142e-02 7.59077892e-02
-9.21813428e-01 2.57115573e-01 2.83048265e-02 -1.62318666e-02
-1.10226369e+00 -8.60282123e-01 -8.04829061e-01 -7.30860829e-02
3.51881593e-01 -4.71132487e-01 8.58034670e-01 -5.64823985e-01
-1.13850856e+00 4.74414915e-01 -4.04136330e-01 -7.01273382e-01
6.95590675e-01 -5.31859756e-01 -6.10190213e-01 1.72468424e-01
1.11818083e-01 3.86880487e-01 8.98087800e-01 -1.21271646e+00
-6.11910820e-01 -1.23714298e-01 -3.41931701e-01 -1.71583921e-01
-4.81266648e-01 -3.21111560e-01 -3.97125959e-01 -8.74015868e-01
9.06441733e-02 -7.98662186e-01 -2.47230791e-02 1.16095833e-01
-5.73065042e-01 -5.12152255e-01 9.53674734e-01 -3.47157300e-01
1.53713751e+00 -2.25504088e+00 -4.56491023e-01 3.28174710e-01
5.04190743e-01 3.15564513e-01 5.09711131e-02 2.61949152e-01
1.01194948e-01 3.79960179e-01 -2.79139906e-01 -3.17288250e-01
1.41545236e-01 2.87607253e-01 -5.99137664e-01 6.79903328e-01
2.98304915e-01 4.15970653e-01 -9.22175109e-01 -4.07149404e-01
2.57179886e-01 3.43412936e-01 -8.61486435e-01 3.32054734e-01
2.72987522e-02 3.68854702e-01 -4.20021445e-01 8.75940621e-01
7.56360292e-01 -2.39102304e-01 -2.41161585e-01 2.25980356e-01
1.26087561e-01 6.39618114e-02 -1.18954766e+00 1.43576932e+00
-1.79850727e-01 9.40955400e-01 -4.29529101e-01 -1.03997397e+00
1.17505944e+00 1.85228109e-01 5.09538412e-01 -5.26202917e-01
-7.03688040e-02 4.24423814e-01 2.89395571e-01 -5.18713653e-01
3.13428521e-01 7.21748322e-02 1.63532600e-01 3.50605994e-01
3.47276181e-02 1.88373670e-01 1.58977002e-01 3.02775383e-01
1.22317851e+00 -2.52898200e-03 2.09260732e-02 -4.01336968e-01
4.85612303e-01 -3.77753735e-01 7.15124547e-01 1.28618383e+00
-6.79311514e-01 5.39195240e-01 8.08310747e-01 -6.27244115e-01
-8.00184250e-01 -1.51051950e+00 -3.16338897e-01 6.54690981e-01
1.02408469e-01 -4.83937263e-01 -5.93166769e-01 -1.01205611e+00
9.62854996e-02 8.54454517e-01 -6.41997516e-01 -2.35832781e-01
-4.28989232e-01 -1.06663823e+00 7.67427206e-01 6.07314110e-01
5.78818262e-01 -7.40940928e-01 -2.65890032e-01 8.75484422e-02
-9.74592119e-02 -1.03862071e+00 -1.85115218e-01 -2.91872174e-01
-9.84599233e-01 -1.02068222e+00 -6.68789029e-01 -2.44953021e-01
5.87131560e-01 1.05724791e-02 1.06289005e+00 1.21768400e-01
-3.68914425e-01 2.28462294e-01 -2.54203081e-01 -5.44815302e-01
-3.31151098e-01 -1.30075067e-01 2.53605038e-01 1.67846307e-01
5.86793423e-01 -4.82163996e-01 -9.68558788e-01 5.73692620e-01
-9.08506453e-01 -8.03212583e-01 4.97057974e-01 8.50094736e-01
6.38240218e-01 1.15892157e-01 6.82904720e-01 -5.90399563e-01
4.95076716e-01 -6.70991421e-01 -6.09251857e-01 -2.66138107e-01
-9.15831983e-01 1.96262792e-01 7.03884900e-01 -3.23359549e-01
-8.58990908e-01 -4.89403158e-01 -1.32850513e-01 -8.99077713e-01
-4.45617825e-01 7.59267062e-02 8.31010416e-02 3.36662322e-01
7.06177831e-01 3.51147562e-01 9.79208574e-02 -4.42378461e-01
-9.47437584e-02 3.87735933e-01 5.33921123e-01 -4.81420487e-01
9.29494441e-01 7.61013269e-01 1.55067712e-01 -1.13316238e+00
-8.83657873e-01 -3.51270407e-01 -1.82890445e-01 -2.88987488e-01
8.64981532e-01 -8.13832402e-01 -4.58058149e-01 5.95927894e-01
-7.68209338e-01 -2.75050461e-01 -1.93992883e-01 5.95039010e-01
-3.46807182e-01 4.60809469e-01 -5.09845257e-01 -1.12225819e+00
-1.81353346e-01 -1.02466083e+00 8.60110223e-01 2.04586789e-01
-2.32337385e-01 -1.17642641e+00 2.32810780e-01 5.26208952e-02
4.57590848e-01 4.75296557e-01 6.17635190e-01 -9.65599895e-01
-5.04211724e-01 -3.18007022e-01 -2.68760264e-01 4.87480104e-01
5.84652610e-02 2.14816302e-01 -1.16618371e+00 -1.67140439e-01
-2.23284319e-01 1.25726253e-01 1.11607909e+00 5.90148270e-01
1.61028135e+00 -6.86493292e-02 -2.02596515e-01 6.74653590e-01
1.35363758e+00 2.40438119e-01 6.97038531e-01 2.15615153e-01
7.27115810e-01 2.56299227e-01 5.51009119e-01 6.24583662e-01
1.06048519e-02 5.00521302e-01 7.09483087e-01 1.44589633e-01
-8.07366148e-02 -2.66934037e-01 5.74937761e-01 4.15686995e-01
-1.60565689e-01 -6.60269201e-01 -9.50069249e-01 6.16770625e-01
-1.69067860e+00 -1.11166322e+00 -4.65820760e-01 2.35687041e+00
2.25134343e-01 5.12316823e-01 2.92742789e-01 2.32003450e-01
5.13856828e-01 3.74366373e-01 -4.83348817e-01 -4.71986651e-01
2.44967211e-02 4.55613770e-02 2.95515448e-01 3.53229374e-01
-1.27299511e+00 7.09316909e-01 6.08525324e+00 9.85239685e-01
-8.98853600e-01 -9.62318555e-02 6.63507223e-01 -6.51587397e-02
-2.28209361e-01 -2.37124339e-01 -9.99169767e-01 7.37111986e-01
1.02071130e+00 3.54300104e-02 -9.20263007e-02 7.82799244e-01
1.69817522e-01 -1.84837475e-01 -9.64018822e-01 1.03969431e+00
1.94003597e-01 -1.08033073e+00 1.16164617e-01 4.79342401e-01
7.83727646e-01 -4.29259501e-02 2.56729782e-01 2.12854460e-01
1.40227824e-01 -9.43426907e-01 3.34617466e-01 5.39833605e-01
5.26293516e-01 -8.48864734e-01 9.22744215e-01 1.59696564e-01
-9.88395095e-01 7.74022937e-02 -4.05367166e-01 2.10629534e-02
1.03466779e-01 1.07974005e+00 -6.99876964e-01 4.43415850e-01
8.86163294e-01 7.38129556e-01 -4.37726796e-01 1.14567280e+00
-1.48454443e-01 1.24494481e+00 -5.13906837e-01 -9.13286861e-03
4.90833789e-01 -1.05047114e-01 1.00115728e+00 1.15795326e+00
5.32571435e-01 -4.09045309e-01 8.13198686e-02 1.01701903e+00
-2.39426866e-02 4.96658646e-02 -8.70679438e-01 5.22073656e-02
3.02178502e-01 8.96827102e-01 -5.56431770e-01 -3.12015623e-01
-4.63643402e-01 8.54924321e-01 -4.89642322e-02 4.48523760e-01
-1.01323974e+00 -5.53427160e-01 9.53280389e-01 1.87749252e-01
3.65158737e-01 -4.14581746e-02 -1.10429771e-01 -1.24667442e+00
1.44181266e-01 -5.04835725e-01 7.00221300e-01 -1.77323550e-01
-1.59484196e+00 5.57322383e-01 -4.55770642e-02 -1.55080521e+00
-3.76956850e-01 -7.69242942e-01 -9.37980294e-01 6.58903420e-01
-1.47687197e+00 -4.11224514e-01 -2.20201582e-01 5.55135429e-01
6.54284179e-01 -3.82799655e-01 6.62529111e-01 3.02170545e-01
-8.52243483e-01 7.84834445e-01 1.53916225e-01 4.86657709e-01
8.04406583e-01 -1.45755887e+00 3.48851562e-01 1.20298004e+00
2.57549733e-01 3.36750895e-01 8.07653069e-01 -5.91520488e-01
-1.10468245e+00 -1.11326861e+00 6.89535379e-01 -6.51830494e-01
7.04292238e-01 -1.14246733e-01 -1.20479155e+00 4.82877731e-01
2.49911062e-02 5.09548664e-01 8.65090013e-01 6.51853457e-02
-4.87286627e-01 -1.96509853e-01 -1.13537025e+00 5.20829320e-01
1.10412467e+00 -2.38786817e-01 -4.97988641e-01 2.38582909e-01
2.84679383e-01 -2.51088619e-01 -8.43804598e-01 3.89041603e-01
4.00633693e-01 -1.30996633e+00 9.14185643e-01 -5.80638111e-01
5.41718006e-01 -3.29174876e-01 -2.52237201e-01 -1.21090174e+00
-9.31820944e-02 -5.49942553e-01 -6.57764733e-01 1.17742503e+00
4.14725661e-01 -9.30817723e-01 7.96450377e-01 2.40723148e-01
-1.73630223e-01 -8.84644270e-01 -8.89975667e-01 -1.21648252e+00
2.40020096e-01 -8.98585439e-01 3.59144747e-01 7.72686422e-01
-3.89922291e-01 -1.50380149e-01 -4.11471099e-01 4.50512946e-01
9.43435967e-01 1.08579502e-01 9.86946046e-01 -1.39205849e+00
-3.36203396e-01 -6.12829447e-01 -6.12701714e-01 -1.08622825e+00
1.35506347e-01 -8.00637543e-01 -1.96288928e-01 -1.22650170e+00
-1.42500758e-01 -3.79586846e-01 -5.61327100e-01 1.72287062e-01
-3.47881943e-01 3.41350138e-01 -4.88130637e-02 2.80261368e-01
-5.34481406e-01 5.71605861e-01 1.05506313e+00 1.78539511e-02
-2.97609448e-01 -2.08931901e-02 -5.57006955e-01 8.89488161e-01
1.00773132e+00 -4.66615021e-01 -3.23703796e-01 -1.86852634e-01
5.02145104e-02 -3.23505491e-01 5.68364799e-01 -1.05087948e+00
-2.03924254e-01 2.51918342e-02 5.95481694e-01 -5.80335319e-01
8.02941248e-03 -4.88875896e-01 -5.17023325e-01 5.46739817e-01
-9.59397182e-02 9.58264917e-02 2.38890901e-01 1.01776016e+00
-3.14083010e-01 -5.93537800e-02 5.98120272e-01 1.99049607e-01
-8.17959607e-01 5.71664274e-01 -4.82869476e-01 5.51908314e-01
8.87555480e-01 -6.03406988e-02 -4.43864256e-01 -5.89150190e-01
-5.72118998e-01 2.37496644e-01 1.05311215e-01 4.50287759e-01
7.02223837e-01 -1.27479148e+00 -7.78376818e-01 5.14543235e-01
2.14535549e-01 -9.71658975e-02 2.53910154e-01 1.04058647e+00
-5.33307552e-01 1.46728441e-01 7.45640621e-02 -9.97276604e-01
-7.94871926e-01 2.58477867e-01 4.53917921e-01 -1.83041140e-01
-6.95986450e-01 7.67532229e-01 4.03264254e-01 -1.18736967e-01
1.90680891e-01 -8.45160782e-02 1.25726968e-01 -4.76063229e-02
5.62430084e-01 6.55161083e-01 -1.24151759e-01 -4.35614467e-01
-4.13137406e-01 5.09934366e-01 -1.60839185e-01 -1.50721483e-02
9.19040740e-01 1.26501828e-01 3.04399163e-01 4.45897967e-01
1.05972826e+00 1.34757563e-01 -1.42039907e+00 7.95717686e-02
-4.36945528e-01 -9.19435740e-01 7.64480084e-02 -4.22915816e-01
-1.26694202e+00 1.14341569e+00 6.07485414e-01 5.53885520e-01
9.34722006e-01 2.50676852e-02 7.72291780e-01 1.96272701e-01
-4.53369543e-02 -9.48366761e-01 2.39913747e-01 3.18298399e-01
5.99405706e-01 -1.32384646e+00 -2.28377402e-01 -2.48119056e-01
-6.14578009e-01 9.79989588e-01 9.15566087e-01 -4.97930706e-01
9.37698603e-01 -5.68408519e-04 3.15220803e-02 -1.62400648e-01
-7.62727082e-01 -5.01987860e-02 5.18131256e-01 5.24620950e-01
2.76585191e-01 -7.08591044e-02 -7.81237334e-02 2.29230404e-01
-1.39817879e-01 -4.32780772e-01 3.95870805e-01 6.32018328e-01
-3.87906075e-01 -7.50557005e-01 -3.06084812e-01 6.89167559e-01
-7.88107693e-01 1.46882057e-01 -9.70060974e-02 9.34421539e-01
-1.01640649e-01 8.38123322e-01 5.26714325e-01 -2.82598555e-01
1.83014348e-01 2.62170821e-01 -2.87664421e-02 -3.34128961e-02
-1.21760972e-01 2.82349139e-01 -2.40143582e-01 -8.06558192e-01
-1.83806017e-01 -7.49183714e-01 -1.15284121e+00 -3.96284997e-01
-2.36172095e-01 2.12123226e-02 2.41203740e-01 8.33478212e-01
4.88893390e-01 4.48333144e-01 6.37344062e-01 -3.42737228e-01
-4.56502706e-01 -7.75440991e-01 -7.25571573e-01 5.08854806e-01
4.96302933e-01 -7.55370855e-01 -1.04160392e+00 -2.77852982e-01] | [7.656364917755127, 2.2893428802490234] |
132eaafc-3c33-4ea5-9337-bfd74e51fc69 | generating-multiple-choice-questions-for | 2303.07069 | null | https://arxiv.org/abs/2303.07069v1 | https://arxiv.org/pdf/2303.07069v1.pdf | Generating multiple-choice questions for medical question answering with distractors and cue-masking | Medical multiple-choice question answering (MCQA) is particularly difficult. Questions may describe patient symptoms and ask for the correct diagnosis, which requires domain knowledge and complex reasoning. Standard language modeling pretraining alone is not sufficient to achieve the best results. \citet{jin2020disease} showed that focusing masked language modeling on disease name prediction when using medical encyclopedic paragraphs as input leads to considerable MCQA accuracy improvement. In this work, we show that (1) fine-tuning on generated MCQA dataset outperforms the masked language modeling based objective and (2) correctly masking the cues to the answers is critical for good performance. We release new pretraining datasets and achieve state-of-the-art results on 4 MCQA datasets, notably +5.7\% with base-size model on MedQA-USMLE. | ['Marie-Francine Moens', 'Kanimozhi Uma', 'Damien Sileo'] | 2023-03-13 | null | null | null | null | ['multiple-choice-qa'] | ['natural-language-processing'] | [ 1.16069885e-02 4.50210214e-01 -3.20618868e-01 -4.20133680e-01
-1.90533769e+00 -4.03068423e-01 1.22358315e-01 4.50769067e-01
-4.67672974e-01 1.02633595e+00 6.69242799e-01 -5.71369767e-01
-3.74197245e-01 -6.90168440e-01 -6.19148731e-01 -1.27165541e-01
1.76351443e-01 1.06359565e+00 2.56905645e-01 -4.43403304e-01
-2.31448844e-01 -5.47138266e-02 -9.11181867e-01 1.20777214e+00
1.29755068e+00 8.10430467e-01 2.09714808e-02 9.49139833e-01
-5.64916134e-01 1.39089680e+00 -5.41417658e-01 -6.31816804e-01
-1.68701068e-01 -6.91147387e-01 -1.12190402e+00 -3.61590356e-01
5.25737941e-01 -5.14469370e-02 -1.20126799e-01 5.28846979e-01
6.65708661e-01 -2.99037695e-01 6.34134352e-01 -7.92109489e-01
-7.72049725e-01 6.75221622e-01 -3.35440263e-02 3.08201522e-01
5.55652976e-01 2.81842917e-01 1.28264380e+00 -6.77288532e-01
7.49987543e-01 1.21852934e+00 6.43945873e-01 1.04102588e+00
-1.09865451e+00 -4.95524138e-01 -1.06095418e-01 2.96641678e-01
-1.18331289e+00 -3.31552535e-01 9.85927358e-02 -3.40597481e-01
1.24480355e+00 6.93732798e-01 -1.56797290e-01 9.63242531e-01
2.74870545e-01 7.56714761e-01 1.12209606e+00 -3.43068480e-01
4.32596281e-02 1.40458941e-01 3.13992620e-01 8.43318760e-01
-2.96042636e-02 -1.22511648e-01 -3.14920247e-01 -5.84604502e-01
1.04999341e-01 -2.64756650e-01 -1.98093265e-01 4.10165161e-01
-1.24493492e+00 9.29288507e-01 6.27179384e-01 2.33437032e-01
-4.80673283e-01 -4.70907837e-02 1.51070386e-01 5.76938093e-01
3.01614583e-01 1.00937319e+00 -9.54016030e-01 9.14137289e-02
-1.01030171e+00 5.28054655e-01 8.50844920e-01 5.75652003e-01
2.61153966e-01 -5.26582241e-01 -7.43216097e-01 8.93115878e-01
2.02145487e-01 7.25424290e-01 3.84374142e-01 -1.01476920e+00
6.57574594e-01 7.79587209e-01 1.43599167e-01 -5.33128381e-01
-8.72924447e-01 -7.68470466e-01 -6.31282151e-01 -4.09469873e-01
7.49090314e-01 -4.37043369e-01 -1.16134560e+00 1.76369762e+00
3.25656720e-02 -2.53194153e-01 3.32552165e-01 5.72948992e-01
1.58902907e+00 4.29325342e-01 6.50287092e-01 4.48281579e-02
1.79842997e+00 -9.29297626e-01 -1.03351831e+00 -3.64160746e-01
1.16093969e+00 -7.83019483e-01 1.13247120e+00 3.22560579e-01
-1.17629302e+00 -1.89500883e-01 -6.20440900e-01 -3.45646113e-01
-3.18733603e-01 7.67908469e-02 4.17603642e-01 5.73845983e-01
-1.00595427e+00 1.37707204e-01 -6.29645944e-01 -2.70494848e-01
4.00205225e-01 2.48644277e-01 -1.88506722e-01 -5.94159424e-01
-1.65760899e+00 1.10976875e+00 1.54788747e-01 -3.31450582e-01
-7.58917153e-01 -1.24185669e+00 -6.51711941e-01 -3.01903803e-02
5.37977874e-01 -1.20464802e+00 1.28471804e+00 -4.02055740e-01
-9.92331266e-01 1.12345898e+00 -3.26302558e-01 -5.74178159e-01
5.45831025e-01 -1.74811572e-01 -7.98709393e-01 3.39750350e-01
2.45445490e-01 1.04383731e+00 2.75394559e-01 -8.09209585e-01
-5.87135017e-01 -2.42032886e-01 1.23017028e-01 5.38911158e-03
1.06001897e-02 6.63388968e-02 -3.73783708e-01 -6.21599078e-01
-1.37506083e-01 -5.98131895e-01 -4.25147623e-01 -3.39028612e-02
-4.78087842e-01 -3.80613118e-01 -1.07846595e-01 -1.14957643e+00
1.41406405e+00 -1.77960360e+00 -2.64220804e-01 -1.67014748e-01
3.52418572e-01 1.01443771e-02 -2.54549026e-01 3.74000520e-01
-1.42854497e-01 3.01083535e-01 -2.60905236e-01 -1.43314585e-01
-6.06041737e-02 2.71566451e-01 -4.24076229e-01 -1.28968865e-01
6.88012600e-01 1.23209667e+00 -7.23979950e-01 -6.92878485e-01
-4.04488355e-01 2.59525031e-01 -9.45668817e-01 1.97215617e-01
-7.94916689e-01 3.43644202e-01 -4.97643918e-01 8.41505706e-01
3.56115937e-01 -8.55273187e-01 -1.14406884e-01 -6.49731001e-03
4.83462811e-01 7.29312062e-01 -6.43697858e-01 1.49904811e+00
-2.68773377e-01 5.68102896e-02 1.12501167e-01 -4.13088232e-01
4.06874806e-01 6.27174616e-01 4.14030761e-01 -9.82821941e-01
-2.74807602e-01 3.85826111e-01 3.01935375e-01 -8.93378437e-01
-8.19730312e-02 -3.89118880e-01 -6.09610714e-02 2.12491080e-01
6.46213666e-02 -3.65091451e-02 1.07733376e-01 3.06733549e-01
1.33275306e+00 -3.50613058e-01 3.07549626e-01 -3.45411181e-01
6.36116683e-01 6.36516213e-01 5.61476588e-01 8.68312061e-01
7.06392974e-02 5.46727717e-01 5.40003598e-01 -1.57483995e-01
-4.72357154e-01 -1.07469213e+00 -2.85218000e-01 1.19212794e+00
-6.65115118e-01 -4.06930208e-01 -6.08500361e-01 -8.72945607e-01
-5.21192793e-03 1.08596456e+00 -7.66624212e-01 7.96931162e-02
-4.94172424e-01 -1.11307323e+00 8.43337059e-01 4.90266711e-01
2.07700789e-01 -8.77941966e-01 -2.59822756e-01 4.22733605e-01
-6.10862613e-01 -1.10492909e+00 -3.05202454e-01 3.67551483e-02
-9.60594594e-01 -1.25570428e+00 -8.63463998e-01 -5.06909072e-01
4.38816935e-01 -4.44100440e-01 1.73832977e+00 2.05365613e-01
-4.32134241e-01 4.02496904e-01 -2.03383207e-01 -5.32819867e-01
-6.67421520e-01 2.00736895e-01 -4.95131344e-01 -4.88607496e-01
6.69998586e-01 2.62845401e-03 -6.85669303e-01 5.62438481e-02
-1.01513016e+00 4.58167195e-02 8.12472105e-01 8.55695248e-01
5.74284434e-01 -5.06801069e-01 9.08669949e-01 -1.29325342e+00
8.03492665e-01 -6.89410627e-01 -1.06725745e-01 6.02839291e-01
-6.27288342e-01 3.08619767e-01 3.45453590e-01 -7.46921524e-02
-7.96195865e-01 -2.45268762e-01 -7.78162956e-01 2.20189050e-01
-2.99679041e-01 7.84240663e-01 1.28350899e-01 3.79708886e-01
1.17576098e+00 -5.14015742e-02 3.64164524e-02 -7.49118686e-01
4.86848384e-01 4.46999967e-01 4.17534262e-01 -4.43168432e-01
3.35506022e-01 3.67846310e-01 -1.98827475e-01 -2.47896641e-01
-1.25760555e+00 -4.37903881e-01 -2.33970687e-01 3.19087863e-01
1.29311740e+00 -1.06709886e+00 -6.78201377e-01 -2.81171679e-01
-9.92686987e-01 -2.82608032e-01 -6.70399293e-02 2.53426164e-01
-2.64109582e-01 -7.83951432e-02 -7.96237826e-01 -5.02144754e-01
-4.84876603e-01 -1.05103576e+00 9.09004867e-01 -1.40925243e-01
-6.21097088e-01 -1.09550393e+00 2.63387889e-01 1.01639938e+00
4.99181479e-01 2.25755751e-01 1.60407043e+00 -1.09512091e+00
-4.36159253e-01 -5.95130660e-02 -1.74594611e-01 1.03130983e-03
2.68449903e-01 -5.05422533e-01 -8.75376523e-01 4.66676168e-02
-1.53819144e-01 -5.55441797e-01 1.04692900e+00 4.09586996e-01
1.07390130e+00 -2.51628816e-01 -2.84725666e-01 9.39856172e-02
1.29030132e+00 5.91110401e-02 4.24882352e-01 -9.52146351e-02
5.09826362e-01 8.19054604e-01 4.59970474e-01 3.11910026e-02
8.01967025e-01 4.29147720e-01 1.56431004e-01 -2.25708231e-01
-3.73863339e-01 -7.98751190e-02 1.24292471e-01 6.21748984e-01
5.92566252e-01 -2.27509782e-01 -1.40845656e+00 7.31976569e-01
-1.49402881e+00 -5.37147284e-01 -2.60345161e-01 1.72965908e+00
1.39897251e+00 2.19469890e-01 1.66929066e-02 -3.02128792e-01
3.99225131e-02 -3.01757812e-01 -5.19345939e-01 -2.81740040e-01
-3.93592298e-01 6.09710634e-01 3.03601801e-01 8.85164797e-01
-8.58861685e-01 6.97747409e-01 7.06647396e+00 9.96120751e-01
-8.58376384e-01 4.76482064e-01 7.96003282e-01 -2.28519425e-01
-6.77551389e-01 -4.66916382e-01 -9.10558999e-01 2.24241972e-01
1.27904105e+00 5.99308275e-02 2.06168909e-02 3.72642905e-01
8.41423795e-02 -5.15710674e-02 -1.16104555e+00 7.92446077e-01
1.22847475e-01 -1.63580287e+00 3.26508582e-01 -2.10976794e-01
6.91374600e-01 1.39258608e-01 1.57602921e-01 6.34004891e-01
4.87434596e-01 -1.58510292e+00 -2.93292664e-02 9.10091043e-01
7.22053945e-01 -4.07490999e-01 9.17682707e-01 4.08240080e-01
-3.81718069e-01 3.08007058e-02 -2.07846947e-02 2.42681757e-01
3.26360837e-02 6.36205256e-01 -1.30620265e+00 6.53098464e-01
5.83146751e-01 2.13285506e-01 -1.01270974e+00 9.92117465e-01
-1.15299709e-01 9.88434434e-01 -1.36044011e-01 -1.11860178e-01
4.23319131e-01 5.56014955e-01 2.38159925e-01 1.29278016e+00
-3.75024080e-02 4.42496270e-01 2.06583086e-02 7.72085905e-01
-2.84746528e-01 4.84410405e-01 -6.38935864e-02 -3.56061339e-01
-4.03245864e-03 6.89073086e-01 -2.11431742e-01 -4.60772246e-01
-4.85645890e-01 5.84588885e-01 2.84132421e-01 1.94791719e-01
-5.06500423e-01 -3.20383580e-03 5.87734103e-01 2.12877795e-01
-1.26321912e-02 3.77795190e-01 -5.14163196e-01 -1.02457297e+00
-2.58819193e-01 -1.45775747e+00 1.19833791e+00 -6.76503539e-01
-1.62319577e+00 6.45733774e-01 -3.25391859e-01 -9.22188878e-01
-6.38937771e-01 -7.90999115e-01 1.01861633e-01 1.07978868e+00
-1.59801400e+00 -1.02214229e+00 6.89420551e-02 6.26312733e-01
3.49545687e-01 -1.69753462e-01 1.37458408e+00 6.26736701e-01
-9.23068300e-02 7.63033152e-01 -1.02899164e-01 2.23702013e-01
1.25982070e+00 -1.43971491e+00 2.18518469e-02 3.85123104e-01
1.78108383e-02 6.08592987e-01 6.83632851e-01 -6.80002928e-01
-1.06209970e+00 -1.07490385e+00 1.52840936e+00 -1.22659862e+00
4.22636002e-01 -4.76637147e-02 -1.14723456e+00 4.63406086e-01
3.09264123e-01 -4.75591391e-01 1.27527189e+00 1.25576571e-01
-4.38816935e-01 4.17211615e-02 -1.39454627e+00 5.84494114e-01
6.25970900e-01 -7.27276266e-01 -9.88224447e-01 7.42851317e-01
1.15325475e+00 -4.67751116e-01 -1.09199417e+00 8.35111082e-01
9.63531435e-02 -3.60482901e-01 1.02090669e+00 -1.59201455e+00
6.99494660e-01 -1.45859599e-01 -3.53813052e-01 -1.02105129e+00
-3.00668657e-01 -2.20868900e-01 8.54901504e-04 6.09812021e-01
1.10327959e+00 -4.34374332e-01 6.88400149e-01 7.66119361e-01
5.47076613e-02 -1.07592845e+00 -8.45132291e-01 -1.31777525e-01
6.07336879e-01 -4.80933309e-01 3.74175996e-01 1.08679974e+00
-1.98891982e-01 4.28645551e-01 -1.81991272e-02 3.83797348e-01
3.17132026e-01 -6.08461089e-02 1.50239676e-01 -9.03680503e-01
-4.66878206e-01 -3.91903639e-01 1.85648441e-01 -9.43151057e-01
-1.19048767e-01 -1.02295589e+00 -2.00968489e-01 -2.02553082e+00
2.55294561e-01 -3.04832757e-01 -4.78622347e-01 6.95436001e-01
-8.14544320e-01 1.79744050e-01 -1.19432539e-01 -3.06416363e-01
-7.17757106e-01 1.35769516e-01 1.24407935e+00 -2.88267195e-01
1.11355424e-01 7.28324503e-02 -1.12149906e+00 3.13044101e-01
5.43674648e-01 -5.96706152e-01 -4.62216288e-01 -7.48752654e-01
5.27798057e-01 5.88664353e-01 1.87003478e-01 -7.09717810e-01
3.28160733e-01 -7.19074579e-03 2.80089080e-01 -7.66710758e-01
3.52835625e-01 -4.90049779e-01 -2.94154644e-01 9.62903440e-01
-9.13389504e-01 2.80047745e-01 4.99856591e-01 4.05262232e-01
-2.87830919e-01 -8.22094679e-02 5.31341136e-01 -4.50827807e-01
-4.17741984e-01 1.05054609e-01 -4.38229322e-01 8.65064323e-01
2.80761838e-01 5.48525572e-01 -4.67102528e-01 -5.88583887e-01
-1.21683407e+00 8.99417400e-01 -2.60498762e-01 4.85527605e-01
4.48384285e-01 -1.04210913e+00 -1.33569884e+00 -3.88826579e-01
3.51585120e-01 -2.38638282e-01 5.54788172e-01 8.43639910e-01
-7.11079061e-01 9.62663949e-01 2.09382996e-01 -5.21138608e-01
-1.34854102e+00 5.32606184e-01 4.71459180e-01 -7.78127789e-01
-2.54067294e-02 1.26996744e+00 5.56458533e-02 -8.26354742e-01
2.59210318e-01 -5.89726925e-01 -3.18013430e-01 3.02022435e-02
8.22145700e-01 -2.54680798e-03 4.30855274e-01 -6.63037151e-02
-5.51357388e-01 1.31018832e-01 -3.27737123e-01 -1.94653556e-01
1.10087347e+00 2.29338512e-01 -2.79905915e-01 2.54868001e-01
9.92268622e-01 2.74810165e-01 -3.44205797e-01 -5.61911464e-01
4.78215039e-01 1.11036047e-01 -4.22923453e-02 -1.74810970e+00
-5.65868020e-01 9.94334042e-01 6.75728440e-01 -1.94766715e-01
1.10494065e+00 2.40693331e-01 7.80100942e-01 6.73269987e-01
-1.70289855e-02 -6.84271872e-01 4.61627692e-02 3.78636092e-01
9.17710125e-01 -1.41448390e+00 -1.58634931e-01 -3.80171895e-01
-8.54599714e-01 7.46227980e-01 5.55865765e-01 1.95578739e-01
7.65016258e-01 1.00892469e-01 7.13990033e-01 -3.95046443e-01
-1.28021610e+00 -2.94958025e-01 8.36485326e-01 1.93489432e-01
7.45171368e-01 2.00539321e-01 -3.49883080e-01 9.39114690e-01
-3.03642303e-01 -3.46168801e-02 3.36749665e-02 5.58345139e-01
-1.74535304e-01 -1.26901400e+00 -4.09589916e-01 7.67536342e-01
-1.04907310e+00 -5.74928701e-01 -4.70924944e-01 6.39048219e-01
2.54218876e-01 1.20611238e+00 -1.19359516e-01 -1.16546310e-01
3.68662834e-01 6.22493327e-01 3.29807818e-01 -9.18175697e-01
-1.01388323e+00 -1.45051554e-01 5.59167206e-01 -5.94623208e-01
-2.73795456e-01 -3.21561396e-01 -1.35360658e+00 6.74496368e-02
2.99245238e-01 3.46672535e-01 1.26404509e-01 1.05846334e+00
7.04348266e-01 8.97061825e-01 -1.26458928e-01 8.21562827e-01
-5.75389504e-01 -1.01598454e+00 1.29390121e-01 3.82081270e-01
5.92041433e-01 -1.84889868e-01 -3.72608043e-02 -1.50790170e-01] | [8.767420768737793, 8.564485549926758] |
0b5c7936-4b13-4acd-af8c-2adf2d353467 | an-emg-gesture-recognition-system-with | 1802.10237 | null | http://arxiv.org/abs/1802.10237v2 | http://arxiv.org/pdf/1802.10237v2.pdf | An EMG Gesture Recognition System with Flexible High-Density Sensors and Brain-Inspired High-Dimensional Classifier | EMG-based gesture recognition shows promise for human-machine interaction.
Systems are often afflicted by signal and electrode variability which degrades
performance over time. We present an end-to-end system combating this
variability using a large-area, high-density sensor array and a robust
classification algorithm. EMG electrodes are fabricated on a flexible substrate
and interfaced to a custom wireless device for 64-channel signal acquisition
and streaming. We use brain-inspired high-dimensional (HD) computing for
processing EMG features in one-shot learning. The HD algorithm is tolerant to
noise and electrode misplacement and can quickly learn from few gestures
without gradient descent or back-propagation. We achieve an average
classification accuracy of 96.64% for five gestures, with only 7% degradation
when training and testing across different days. Our system maintains this
accuracy when trained with only three trials of gestures; it also demonstrates
comparable accuracy with the state-of-the-art when trained with one trial. | ['Luca Benini', 'Fred Burghardt', 'Natasha Yamamoto', 'Simone Benatti', 'Jonathan Ting', 'Jan M. Rabaey', 'Alisha Menon', 'Ali Moin', 'Abbas Rahimi', 'Yasser Khan', 'Senam Tamakloe', 'Andy Zhou', 'Ana C. Arias'] | 2018-02-28 | null | null | null | null | ['emg-gesture-recognition'] | ['medical'] | [ 7.50415027e-01 -2.93900400e-01 -1.59479193e-02 -2.39085674e-01
-1.28570390e+00 -2.97490478e-01 -1.81346878e-01 -3.77254516e-01
-9.33723748e-01 5.04562855e-01 -5.04449755e-02 4.86809671e-01
-5.37950173e-02 1.04017928e-01 -6.90323174e-01 -6.30326033e-01
-5.27874768e-01 1.79167420e-01 2.50142395e-01 2.99058318e-01
4.67864089e-02 2.86551267e-01 -1.51702845e+00 4.12710905e-01
2.90514499e-01 1.29691029e+00 2.71495461e-01 8.43953252e-01
4.79665369e-01 2.11382285e-01 -9.04917777e-01 1.01170756e-01
3.13582569e-01 -3.55356753e-01 -1.61379933e-01 -2.18034282e-01
2.76805282e-01 -5.53234994e-01 -3.35779727e-01 5.31120956e-01
1.29200280e+00 -3.06135297e-01 5.27749956e-01 -1.04401481e+00
5.95901348e-02 4.42825228e-01 -4.38072592e-01 2.38575652e-01
5.90826571e-01 3.16413403e-01 3.76558483e-01 -1.06280565e+00
6.25980020e-01 6.15574539e-01 1.16458654e+00 9.89637494e-01
-1.37218869e+00 -9.04869497e-01 -2.65933692e-01 7.17213899e-02
-1.48245025e+00 -6.25043273e-01 5.02997637e-01 -3.42161238e-01
1.17385983e+00 3.04247886e-01 9.98100460e-01 1.57774580e+00
6.83765411e-01 8.15252185e-01 8.57482195e-01 4.70912233e-02
5.48486471e-01 -5.39689004e-01 -4.02707607e-02 7.84674883e-02
2.39001989e-01 -1.21852815e-01 -1.35396147e+00 -1.11157507e-01
9.64388132e-01 2.06207037e-01 -3.22524846e-01 7.68685862e-02
-1.25134134e+00 5.23631983e-02 2.03994632e-01 2.62974292e-01
-8.62535477e-01 5.08488536e-01 2.78457165e-01 4.44007844e-01
1.35568112e-01 4.98439431e-01 -3.12874407e-01 -1.23245609e+00
-1.00983238e+00 5.93445860e-02 9.58171248e-01 9.58929002e-01
-2.34675661e-01 6.18775561e-02 -3.84107530e-02 8.77651930e-01
6.25618473e-02 6.37317717e-01 7.79928744e-01 -7.98978627e-01
4.32785034e-01 1.09506600e-01 -3.00874710e-02 -6.01268709e-01
-8.83280337e-01 -1.88599452e-01 -6.68371618e-01 2.61939257e-01
3.47909659e-01 -7.32209086e-01 -9.02571321e-01 1.36764634e+00
-8.48002061e-02 3.13071996e-01 -2.53931701e-01 1.26019609e+00
3.08863729e-01 1.66789308e-01 -6.73883036e-02 -3.57065499e-01
7.97810376e-01 -2.79436290e-01 -8.50558102e-01 -4.65534747e-01
1.42698765e-01 -3.27752829e-01 1.18204260e+00 9.58313823e-01
-1.11282372e+00 -1.20117068e-01 -1.31494796e+00 4.02795196e-01
3.42028260e-01 -1.23961970e-01 6.09552026e-01 7.49611318e-01
-7.11754978e-01 8.45463455e-01 -1.64041960e+00 -3.85126501e-01
6.96756721e-01 1.06006134e+00 -3.19652021e-01 2.96176195e-01
-5.43280780e-01 5.84428072e-01 -3.42903644e-01 2.32866749e-01
-4.61929202e-01 -6.34450674e-01 -2.74619132e-01 -4.65171158e-01
-1.26957700e-01 -2.48727173e-01 1.14937913e+00 -6.59877717e-01
-2.14104271e+00 4.47703868e-01 -5.23177795e-02 -3.60131979e-01
4.27422315e-01 -7.14125991e-01 -5.89155376e-01 2.18489051e-01
-1.79419309e-01 3.05292010e-01 1.06392133e+00 -5.08769691e-01
-2.06661239e-01 -9.22141612e-01 -9.97894585e-01 1.72495186e-01
-6.96581244e-01 8.88029262e-02 -2.06536755e-01 -6.35380566e-01
4.68160808e-01 -1.04778838e+00 1.36072904e-01 4.20659572e-01
-1.17681280e-01 3.63845170e-01 6.34743452e-01 -5.22964239e-01
1.20210278e+00 -2.14333463e+00 2.16278657e-01 3.43341202e-01
-4.82277572e-02 1.35024339e-01 -1.71788514e-01 3.16432774e-01
4.31756765e-01 -3.68487120e-01 -2.38060847e-01 2.70517892e-03
-2.40488455e-01 6.13107644e-02 2.17118487e-01 7.59897590e-01
1.71367273e-01 9.02848899e-01 -5.10262668e-01 2.65747029e-03
6.21976182e-02 6.04405880e-01 -3.52822006e-01 3.72929841e-01
5.19232213e-01 6.67096138e-01 -2.68754423e-01 1.08983147e+00
7.90211633e-02 1.88535098e-02 2.65927136e-01 -3.17169487e-01
1.75547913e-01 1.26170412e-01 -1.38004112e+00 2.41681457e+00
-2.74898380e-01 7.83635378e-01 4.48345363e-01 -7.15704620e-01
8.71768415e-01 3.83247793e-01 8.64312589e-01 -6.83403492e-01
4.74707216e-01 6.30416572e-01 2.84102172e-01 -1.04979634e+00
-3.09793890e-01 -3.30055552e-03 -2.13104159e-01 5.41992784e-01
1.47941664e-01 -2.75588110e-02 -4.82002079e-01 -3.72624695e-01
1.75476646e+00 1.34923905e-01 -2.65620053e-01 -7.79675916e-02
-5.21254539e-01 -1.50903106e-01 3.98994565e-01 6.64599419e-01
-9.75768790e-02 7.46466875e-01 -1.61656439e-01 -1.90949321e-01
-5.95184863e-01 -1.31684637e+00 -2.63273746e-01 1.01038933e+00
2.25441828e-01 -2.16668218e-01 -7.43300438e-01 1.42534211e-01
3.60983461e-01 2.59007141e-02 -1.59804493e-01 -2.36539111e-01
-5.63393116e-01 -7.25827992e-01 8.38058114e-01 9.69406188e-01
2.32095122e-01 -9.91207659e-01 -1.41622877e+00 8.07797968e-01
2.84980923e-01 -1.02581918e+00 -2.78166175e-01 6.24619901e-01
-1.15443802e+00 -7.21202075e-01 -8.97443295e-01 -7.76096582e-01
2.44432941e-01 -4.24629450e-01 4.15509611e-01 -5.49276173e-01
-8.17314208e-01 6.25641942e-01 -2.13694304e-01 -7.20114172e-01
4.32311088e-01 7.66050965e-02 6.43063545e-01 -1.44672558e-01
6.81335092e-01 -1.11361587e+00 -7.31473923e-01 2.91387767e-01
-2.66765416e-01 -4.30061311e-01 9.45868015e-01 8.77604425e-01
7.20414221e-01 -7.50702083e-01 8.00961554e-01 -1.78393498e-01
9.59712327e-01 -2.88410395e-01 1.24016389e-01 -1.60842106e-01
-5.81441462e-01 -1.45604774e-01 1.47747234e-01 -1.14556360e+00
-4.53628570e-01 5.63900352e-01 -1.09542333e-01 -4.10985619e-01
-5.28487787e-02 2.88556278e-01 1.74177215e-01 -3.47207040e-01
9.78047013e-01 3.48485336e-02 5.52617550e-01 -5.40087223e-01
-1.59131885e-02 1.43700457e+00 9.82219279e-01 -2.34146550e-01
2.68915921e-01 2.86525130e-01 -3.40664864e-01 -1.12110317e+00
1.31289080e-01 -3.90002728e-01 -6.94002569e-01 -5.63859403e-01
5.23326635e-01 -1.03375459e+00 -8.58416378e-01 1.09336138e+00
-8.62850726e-01 -7.47508109e-01 -1.20051466e-01 1.04315841e+00
-6.66268587e-01 -4.19383526e-01 -8.07774901e-01 -9.43755925e-01
-9.58368301e-01 -7.49955952e-01 1.31602263e+00 3.65008526e-02
-8.78682315e-01 9.26618639e-04 -6.67292476e-02 -8.56076255e-02
4.79432106e-01 4.58717257e-01 -3.02997814e-03 -3.22825462e-01
1.48709584e-02 -7.63749957e-01 4.34266329e-01 -4.55114059e-02
2.42144898e-01 -5.97241223e-01 -1.06026328e+00 -5.07283330e-01
1.33076563e-01 -6.69168830e-01 3.25339884e-01 5.79936445e-01
1.10855317e+00 1.92203932e-02 -6.23853445e-01 6.14118159e-01
1.12591565e+00 3.28596324e-01 7.38562703e-01 7.63082132e-02
5.90671539e-01 8.06226358e-02 3.60583901e-01 4.58509028e-01
-2.47277007e-01 8.23308051e-01 -4.86015603e-02 1.10544853e-01
-6.39455393e-02 6.24064095e-02 5.19297540e-01 8.40073347e-01
-2.43482247e-01 4.71998751e-02 -7.87814081e-01 3.16109180e-01
-1.70801151e+00 -6.88300490e-01 7.90738016e-02 2.21676898e+00
9.75741804e-01 2.13517010e-01 4.72401589e-01 3.33801687e-01
4.83054876e-01 -4.12029326e-01 -1.30857050e+00 -2.90586859e-01
1.68416321e-01 7.78988242e-01 6.43306315e-01 -1.88248247e-01
-7.46292949e-01 4.64037627e-01 6.89093971e+00 3.89116704e-01
-1.64213419e+00 1.93484157e-01 -9.89940092e-02 -1.10096216e+00
5.99314988e-01 -1.02226293e+00 -4.52096820e-01 6.20358527e-01
1.16115439e+00 1.31022871e-01 5.22128463e-01 7.55526960e-01
1.13841645e-01 -9.48544964e-02 -1.22955000e+00 1.58531308e+00
1.81991249e-01 -1.04676485e+00 -8.12301397e-01 -3.80209535e-02
3.17728192e-01 5.04037976e-01 -2.38347977e-01 -1.49749756e-01
-5.92603683e-01 -1.10916471e+00 6.78545475e-01 5.16344070e-01
1.49503016e+00 -3.93723696e-01 4.59068418e-01 3.70478511e-01
-1.01263726e+00 -2.62747198e-01 8.28815699e-02 -5.21140933e-01
2.31601655e-01 1.77583203e-01 -4.01742339e-01 -2.77190328e-01
1.01499951e+00 7.20649123e-01 6.08410239e-02 9.81385410e-01
8.69005173e-02 7.37398565e-01 -8.42351019e-01 -6.74584270e-01
-3.42884034e-01 3.48751009e-01 6.01699829e-01 1.17165232e+00
3.84864151e-01 4.20917094e-01 -1.37266234e-01 4.29024398e-01
4.04596142e-02 -3.26606005e-01 -5.25909483e-01 -3.84634035e-03
7.35450566e-01 8.23003292e-01 -3.86882573e-01 1.52569622e-01
-2.26352677e-01 1.53750408e+00 -2.78837909e-03 2.55980521e-01
-4.12543982e-01 -9.97908413e-01 7.92704165e-01 1.10072210e-01
1.20733812e-01 -5.43351829e-01 -7.91464508e-01 -9.34271514e-01
7.92327285e-01 -6.58372819e-01 -6.91712946e-02 -4.36081856e-01
-1.40606153e+00 4.26402420e-01 -4.44762230e-01 -1.53392398e+00
-5.45541644e-01 -7.50624776e-01 -4.35684115e-01 5.24400651e-01
-5.65743446e-01 -4.95847583e-01 -3.37772340e-01 6.87952638e-01
5.74962795e-01 -8.61638263e-02 1.33378732e+00 3.57614815e-01
-4.79174942e-01 9.89239752e-01 2.09530577e-01 1.36276916e-01
7.44004130e-01 -8.13189328e-01 3.61856610e-01 5.74095726e-01
2.98459549e-02 5.62842667e-01 3.68370682e-01 -7.23891497e-01
-2.42659998e+00 -6.73399508e-01 2.77997315e-01 -2.31777325e-01
5.06823003e-01 -8.52848411e-01 -7.48598039e-01 3.78864110e-01
-2.87816674e-01 4.15439792e-02 8.78692091e-01 5.17268442e-02
1.11962454e-02 -4.29072350e-01 -1.23289192e+00 5.39692342e-01
1.68044102e+00 -5.03166497e-01 -6.50640786e-01 2.03262702e-01
-5.55458553e-02 -7.79065430e-01 -1.22063410e+00 2.80526996e-01
1.76060617e+00 -1.35742456e-01 5.52265823e-01 -1.93860710e-01
-6.86347187e-02 1.24701656e-01 -5.98075800e-02 -1.24097061e+00
-2.01937452e-01 -1.08938527e+00 -2.57446796e-01 6.77494645e-01
6.53805256e-01 -5.42211950e-01 1.05449915e+00 1.00963485e+00
-8.80869925e-02 -1.06390750e+00 -1.30254745e+00 -1.19129336e+00
-3.41505677e-01 -8.10701013e-01 1.41878694e-01 2.56188899e-01
9.81447041e-01 2.98371911e-01 -3.37711871e-01 -1.40086055e-01
6.31712258e-01 -1.82604834e-01 5.11167586e-01 -1.14870381e+00
-2.76227266e-01 -1.31405801e-01 -1.03975713e+00 -1.09819973e+00
-4.83544976e-01 -5.79874694e-01 6.35239005e-01 -1.18217659e+00
-1.31240517e-01 -8.27227011e-02 -2.98298299e-01 5.20044208e-01
2.12695092e-01 6.88899636e-01 -1.43024176e-01 3.77704293e-01
-3.38595927e-01 2.29752988e-01 6.42782509e-01 -5.91674410e-02
-6.21536255e-01 -1.48202879e-02 -1.96133718e-01 4.26368684e-01
6.48579597e-01 -6.73246801e-01 -5.10709397e-02 -7.28192627e-01
-3.63397509e-01 -3.71527150e-02 1.80806383e-01 -1.63966274e+00
6.91482008e-01 2.95209974e-01 8.46910655e-01 -1.17656782e-01
7.22129524e-01 -9.31057334e-01 2.91746646e-01 7.90736139e-01
-4.00088638e-01 -1.34427741e-01 3.05404544e-01 6.02693498e-01
1.55546188e-01 5.81238866e-01 4.07555670e-01 3.28984410e-01
-5.98561287e-01 6.16661794e-02 -7.57261157e-01 -4.80996147e-02
8.93793881e-01 -6.84091270e-01 3.95331085e-02 -1.77932873e-01
-9.14152205e-01 8.76262859e-02 1.12297580e-01 4.96382743e-01
9.34444070e-01 -1.37315536e+00 -5.03102899e-01 6.93907320e-01
3.41251381e-02 -4.36355293e-01 9.60220490e-03 1.06777680e+00
-2.29534596e-01 2.40658261e-02 -5.94584823e-01 -1.05026186e+00
-1.26971114e+00 -5.57722867e-01 2.92738050e-01 6.47547781e-01
-1.03735685e+00 1.18623352e+00 -1.13353682e+00 2.01447830e-01
8.67974758e-01 -3.30324441e-01 3.82327557e-01 -1.19165950e-01
7.63982534e-01 4.79828358e-01 2.94042140e-01 7.16674179e-02
-7.60919571e-01 9.10401165e-01 2.64324188e-01 -5.79347551e-01
1.58253407e+00 3.00360233e-01 5.88562191e-01 1.06035566e+00
1.11723626e+00 -5.83407164e-01 -1.58571351e+00 1.66076511e-01
-1.39179707e-01 -4.41510499e-01 8.52478519e-02 -1.05983365e+00
-8.64738524e-01 7.95821786e-01 1.38912117e+00 -2.95482606e-01
1.09494698e+00 -2.36896724e-01 1.15322292e+00 6.47188187e-01
8.66071343e-01 -1.42307031e+00 6.31155670e-02 6.36335164e-02
9.16204333e-01 -9.13356245e-01 -6.20826744e-02 2.24647904e-03
-6.72003269e-01 1.12989044e+00 3.89749557e-01 -6.14859462e-01
7.54702389e-01 1.14942336e+00 2.69221514e-01 -2.74022341e-01
-4.45663363e-01 2.46603280e-01 1.74122378e-01 9.30489540e-01
4.30072367e-01 2.68321753e-01 -5.08876801e-01 1.03508937e+00
-8.71880427e-02 7.55259693e-01 -1.13628291e-01 1.43966746e+00
-3.13250095e-01 -6.30835652e-01 -1.03992894e-02 1.15469933e+00
-4.93286610e-01 8.23874325e-02 -4.43960845e-01 4.38502342e-01
-2.40179956e-01 1.03205085e+00 2.66231835e-01 -1.17726767e+00
7.06271589e-01 3.31057638e-01 8.21130574e-01 -6.30977988e-01
-9.02697861e-01 3.98454458e-01 -2.47090198e-02 -1.19278765e+00
-2.62947500e-01 -9.90335405e-01 -1.58054936e+00 1.93106458e-01
-2.61182189e-01 -5.09508610e-01 1.05661583e+00 7.75837779e-01
9.28219259e-01 5.03502965e-01 3.38010460e-01 -1.35166252e+00
-7.66592741e-01 -1.24259353e+00 -1.04476953e+00 6.46881461e-02
1.75296739e-01 -5.69044232e-01 -1.73610628e-01 9.91573557e-02] | [6.808749198913574, 0.1495652049779892] |
307beb6e-b789-40c1-a866-1e3392fa10da | self-constrained-inference-optimization-on | 2207.02425 | null | https://arxiv.org/abs/2207.02425v1 | https://arxiv.org/pdf/2207.02425v1.pdf | Self-Constrained Inference Optimization on Structural Groups for Human Pose Estimation | We observe that human poses exhibit strong group-wise structural correlation and spatial coupling between keypoints due to the biological constraints of different body parts. This group-wise structural correlation can be explored to improve the accuracy and robustness of human pose estimation. In this work, we develop a self-constrained prediction-verification network to characterize and learn the structural correlation between keypoints during training. During the inference stage, the feedback information from the verification network allows us to perform further optimization of pose prediction, which significantly improves the performance of human pose estimation. Specifically, we partition the keypoints into groups according to the biological structure of human body. Within each group, the keypoints are further partitioned into two subsets, high-confidence base keypoints and low-confidence terminal keypoints. We develop a self-constrained prediction-verification network to perform forward and backward predictions between these keypoint subsets. One fundamental challenge in pose estimation, as well as in generic prediction tasks, is that there is no mechanism for us to verify if the obtained pose estimation or prediction results are accurate or not, since the ground truth is not available. Once successfully learned, the verification network serves as an accuracy verification module for the forward pose prediction. During the inference stage, it can be used to guide the local optimization of the pose estimation results of low-confidence keypoints with the self-constrained loss on high-confidence keypoints as the objective function. Our extensive experimental results on benchmark MS COCO and CrowdPose datasets demonstrate that the proposed method can significantly improve the pose estimation results. | ['Zhihai He', 'Zeng Li', 'Shuoshuo Chen', 'Zhehan Kan'] | 2022-07-06 | null | null | null | null | ['inference-optimization', 'multi-person-pose-estimation'] | ['audio', 'computer-vision'] | [-2.42127895e-01 1.35565653e-01 -3.30793470e-01 -3.38129610e-01
-5.77846110e-01 -3.89306456e-01 1.75293639e-01 2.13544935e-01
-3.57977837e-01 6.71723962e-01 1.14349894e-01 3.65902752e-01
-2.27149665e-01 -6.17353857e-01 -9.45346534e-01 -6.00253880e-01
-3.52519631e-01 7.12356865e-01 4.21405733e-01 -2.54541814e-01
3.40454723e-03 4.41528231e-01 -1.43393695e+00 -1.20118715e-01
6.93536103e-01 1.18736851e+00 6.76945820e-02 4.69349325e-01
6.54366076e-01 2.58166075e-01 -4.47829753e-01 -2.39554122e-01
2.94836253e-01 -1.71540409e-01 -5.30683875e-01 -3.48197877e-01
4.10175383e-01 -1.35206133e-01 -5.51367104e-02 8.90130222e-01
7.08473921e-01 3.21964860e-01 3.57831776e-01 -1.29853249e+00
1.94252133e-01 2.79657632e-01 -4.02725101e-01 -1.38982147e-01
7.06963599e-01 1.65459260e-01 9.98950005e-01 -1.03360951e+00
6.64480865e-01 1.17905617e+00 9.90310192e-01 4.66941774e-01
-8.85187447e-01 -8.92452300e-01 2.95898855e-01 2.50587702e-01
-1.82385135e+00 -2.23750576e-01 8.75301838e-01 -4.82894063e-01
4.22658414e-01 2.65604734e-01 1.16125333e+00 8.95759046e-01
4.46514308e-01 7.60012329e-01 6.02336049e-01 -1.20118394e-01
1.20419852e-01 -1.50368333e-01 -1.34944409e-01 9.96201575e-01
1.83682784e-01 2.74744689e-01 -8.65187705e-01 -1.96860954e-01
7.58777976e-01 -4.49134670e-02 -3.72058719e-01 -6.70403957e-01
-1.35271060e+00 4.67010647e-01 9.39691961e-01 -1.82844758e-01
-3.99252266e-01 1.92760587e-01 2.07492903e-01 -3.26925367e-01
1.41481459e-01 6.25825226e-01 -7.30281830e-01 -2.26897039e-02
-7.91742146e-01 5.57915449e-01 6.53246939e-01 8.76958609e-01
7.79832423e-01 -3.99989963e-01 -2.50316650e-01 6.35646343e-01
4.62988436e-01 5.36649704e-01 2.10623935e-01 -7.50328481e-01
6.33862615e-01 7.46661007e-01 1.72079191e-01 -1.51959527e+00
-7.77438581e-01 -5.71770549e-01 -8.65868330e-01 -1.38227921e-02
4.23888654e-01 -1.66454211e-01 -7.83414662e-01 1.86164176e+00
8.81426692e-01 1.27905279e-01 -4.59918261e-01 1.22265685e+00
5.66325128e-01 3.10084373e-01 -1.54324681e-01 2.17257831e-02
1.25006235e+00 -7.43114531e-01 -3.06957155e-01 -2.81658232e-01
4.70738858e-01 -5.14807582e-01 6.87011719e-01 3.12321186e-01
-7.26238430e-01 -9.17150855e-01 -1.13086557e+00 1.85098901e-01
-3.67174530e-03 4.71909910e-01 3.89164001e-01 2.53931761e-01
-3.72918427e-01 7.59098589e-01 -1.01642716e+00 -1.39961988e-01
1.75313786e-01 6.11586988e-01 -5.31865478e-01 2.04515323e-01
-1.30637991e+00 8.74148190e-01 6.62936330e-01 5.57399094e-01
-7.53941417e-01 -8.09583426e-01 -1.00489378e+00 -3.67620170e-01
6.25358701e-01 -7.71312892e-01 8.31318498e-01 -4.37707037e-01
-1.39538682e+00 5.39017797e-01 5.56991957e-02 -2.63807148e-01
9.32712257e-01 -7.04826951e-01 -3.81971337e-02 2.24867407e-02
2.33031854e-01 7.85854459e-01 7.55938530e-01 -1.18727493e+00
-6.87369883e-01 -4.87246394e-01 -1.08013026e-01 3.42281818e-01
1.09207451e-01 -5.37377477e-01 -9.08461809e-01 -7.56871104e-01
3.90569299e-01 -1.43695688e+00 -1.88251600e-01 2.76828587e-01
-7.24430025e-01 -7.05064461e-02 5.10295093e-01 -7.97673702e-01
1.13693750e+00 -1.83468056e+00 5.39150655e-01 7.59424686e-01
9.13449079e-02 -5.80478795e-02 1.40293345e-01 8.07347074e-02
-1.94415003e-02 -2.44133249e-01 7.36752301e-02 -2.46524885e-01
-2.61969626e-01 1.50053695e-01 -3.98280844e-02 7.07243323e-01
1.67262346e-01 1.00857496e+00 -8.11778069e-01 -6.02350652e-01
3.09891433e-01 4.40805644e-01 -7.48871565e-01 4.61290509e-01
-1.77631214e-01 7.96857059e-01 -5.43287635e-01 7.31797218e-01
4.65624809e-01 -2.19017476e-01 9.71628726e-02 -7.31770992e-01
2.37967268e-01 9.81289893e-02 -1.44461048e+00 1.81580973e+00
-5.47724627e-02 6.57892078e-02 -1.64840847e-01 -6.13014698e-01
8.14884424e-01 2.39752373e-03 6.49431586e-01 -2.59322762e-01
2.09842309e-01 -5.06854281e-02 1.50245763e-02 -1.18884251e-01
4.26751107e-01 3.12145305e-04 -2.15177968e-01 2.86750458e-02
-7.97834713e-04 2.67012753e-02 -1.19403459e-01 -1.69559091e-01
6.99802995e-01 4.79015410e-01 3.92479748e-01 -9.81386602e-02
7.07259476e-01 -1.98081627e-01 9.01846707e-01 3.64902586e-01
-1.16581827e-01 7.70915866e-01 1.56014919e-01 -5.30835569e-01
-9.00416195e-01 -1.10513175e+00 2.85870116e-02 9.25890148e-01
5.05639374e-01 -7.82497764e-01 -6.69656694e-01 -7.34970868e-01
2.35542372e-01 -3.17365676e-02 -8.03411543e-01 -5.43796957e-01
-8.53703678e-01 -4.28666651e-01 5.09176850e-01 8.63827050e-01
4.54214424e-01 -6.96116805e-01 -7.31911361e-01 4.14016284e-02
-4.48641539e-01 -9.72433388e-01 -5.15459418e-01 1.27922803e-01
-6.98019207e-01 -1.23025608e+00 -7.40656495e-01 -5.09765625e-01
8.92275214e-01 -2.40919784e-01 7.88695514e-01 4.81659174e-01
-1.74804047e-01 1.23131461e-01 -3.29190254e-01 -1.78195059e-01
3.75481583e-02 1.70170292e-01 5.65097809e-01 -1.25326633e-01
-2.78163135e-01 -3.34852517e-01 -8.10244977e-01 9.14001644e-01
-2.24546969e-01 5.57995476e-02 4.56807792e-01 9.29958403e-01
9.84517455e-01 6.94822297e-02 1.60948336e-01 -3.63587648e-01
1.27448365e-01 -5.18370531e-02 -4.97080147e-01 2.86197335e-01
-3.23427767e-01 2.01308668e-01 3.87928873e-01 -6.11935616e-01
-6.42117798e-01 5.61234236e-01 -2.91342467e-01 -5.32394171e-01
1.03743888e-01 6.48589373e-01 -3.62173229e-01 -2.70639777e-01
5.87895632e-01 3.67246196e-02 3.31963301e-02 -3.99935603e-01
1.52252570e-01 1.04860887e-01 7.20122278e-01 -8.78461599e-01
1.04586911e+00 3.00520271e-01 2.40429536e-01 -6.10529244e-01
-9.04930472e-01 -5.55023670e-01 -1.00180876e+00 -5.13633311e-01
9.54184651e-01 -1.02025163e+00 -1.06288481e+00 4.39394653e-01
-9.47946787e-01 -1.61634639e-01 -2.35543232e-02 5.50310194e-01
-6.53998315e-01 4.47788954e-01 -3.18131268e-01 -7.16896415e-01
-3.73039752e-01 -1.19677544e+00 1.36906385e+00 1.23268053e-01
-6.36855364e-01 -6.06566310e-01 1.22001609e-02 2.80111521e-01
-3.80920559e-01 4.84336019e-01 5.02504289e-01 -4.67096239e-01
-4.44023073e-01 -5.83753884e-01 2.57183552e-01 1.57360375e-01
1.02721862e-01 -1.15112737e-01 -5.11781156e-01 -5.54006219e-01
-3.68476510e-01 -3.95031065e-01 4.80640769e-01 3.78392756e-01
1.13177288e+00 -1.10110179e-01 -6.60213828e-01 7.73297787e-01
7.85584509e-01 -3.79338861e-01 3.45979810e-01 2.69661993e-01
9.31589127e-01 6.21895134e-01 1.26217413e+00 5.12729943e-01
4.02741820e-01 9.14258003e-01 4.23860401e-01 2.14077070e-01
1.64522067e-01 -8.07146609e-01 2.27490976e-01 5.48627019e-01
-3.66469443e-01 1.46182522e-01 -1.03436100e+00 1.62375525e-01
-2.02459073e+00 -6.77798390e-01 2.00346887e-01 2.42765141e+00
8.47659528e-01 3.48221034e-01 2.92777479e-01 1.58642203e-01
7.52907515e-01 -9.74927563e-03 -6.69100106e-01 5.56051612e-01
3.29104036e-01 -7.26223737e-02 3.55922699e-01 4.00367588e-01
-1.27731013e+00 8.65210235e-01 6.04972553e+00 7.90499687e-01
-1.00731027e+00 -4.02411848e-01 2.92936802e-01 3.02932765e-02
2.89148927e-01 -1.95130572e-01 -1.14682877e+00 4.50158805e-01
2.50072896e-01 8.59365165e-02 3.89804244e-02 1.03513277e+00
-4.30515641e-03 -2.58008003e-01 -1.35931730e+00 9.62853611e-01
-1.21233650e-01 -1.08340180e+00 -7.18662739e-02 -6.43041059e-02
5.19021988e-01 -3.61662298e-01 -7.97962174e-02 9.81670395e-02
5.93214110e-02 -9.01277900e-01 9.44738388e-01 7.03781545e-01
6.47544801e-01 -1.08885181e+00 7.69065559e-01 7.51645386e-01
-1.70978844e+00 -6.39049560e-02 -3.08871061e-01 8.78806226e-03
1.32478401e-01 3.89069885e-01 -7.90866911e-01 7.08472073e-01
8.72834802e-01 7.45635450e-01 -5.67135513e-01 1.03329051e+00
-5.87808847e-01 2.01274052e-01 -6.19470417e-01 7.30636045e-02
-3.72050941e-01 1.44318417e-01 6.12502575e-01 7.32047558e-01
1.43268421e-01 9.50965807e-02 6.91582739e-01 7.15839624e-01
2.50762314e-01 -4.40311506e-02 -3.05263884e-02 3.54259312e-01
4.83685702e-01 1.18157673e+00 -7.36710489e-01 -1.08875617e-01
2.85134137e-01 9.66369629e-01 3.94709051e-01 -3.59348655e-02
-1.03873825e+00 -8.05903375e-02 7.31947482e-01 3.22466254e-01
1.06680989e-01 -3.74345303e-01 -1.57026321e-01 -1.18875325e+00
2.26900563e-01 -9.69261706e-01 5.16539574e-01 -6.06881261e-01
-1.10760868e+00 2.87309527e-01 2.25915924e-01 -1.42381108e+00
-4.51944262e-01 -4.31401312e-01 -2.43328556e-01 7.30950773e-01
-7.85704970e-01 -1.32173729e+00 -4.72523659e-01 6.00051820e-01
1.18522204e-01 1.46472201e-01 5.05089521e-01 5.75299412e-02
-4.08395618e-01 1.00260925e+00 -4.59189147e-01 5.11369586e-01
7.67654419e-01 -9.86090779e-01 1.95125133e-01 6.49811149e-01
-1.38081042e-02 9.14756119e-01 8.84596288e-01 -1.12760901e+00
-1.32945776e+00 -1.07791495e+00 4.01039958e-01 -6.40733123e-01
3.17750663e-01 -5.58444321e-01 -7.28831947e-01 5.31057656e-01
-8.94844294e-01 3.15597475e-01 5.31504750e-01 3.16813231e-01
-2.07869530e-01 -1.40791342e-01 -9.57747161e-01 4.99345452e-01
1.17068398e+00 -3.64034355e-01 -7.88453221e-01 2.67426848e-01
6.56265140e-01 -9.26162243e-01 -1.02715397e+00 9.12347317e-01
1.10127401e+00 -5.68116486e-01 1.32573760e+00 -4.28857565e-01
2.56957620e-01 -5.69592595e-01 -7.41513744e-02 -1.16629076e+00
-3.39888245e-01 -2.11073950e-01 -2.75897771e-01 8.76932442e-01
2.08561331e-01 -2.29842544e-01 1.30005729e+00 5.87574065e-01
1.91608891e-01 -9.62032020e-01 -1.12595725e+00 -7.81516314e-01
-2.01933995e-01 -5.49899876e-01 5.07854104e-01 5.54983139e-01
-1.00725904e-01 1.06033348e-01 -6.79120123e-01 5.26292562e-01
7.13403702e-01 9.98423770e-02 1.23466539e+00 -1.28669572e+00
-4.65982199e-01 2.88470946e-02 -7.93300867e-01 -1.42036569e+00
1.03515185e-01 -4.87530679e-01 5.13009846e-01 -1.09364951e+00
6.27127513e-02 -5.68473637e-01 -1.94864273e-01 4.94906902e-01
-5.05925894e-01 2.75144577e-01 2.93799073e-01 2.67847776e-01
-6.86251342e-01 7.23844051e-01 1.27109063e+00 1.19002629e-02
-2.33678415e-01 3.29049110e-01 -1.86948314e-01 9.68205094e-01
4.62568611e-01 -4.45916146e-01 -2.11821064e-01 2.05301747e-01
4.29661751e-01 1.08430989e-01 6.80850089e-01 -1.44157386e+00
3.49015474e-01 -1.69653758e-01 9.95047092e-01 -1.09040487e+00
4.83946830e-01 -8.28000546e-01 3.32791865e-01 8.61272037e-01
-2.07728282e-01 -7.60877728e-02 -1.44243706e-03 7.14554846e-01
-7.23321214e-02 2.48543873e-01 7.47240007e-01 -1.53120039e-02
-6.57404423e-01 6.19062781e-01 1.74391448e-01 -1.40494350e-02
1.06692469e+00 -4.22517747e-01 2.57364929e-01 -3.34130287e-01
-9.79323268e-01 5.36436915e-01 5.16338110e-01 5.53838909e-01
7.86510050e-01 -1.56183875e+00 -3.25429261e-01 2.85047948e-01
3.61561686e-01 4.10221606e-01 2.01066643e-01 8.97004247e-01
-3.01554769e-01 1.83856994e-01 -3.17645401e-01 -1.04266989e+00
-1.49812758e+00 3.79841268e-01 5.32282352e-01 -1.95752218e-01
-4.63492006e-01 1.07294357e+00 1.95608094e-01 -6.76252782e-01
3.00671965e-01 -4.73330379e-01 -7.49407709e-02 -2.93229353e-02
3.53159010e-01 4.17558730e-01 -6.89167827e-02 -1.00877881e+00
-7.76019216e-01 9.84667003e-01 1.39058158e-01 1.24158524e-01
1.21935594e+00 -1.17215654e-02 9.73324105e-03 3.84059310e-01
1.16147041e+00 -5.80997160e-03 -1.34609556e+00 -1.48859650e-01
-2.36549973e-01 -4.78976429e-01 -3.99570405e-01 -6.71813309e-01
-9.21061039e-01 6.16572499e-01 5.79997003e-01 -5.81333995e-01
8.38750839e-01 1.03894182e-01 6.70441687e-01 4.64468569e-01
6.70366764e-01 -1.33005297e+00 3.60179931e-01 5.35076082e-01
1.16138625e+00 -1.19296265e+00 4.48379129e-01 -6.13879204e-01
-4.14675504e-01 9.35185373e-01 1.03178942e+00 -2.67851144e-01
7.80536175e-01 9.28123817e-02 -2.10786849e-01 -1.66684285e-01
-3.57121885e-01 -3.25567164e-02 1.04456317e+00 6.47546709e-01
2.49991730e-01 7.68911019e-02 -1.72350958e-01 7.78449118e-01
-6.70260727e-01 -1.15509793e-01 -3.61463100e-01 8.37891221e-01
-5.68132818e-01 -9.29395497e-01 -6.47418439e-01 2.11128980e-01
-1.72935441e-01 3.33487481e-01 -4.80021238e-01 8.58070254e-01
4.07910526e-01 5.29181421e-01 -2.75394082e-01 -1.02736270e+00
4.59055066e-01 -1.87994242e-01 4.84828442e-01 -5.40756941e-01
-5.19472718e-01 -7.43135735e-02 8.75134766e-03 -9.08256292e-01
-1.00398511e-01 -5.63750744e-01 -1.51140535e+00 -1.66566312e-01
-6.54404461e-01 1.48465827e-01 3.49295139e-01 9.83354747e-01
1.54267803e-01 3.61273527e-01 3.09049398e-01 -1.11229289e+00
-5.33935487e-01 -7.24501669e-01 -1.35305673e-01 3.96758437e-01
3.40457648e-01 -1.14054441e+00 5.82646020e-02 -2.37364098e-01] | [7.057668685913086, -0.862555205821991] |
ff6ad74a-eae9-4007-861a-56bc2ed37bda | guided-slot-attention-for-unsupervised-video | 2303.08314 | null | https://arxiv.org/abs/2303.08314v1 | https://arxiv.org/pdf/2303.08314v1.pdf | Guided Slot Attention for Unsupervised Video Object Segmentation | Unsupervised video object segmentation aims to segment the most prominent object in a video sequence. However, the existence of complex backgrounds and multiple foreground objects make this task challenging. To address this issue, we propose a guided slot attention network to reinforce spatial structural information and obtain better foreground--background separation. The foreground and background slots, which are initialized with query guidance, are iteratively refined based on interactions with template information. Furthermore, to improve slot--template interaction and effectively fuse global and local features in the target and reference frames, K-nearest neighbors filtering and a feature aggregation transformer are introduced. The proposed model achieves state-of-the-art performance on two popular datasets. Additionally, we demonstrate the robustness of the proposed model in challenging scenes through various comparative experiments. | ['Sangyoun Lee', 'Jungho Lee', 'Chaewon Park', 'Dogyoon Lee', 'Suhwan Cho', 'Minhyeok Lee'] | 2023-03-15 | null | null | null | null | ['video-object-segmentation', 'video-semantic-segmentation', 'unsupervised-video-object-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 0.3250427 -0.3860547 -0.22344491 -0.44157916 -0.66072166 -0.31183377
0.22643471 -0.11953242 -0.45530334 0.53283226 0.1297529 0.02806459
-0.08414364 -0.5188712 -0.51505 -0.8900051 0.18748954 0.01905362
1.0383058 0.1544415 0.30088368 0.46071634 -1.4619257 0.4471816
0.9595632 1.0703572 0.6234672 0.41571346 -0.38461497 0.92432415
-0.66461825 -0.06896529 0.26687703 -0.3704165 -0.66933835 0.74369407
0.6100017 -0.50341994 -0.46751088 1.2854192 0.24512073 0.4964162
0.2464346 -1.0898501 -0.3868533 0.44461006 -0.850736 0.97322243
0.01361611 0.12953825 0.8353882 -0.99853635 0.5148477 1.1785346
0.04600498 0.27119055 -1.0130477 -0.49191448 0.856283 0.5341719
-1.4517386 -0.34442192 0.94439125 -0.27785978 0.5293903 0.20761721
0.68610317 0.5847419 -0.07512043 1.212516 0.5392532 -0.10863742
-0.03001497 -0.03274489 0.3279496 0.66341007 0.13068739 -0.19755213
-0.43266776 0.22036283 0.9293096 0.33462304 -0.49255198 -0.47218487
-1.1312248 0.47888163 0.49956146 0.2650514 -0.50341725 0.04385569
0.18014799 -0.45112342 0.26258948 -0.0176201 -0.43612275 0.04976406
-1.0099093 0.07115852 0.1338223 1.0886686 0.79349756 0.05280134
-0.62445265 0.83221406 0.4738339 0.29470995 0.17736241 -1.0983037
0.45429507 0.7286804 0.12581052 -1.2147126 -0.15411736 -0.5307551
-0.6416852 -0.2119727 0.34545127 0.05545018 -1.1775808 1.2496383
0.69512266 0.7268876 -0.13017146 1.3117543 0.9688279 0.87882483
0.29868382 -0.51207 1.2796104 -1.3725994 -0.9462584 -0.32945567
-0.02552242 -0.7888487 0.73843545 0.10919369 -1.1664238 -0.89057547
-0.70544136 -0.11193332 -0.03248678 0.27161378 0.3998046 0.304175
-0.58553326 0.2189008 -0.86622363 -0.13503033 0.7335551 0.35115644
0.07493819 -0.15061058 -0.9070281 0.28139442 0.60906667 0.3763498
-0.9045972 -0.43134943 -0.66484845 0.1735993 0.8819386 -0.41058344
1.06021 -1.0517797 -1.2438815 0.5034837 -0.43973124 -0.29971603
0.37407097 -0.43268913 -0.33883265 0.47587475 0.18830995 0.7520091
0.77312696 -1.1809697 -1.2216994 -0.3147861 0.04960928 0.41388306
-0.08688404 0.22492193 -1.3730614 -0.88998836 0.4960411 -0.5111687
-0.46389264 -0.1062775 -0.29627317 -0.12300278 1.277012 -0.7039885
1.3358797 -2.3389273 0.18763055 0.12412114 0.2899322 0.38887554
-0.14670008 -0.34601542 0.12636384 -0.15639132 0.05611851 0.06895011
-0.26959568 0.19196709 -0.0232718 0.381809 0.35042766 0.8093109
-0.79799604 -0.8603816 0.45212194 0.33968213 -0.4901563 0.3305129
-0.29345524 0.5738919 -0.64888436 0.8242142 0.7961321 -0.4466627
-0.08507114 -0.4752023 -0.09703037 0.00794427 -1.4149531 1.4611666
0.3929249 0.6252431 0.12196043 -0.9648681 0.6659727 -0.06797182
0.64023566 -0.75910825 0.32535326 -0.11233408 0.14903603 -0.54677683
0.6524694 0.44839942 0.38024586 -0.09376939 -0.09326194 0.42461237
0.4730215 0.18285291 0.590863 0.22186731 -0.00977849 -0.3271358
0.84417003 -0.17187539 1.1523317 0.57959604 -0.58301425 0.6715801
0.30737162 -0.51677483 -0.61160856 -1.0171672 0.11353216 1.3108287
0.95100224 -0.21225984 -0.8826623 -0.75665313 -0.32380033 0.469278
-0.45162794 0.12922095 -0.6609132 -0.6870098 0.03901315 0.6389516
0.7617555 -0.9474119 -0.6966741 0.29789242 -0.4769653 -1.3940034
-0.8353195 -0.16834109 -0.79742575 -1.242226 -0.8435231 -0.9644909
0.75479954 0.7020703 0.87346095 0.28967836 -0.2980044 0.13869458
-0.3293813 -0.05700762 0.15935382 -0.07406907 -0.19975375 0.4254235
0.31250408 -0.07196432 -0.8795147 0.67541754 -0.9797532 0.230176
0.522685 0.6066453 0.779623 0.22563681 0.17556037 -0.55218476
0.02697525 -0.12279311 -0.8679359 0.3918965 0.01797059 -0.32370436
0.13694932 -0.4791835 -1.145487 0.13074976 0.29916084 -0.6791813
-0.2307115 0.15897848 -0.59464514 0.10431796 0.1768596 0.3230483
-0.4550002 -0.38458258 0.23872262 0.43970203 0.67794293 -0.45915717
0.6363405 0.40024245 -0.3620978 -0.7146284 -0.84615195 -0.7026859
-0.7986293 -0.43800503 1.1483557 -0.94676167 -0.48596275 0.5027022
-1.1569752 -0.22593787 -0.07418348 0.45856148 -0.20318179 0.45251623
-0.57712704 -0.81796503 -0.09533454 -1.4774361 1.025785 0.78459287
0.3226139 -0.58112407 -0.6241261 0.3466836 0.13703465 -0.05033749
0.49698132 -0.5636448 -1.4025236 0.17901273 -0.69824535 0.11551102
0.38361725 0.29329416 -0.62950706 -0.0522871 -0.20190653 0.37314457
0.96408725 0.6394043 1.2727157 -0.16207628 -0.46545926 0.6172697
1.0492126 0.63297397 0.5934676 0.27867994 0.95273167 0.47816145
1.0968502 0.4546641 0.18424802 0.636664 0.24890949 -0.20937699
-0.06466182 0.11622761 0.09688328 0.49086463 0.06825034 -0.17661111
-0.7867198 0.61391425 -1.9940066 -1.0057193 -0.06371773 1.9179487
0.5281951 0.39713812 0.18901421 -0.20384067 1.0510052 0.24755007
-0.6107916 0.5280565 -0.31692794 -0.2957943 0.40363932 0.41735685
-1.43536 1.2199612 5.9322696 0.9449629 -0.9962274 -0.07855362
1.0876255 -0.2290909 0.02454874 -0.13019918 -0.8862169 0.64228195
0.17637263 -0.03555125 0.33492926 0.7936695 0.33636013 -0.39911512
-0.8328249 1.0797545 0.00701525 -1.4184147 0.09337715 -0.2226098
0.798737 -0.1982589 -0.02414997 0.16375048 -0.07002862 -0.5938653
0.7799447 0.51908034 0.27921185 -0.8178377 0.6216496 0.10995504
-1.5025704 -0.22615716 -0.28078437 0.22016059 0.3131479 0.39343372
-0.3717442 0.4438213 0.9270477 0.6833762 -0.65686244 1.3780271
0.01566896 0.412401 -0.26240838 0.08292656 0.37610307 -0.28948334
0.5036229 1.1628693 -0.02311383 0.6000812 0.6260568 0.7459518
0.17220064 0.16428423 0.10613891 0.03465358 0.37409815 1.2059587
-1.2636757 -0.5975635 -0.4598985 0.9494108 0.09411849 0.73279256
-1.2107143 -0.2276541 0.6519413 0.00987414 0.7695627 -0.23183063
0.01692588 -1.1309236 -0.00806354 -0.85276574 0.44857422 -0.74500406
-0.86913526 0.51415986 0.10327 -1.0816187 0.27972972 -0.37992546
-0.54352033 0.51540273 -1.4227391 -0.83718663 -0.52723426 0.655445
1.0177858 -0.07153948 -0.06625643 0.57909155 -0.9926562 0.3430286
-0.07260428 0.4763793 0.3506215 -0.8310554 0.19626485 1.171921
0.24704257 0.45391366 0.46006924 -0.7613586 -1.0140811 -1.2930692
0.21411173 -0.01444044 0.32808062 -0.05836434 -1.0683978 0.46480215
0.06102713 0.35032502 0.40910438 -0.25016344 0.07502966 -0.23496236
-0.79399765 0.7366364 1.1341825 -0.17681262 -0.45105812 0.24997419
0.85217947 -0.66796005 -0.35774365 0.6247666 0.3664691 -0.9670353
0.974132 -0.4552797 0.03988189 -0.8833315 -0.331858 -0.62100875
-0.4113824 -0.5618423 -0.11806946 1.2936205 0.10085665 -0.10081399
0.96165913 0.63091433 -0.00707611 -0.79287237 -0.7413343 -0.3877282
-0.6703031 -0.3211211 0.49842072 0.6398914 -0.58281267 0.25435588
-0.273655 0.55827993 0.60476506 0.27741382 0.7510279 -1.022227
-0.0671572 -0.4987578 -0.4979999 -1.4893571 -0.03199063 -0.28418058
0.29031146 -1.4521401 0.44448754 -0.22581933 -0.6868194 0.1248381
-0.72450674 0.22911379 0.4568628 0.1324917 -1.331677 0.591233
1.2785348 -0.2831366 -0.4386017 0.03144282 -0.40875083 0.81023616
0.5652686 -0.3200382 -0.33216134 -0.5315869 -0.6081104 0.06514361
0.2568823 -1.1290163 0.36230752 -0.5150868 0.69573045 -1.0619025
0.3625495 -0.92332655 -0.08751921 0.18603101 -0.2169126 -0.05585763
0.31222034 0.728518 -0.272103 -0.04618119 0.8033037 0.05951713
-1.1652828 0.5554941 -0.2651025 -0.01699294 1.2817315 -0.45256543
-0.17276224 -0.14134818 -0.8227208 0.5327396 0.30586892 0.49933603
0.77109015 -1.2453504 -0.441277 0.36321065 -0.03021366 0.2847134
0.6548849 0.8047792 -0.552516 0.25626904 -0.09324111 -0.99717873
-1.5753013 0.71524024 0.4556177 0.0837853 -0.4708122 0.9730822
0.71990687 0.18193538 0.5647889 -0.53908646 -0.32106754 -0.02451267
0.7412265 0.3275914 -0.4179093 -0.9106054 -0.39156994 0.6380711
-0.26172444 0.10988887 0.90264654 -0.5148082 -0.05262221 0.18602544
0.8970614 -0.01581296 -1.6821373 -0.5604982 -0.00816496 -0.9827363
0.00795282 -0.42641765 -1.4474297 0.7949291 0.7330503 0.077508
1.2972618 -0.06509511 0.6394898 0.23452857 0.02958871 -1.1200513
0.31790942 0.41327652 0.49359873 -1.2097347 0.06871817 -0.79349154
-0.5881637 0.8154427 1.137476 0.11695688 0.37417093 0.0264806
0.24835332 0.08292819 -0.46069172 -0.5119151 0.51179 0.44416988
0.16198228 -0.40857023 -0.0734458 0.5717725 0.5558021 -0.13609138
0.13925257 0.9408744 -0.7426364 -0.75461316 -0.64020056 0.36533806
-0.6970415 0.01449359 -0.07687802 0.5879059 0.2105263 1.0881916
0.36558914 -0.11461847 0.19728133 -0.24026594 0.25721875 -0.40424323
-0.28704745 0.88007504 -0.2845364 -0.6478589 -0.755067 -0.62030005
-1.4222039 0.13735543 -0.64997727 0.19673486 0.04981346 0.9143068
0.39174873 0.96177435 0.4377469 -0.9688227 0.04114875 -0.73214287
-0.48238245 0.42901722 0.33108723 -0.7215055 0.15335774 0.28394172] | [9.252098083496094, -0.29938170313835144] |
7b9562ab-a8c0-4dc6-8165-28d8ac29b8f5 | weakly-supervised-action-localization-with-2 | 2004.00163 | null | https://arxiv.org/abs/2004.00163v2 | https://arxiv.org/pdf/2004.00163v2.pdf | Weakly-Supervised Action Localization with Expectation-Maximization Multi-Instance Learning | Weakly-supervised action localization requires training a model to localize the action segments in the video given only video level action label. It can be solved under the Multiple Instance Learning (MIL) framework, where a bag (video) contains multiple instances (action segments). Since only the bag's label is known, the main challenge is assigning which key instances within the bag to trigger the bag's label. Most previous models use attention-based approaches applying attentions to generate the bag's representation from instances, and then train it via the bag's classification. These models, however, implicitly violate the MIL assumption that instances in negative bags should be uniformly negative. In this work, we explicitly model the key instances assignment as a hidden variable and adopt an Expectation-Maximization (EM) framework. We derive two pseudo-label generation schemes to model the E and M process and iteratively optimize the likelihood lower bound. We show that our EM-MIL approach more accurately models both the learning objective and the MIL assumptions. It achieves state-of-the-art performance on two standard benchmarks, THUMOS14 and ActivityNet1.2. | ['Huijuan Xu', 'Fang Wan', 'Zhekun Luo', 'Baifeng Shi', 'Devin Guillory', 'Wei Ke', 'Trevor Darrell'] | 2020-03-31 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/6965_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123740715.pdf | eccv-2020-8 | ['weakly-supervised-action-localization'] | ['computer-vision'] | [ 5.83578467e-01 3.03741872e-01 -8.17897499e-01 -5.57008684e-01
-1.23141336e+00 -3.60974193e-01 4.99867171e-01 -1.14321388e-01
-4.33280766e-01 8.10601890e-01 2.24349812e-01 4.43622768e-02
2.78292060e-01 -5.00834882e-01 -1.11428082e+00 -9.28041160e-01
-4.82742637e-02 4.78269637e-01 1.71747997e-01 5.06790221e-01
1.19831435e-01 -5.11143319e-02 -1.38919508e+00 7.16034174e-01
4.52421427e-01 1.10968363e+00 1.86728835e-01 7.37677813e-01
-1.06291123e-01 1.63850200e+00 -6.70166731e-01 -2.27642789e-01
2.14985207e-01 -6.88609600e-01 -1.09350610e+00 6.35356009e-01
5.02174437e-01 -5.34867942e-01 -3.53394091e-01 8.98259699e-01
1.43431842e-01 2.92884290e-01 7.32052088e-01 -1.71793163e+00
-4.51359689e-01 4.33240861e-01 -6.42078578e-01 5.17405234e-02
2.66731232e-01 1.10184625e-01 1.20932865e+00 -7.36282170e-01
5.97279429e-01 1.09397304e+00 3.74269426e-01 8.79611909e-01
-1.00676060e+00 -4.53573406e-01 7.20660150e-01 4.50598150e-01
-1.17500293e+00 -2.99452662e-01 5.50938666e-01 -5.44138491e-01
8.37961733e-01 5.99047057e-02 6.84379995e-01 1.11484957e+00
-1.73046678e-01 1.47666740e+00 8.85763705e-01 -4.34907168e-01
4.67941940e-01 -4.82388362e-02 9.01450813e-02 6.98117077e-01
-1.43215373e-01 -4.30603951e-01 -7.15842485e-01 -6.42054081e-02
7.28286207e-01 1.05595514e-01 -1.32440656e-01 -5.20424426e-01
-1.25374663e+00 7.82834172e-01 2.07103401e-01 -1.51040629e-02
-6.01297379e-01 8.84029806e-01 3.54909152e-01 -3.10866594e-01
4.70590770e-01 8.19691345e-02 -5.58539510e-01 -2.36484677e-01
-8.42635274e-01 2.47959763e-01 5.76154768e-01 1.22610021e+00
7.78655887e-01 -3.63773704e-01 -5.63940704e-01 6.97229147e-01
3.29070389e-01 1.49730921e-01 3.42957169e-01 -1.16780043e+00
6.33656979e-01 4.72446263e-01 2.78409272e-01 -5.11212826e-01
2.76642684e-02 -1.57824099e-01 -3.13749164e-01 -6.91903830e-02
3.91565681e-01 -4.57715914e-02 -1.15498817e+00 1.94512165e+00
4.56549942e-01 7.43779242e-01 3.05191008e-03 8.71909261e-01
6.72122836e-01 7.18495727e-01 5.19142210e-01 -2.02243865e-01
9.74286437e-01 -1.47502303e+00 -7.74460077e-01 -5.99928141e-01
9.60041344e-01 -1.63743526e-01 8.03148746e-01 1.89193293e-01
-1.06085742e+00 -4.46071386e-01 -7.62975931e-01 3.31984907e-02
-1.34045392e-01 4.17845458e-01 7.07832396e-01 1.85069948e-01
-8.42371523e-01 3.80069077e-01 -9.65182245e-01 -2.09695131e-01
7.87667572e-01 2.77618289e-01 -4.12640274e-01 -2.09016219e-01
-9.40692961e-01 6.04914486e-01 4.02909726e-01 7.55530298e-02
-1.46722376e+00 -2.93946892e-01 -1.02870882e+00 -5.53527363e-02
6.66668057e-01 -4.98516649e-01 1.43746483e+00 -1.34834802e+00
-1.16919041e+00 1.01601112e+00 -4.70989913e-01 -5.57211578e-01
3.77456754e-01 -3.12695235e-01 8.43408704e-02 2.96815634e-01
2.76907355e-01 1.11496770e+00 9.19125259e-01 -1.34862578e+00
-9.90180135e-01 -1.25707805e-01 5.33834755e-01 3.79376173e-01
-3.45128737e-02 -9.21444688e-03 -6.88219726e-01 -5.26741803e-01
1.04229324e-01 -9.49544489e-01 -3.34816754e-01 -1.92801997e-01
-4.07754421e-01 -3.36870819e-01 4.64566171e-01 -6.35778725e-01
1.20484757e+00 -2.08866954e+00 1.89313173e-01 -1.78102657e-01
5.96782975e-02 6.28968747e-03 -1.26629442e-01 5.67671396e-02
-5.53791039e-02 3.98167744e-02 -9.99798253e-02 -6.69468641e-01
1.01836778e-01 4.71642643e-01 -2.30094314e-01 5.84162414e-01
2.70858347e-01 9.85560119e-01 -1.07767963e+00 -7.15915263e-01
1.64685458e-01 2.94026196e-01 -7.42882013e-01 4.95581508e-01
-6.16035998e-01 5.05791187e-01 -4.45442677e-01 7.14341998e-01
2.56648332e-01 -5.48116505e-01 2.84127921e-01 -2.34087870e-01
2.43854761e-01 2.07742527e-01 -1.14179087e+00 1.61957383e+00
-2.02683732e-01 5.04192173e-01 -1.90752521e-01 -1.16198218e+00
3.71890038e-01 3.95949990e-01 8.16466689e-01 -1.47134662e-01
-2.22873427e-02 -1.31804124e-01 -3.47228378e-01 -7.18241096e-01
1.07149608e-01 -1.58599854e-01 -6.73881844e-02 4.66292858e-01
3.62358958e-01 3.69456738e-01 2.79780835e-01 2.84385562e-01
1.24311495e+00 6.18431449e-01 2.83206433e-01 2.47751340e-01
4.33339596e-01 -4.56105731e-02 7.25595117e-01 9.77759957e-01
-3.20219278e-01 6.02656603e-01 7.35282838e-01 -2.93804675e-01
-5.74675262e-01 -7.63576627e-01 2.27800652e-01 1.39119852e+00
1.52329326e-01 -5.40008485e-01 -9.60018277e-01 -1.22252095e+00
-3.01758587e-01 6.60016358e-01 -8.46636236e-01 -1.95434928e-01
-5.03520429e-01 -5.67589164e-01 2.10909322e-01 8.04488063e-01
3.97618383e-01 -1.25237107e+00 -5.81681907e-01 2.30381772e-01
-6.32666528e-01 -1.27242553e+00 -5.34799635e-01 3.31132531e-01
-6.53809905e-01 -1.22189045e+00 -5.61389446e-01 -6.12923622e-01
9.17509675e-01 1.93775985e-02 1.17609608e+00 8.50122273e-02
-1.71593502e-01 6.84594274e-01 -5.42847514e-01 -3.22521389e-01
-1.09347172e-01 -1.25927240e-01 -2.17455193e-01 4.13462698e-01
5.83541751e-01 -3.54545750e-02 -5.32686055e-01 2.33458847e-01
-7.57484198e-01 3.15019429e-01 5.25495589e-01 7.74881959e-01
9.82629240e-01 -3.78766358e-02 4.87022907e-01 -8.38452518e-01
-9.60031897e-02 -6.17594063e-01 -2.56393164e-01 4.69146788e-01
-1.11141086e-01 -7.84081444e-02 1.24222487e-01 -5.94738483e-01
-8.12004030e-01 4.39805478e-01 1.02843590e-01 -7.05030382e-01
-4.03888524e-01 3.24690968e-01 -4.35262501e-01 3.12120706e-01
1.37292564e-01 1.46412253e-01 -3.29209626e-01 -3.70596260e-01
2.48631120e-01 5.23808420e-01 3.53504688e-01 -5.73809803e-01
2.19716579e-01 5.61374843e-01 -1.27491012e-01 -4.00405169e-01
-1.61436081e+00 -6.85161650e-01 -8.16756487e-01 -5.92611313e-01
1.27433228e+00 -1.17345929e+00 -5.90795636e-01 4.86652613e-01
-1.13469207e+00 -8.10685813e-01 -3.28957647e-01 6.16885185e-01
-9.77888703e-01 1.14169016e-01 -5.48709333e-01 -9.90319669e-01
1.78535700e-01 -1.17123890e+00 1.42849708e+00 2.75608264e-02
-1.10700585e-01 -9.30682659e-01 -1.07824979e-02 6.11213803e-01
-1.82347789e-01 2.43675530e-01 5.47282636e-01 -5.91298044e-01
-8.24901938e-01 -3.43868226e-01 -2.54110303e-02 4.89135683e-01
1.22131079e-01 -3.59006792e-01 -1.10398352e+00 -1.67644963e-01
-1.65227070e-01 -5.71097076e-01 9.99979675e-01 5.91072559e-01
1.60807872e+00 -4.34995979e-01 -3.89243215e-01 4.32379961e-01
1.18336332e+00 1.69102877e-01 8.09150279e-01 2.09813505e-01
8.67314875e-01 4.68815148e-01 1.04648626e+00 4.83665138e-01
4.16342825e-01 8.07512224e-01 6.28733814e-01 -4.58850525e-02
3.52474675e-02 -4.43525940e-01 6.49060547e-01 2.68647522e-01
3.63516598e-03 -4.61699277e-01 -5.71551383e-01 5.76875865e-01
-2.25343490e+00 -1.19816506e+00 3.25389579e-02 2.17940521e+00
9.73096609e-01 8.00988078e-02 1.39714912e-01 -1.12281553e-01
8.24941933e-01 2.95023501e-01 -6.06830537e-01 1.90465346e-01
1.92843601e-01 -5.49402982e-02 5.72311938e-01 6.02380216e-01
-1.54919231e+00 1.08727741e+00 6.39087343e+00 6.86073124e-01
-6.24300003e-01 3.87814790e-01 8.57699037e-01 -4.19459939e-01
2.38075495e-01 1.07320040e-01 -1.17731905e+00 5.96590936e-01
7.69800127e-01 3.05032670e-01 3.09132189e-01 1.02519798e+00
1.80921733e-01 -3.34124416e-01 -1.49567020e+00 1.00219691e+00
4.43220884e-01 -1.13543725e+00 -1.35244550e-02 1.49926201e-01
7.79689848e-01 -1.84546798e-01 -1.72442645e-01 5.22912443e-01
2.78896868e-01 -9.11426604e-01 9.70285773e-01 6.23284638e-01
6.71924472e-01 -4.40200180e-01 5.21250546e-01 3.82496834e-01
-1.09323835e+00 -2.26369008e-01 -2.96539634e-01 -4.74959835e-02
3.69148165e-01 3.64389658e-01 -7.28591502e-01 3.61104496e-02
5.89941919e-01 9.66779530e-01 -4.56098080e-01 9.76053298e-01
-5.42272925e-01 8.88430595e-01 2.88765654e-02 2.25450560e-01
5.12845993e-01 -5.63423336e-02 2.13462159e-01 1.14994717e+00
2.67011523e-02 1.50713533e-01 6.18347406e-01 6.93162084e-01
-1.34017661e-01 -7.59880766e-02 -2.99819320e-01 -1.81675524e-01
1.62705645e-01 1.14174032e+00 -7.04202533e-01 -6.58760548e-01
-5.89709997e-01 1.14080536e+00 4.30684835e-01 5.23070157e-01
-1.21913803e+00 2.26970851e-01 6.80029988e-01 9.54719707e-02
3.80040854e-01 1.50127381e-01 9.75643322e-02 -1.04541743e+00
3.78379896e-02 -7.55607784e-01 5.75248778e-01 -1.01496899e+00
-1.00076091e+00 1.46351948e-01 1.35710612e-01 -1.22960496e+00
-2.90149063e-01 -5.72708368e-01 -3.12267661e-01 5.07433891e-01
-1.28734529e+00 -1.27767682e+00 -2.90393054e-01 5.64004302e-01
7.93642104e-01 1.89459264e-01 5.66783786e-01 2.76850492e-01
-7.11681783e-01 3.63678515e-01 -3.61737460e-01 3.21982980e-01
6.54524446e-01 -1.41971493e+00 -8.81897360e-02 6.55096889e-01
3.27010006e-01 1.76198527e-01 5.39080620e-01 -7.04415739e-01
-9.79812205e-01 -1.43522477e+00 9.74729359e-01 -7.56365180e-01
4.95342165e-01 -3.80354196e-01 -6.45753145e-01 1.30324578e+00
-9.03137401e-02 3.23720187e-01 7.73920476e-01 -2.12385297e-01
-8.05250108e-02 1.77702978e-01 -8.64537239e-01 3.45440000e-01
1.19017458e+00 -4.26323593e-01 -3.00806463e-01 8.14375818e-01
5.54378331e-01 -4.41729844e-01 -6.14381015e-01 3.34424227e-01
2.92487621e-01 -6.97211206e-01 8.58316958e-01 -1.19938922e+00
6.43954217e-01 -3.39623988e-01 -3.30580771e-01 -9.92299199e-01
-2.89433122e-01 -2.58684903e-01 -6.78216279e-01 1.17252445e+00
4.34923202e-01 -6.24052882e-02 9.31115448e-01 7.91674674e-01
-5.52258044e-02 -1.01115870e+00 -8.36643696e-01 -6.35723650e-01
-3.55464160e-01 -5.64943910e-01 4.04658854e-01 7.67327785e-01
-2.22354792e-02 2.09195212e-01 -7.09357500e-01 1.93580031e-01
6.05558455e-01 -5.92870601e-02 8.29337597e-01 -7.38063633e-01
-5.70726216e-01 4.40251678e-02 -4.67454344e-01 -1.29303873e+00
6.39426708e-01 -7.22258151e-01 4.72230732e-01 -1.78201509e+00
6.20010734e-01 -3.50082129e-01 -5.20716190e-01 9.49966848e-01
-3.60036373e-01 2.29439676e-01 1.72167078e-01 1.44024223e-01
-1.37256658e+00 3.55254441e-01 9.76346731e-01 -1.55261263e-01
2.68943422e-02 7.96438679e-02 -4.90894288e-01 9.15049374e-01
7.15266109e-01 -6.67652845e-01 -4.48001027e-01 -3.90985608e-01
1.72362819e-01 -3.76808159e-02 5.92975140e-01 -9.59820688e-01
1.28722206e-01 -4.78708476e-01 3.85698497e-01 -5.35589457e-01
5.70863664e-01 -7.92411983e-01 -2.15311777e-02 1.21653318e-01
-8.43032122e-01 -4.52266723e-01 -2.67851919e-01 8.32647085e-01
-1.48157537e-01 -3.93683225e-01 7.44047821e-01 -3.46762061e-01
-8.84908676e-01 5.64372838e-01 -3.38293850e-01 1.21983580e-01
1.36712754e+00 -1.35567307e-01 -7.96080977e-02 -5.28359830e-01
-1.09792554e+00 3.93529058e-01 4.09729928e-01 3.26890171e-01
4.50936884e-01 -1.50101507e+00 -5.55338800e-01 -2.58743903e-03
2.01080680e-01 1.17035080e-02 2.45450988e-01 1.06173265e+00
1.61827879e-03 2.85710484e-01 2.49731019e-01 -6.87496483e-01
-1.25050437e+00 6.03906631e-01 4.86626148e-01 -3.93193722e-01
-4.34986800e-01 1.06364083e+00 5.54974020e-01 -1.42155096e-01
6.45231307e-01 -3.23642753e-02 -1.51896000e-01 -5.33917658e-02
6.63246930e-01 3.22938740e-01 -3.71583670e-01 -8.91518772e-01
-3.70865375e-01 1.94097713e-01 2.52319630e-02 -1.40859887e-01
1.14449251e+00 -4.87408862e-02 1.14917757e-05 6.28315747e-01
1.20716500e+00 -4.25587535e-01 -1.76132393e+00 -1.91722989e-01
-6.88566417e-02 -6.71088338e-01 -7.50619397e-02 -7.41192520e-01
-1.08786392e+00 7.45058239e-01 3.36963594e-01 -1.47799507e-01
8.79353404e-01 4.37597930e-01 5.85276723e-01 1.81708246e-01
4.57620382e-01 -1.41462743e+00 4.44638342e-01 3.15832525e-01
6.12027526e-01 -1.23226607e+00 -1.43884838e-01 -4.40538228e-01
-9.18614864e-01 6.71649277e-01 9.58079994e-01 4.25583236e-02
4.18200552e-01 1.36809096e-01 -1.14264853e-01 -2.20867425e-01
-8.75030100e-01 -3.87479156e-01 2.29087800e-01 4.40351218e-01
4.04455245e-01 -3.46030458e-03 -1.30372375e-01 6.77647173e-01
4.46139514e-01 2.05236614e-01 2.61205435e-01 1.15496993e+00
-5.55497885e-01 -9.63783622e-01 -1.96806744e-01 6.53032899e-01
-6.11129522e-01 -1.69484075e-02 -3.95981818e-01 4.37715352e-01
3.12203258e-01 9.13167715e-01 1.32703826e-01 -2.56579667e-01
8.45959876e-03 3.80935341e-01 6.15622461e-01 -9.42859411e-01
-1.22419149e-01 1.18885949e-01 5.92620783e-02 -8.85001779e-01
-9.10662293e-01 -8.24518502e-01 -1.31352460e+00 2.32671410e-01
-4.59058881e-01 8.24417919e-02 4.49317008e-01 1.20333278e+00
1.76471427e-01 5.36897421e-01 4.79029119e-01 -9.17856097e-01
-4.43977565e-01 -9.76945579e-01 -5.76336920e-01 5.67269325e-01
2.19824612e-01 -8.62800658e-01 -5.36787510e-01 6.73488319e-01] | [8.642057418823242, 0.7422336339950562] |
351973fc-39d6-4aab-960b-c28a61ecd821 | interpretable-summaries-of-black-box-incident | 2108.03013 | null | https://arxiv.org/abs/2108.03013v1 | https://arxiv.org/pdf/2108.03013v1.pdf | Interpretable Summaries of Black Box Incident Triaging with Subgroup Discovery | The need of predictive maintenance comes with an increasing number of incidents reported by monitoring systems and equipment/software users. In the front line, on-call engineers (OCEs) have to quickly assess the degree of severity of an incident and decide which service to contact for corrective actions. To automate these decisions, several predictive models have been proposed, but the most efficient models are opaque (say, black box), strongly limiting their adoption. In this paper, we propose an efficient black box model based on 170K incidents reported to our company over the last 7 years and emphasize on the need of automating triage when incidents are massively reported on thousands of servers running our product, an ERP. Recent developments in eXplainable Artificial Intelligence (XAI) help in providing global explanations to the model, but also, and most importantly, with local explanations for each model prediction/outcome. Sadly, providing a human with an explanation for each outcome is not conceivable when dealing with an important number of daily predictions. To address this problem, we propose an original data-mining method rooted in Subgroup Discovery, a pattern mining technique with the natural ability to group objects that share similar explanations of their black box predictions and provide a description for each group. We evaluate this approach and present our preliminary results which give us good hope towards an effective OCE's adoption. We believe that this approach provides a new way to address the problem of model agnostic outcome explanation. | ['Mehdi Kaytoue', 'Céline Robardet', 'Marc Plantevit', 'Anes Bendimerad', 'Youcef Remil'] | 2021-08-06 | null | null | null | null | ['subgroup-discovery'] | ['methodology'] | [ 2.13953018e-01 4.56138939e-01 -6.63439482e-02 -6.37025595e-01
-5.91417924e-02 -1.62552238e-01 1.01608515e-01 5.90353966e-01
1.73269287e-01 7.41141438e-01 9.63319372e-03 -6.02687299e-01
-1.03200936e+00 -7.92133927e-01 -2.05996230e-01 -3.73440236e-01
-2.09617272e-01 1.02423847e+00 3.41140516e-02 -3.45783770e-01
6.06088638e-01 7.61728227e-01 -1.76832163e+00 8.80800128e-01
8.42875123e-01 1.10814071e+00 1.75829068e-01 4.96751249e-01
-1.01329975e-01 1.07424092e+00 -7.01026261e-01 -2.70589828e-01
1.66139916e-01 -2.74165154e-01 -1.01853704e+00 3.48436236e-01
-2.60269642e-01 -8.55144113e-02 3.70068550e-01 4.24509317e-01
-7.14446511e-03 -6.81681931e-02 5.49957812e-01 -1.69505286e+00
-1.70395538e-01 5.14639854e-01 -2.24453077e-01 2.81044990e-01
5.28189719e-01 -6.25759661e-02 9.14902389e-01 -6.17696345e-01
5.06621480e-01 8.27956021e-01 6.65218353e-01 4.17822808e-01
-1.21200991e+00 -4.13239479e-01 2.50593305e-01 7.90376306e-01
-1.26561320e+00 -1.71270058e-01 5.55772424e-01 -4.11138624e-01
1.42311406e+00 9.76735771e-01 5.34862876e-01 4.80245471e-01
3.24198842e-01 1.17893569e-01 8.86069119e-01 -6.83045268e-01
2.00787127e-01 6.75525069e-01 5.17577410e-01 6.12792671e-01
4.74107772e-01 -1.94775417e-01 -4.85538542e-01 -3.77840817e-01
2.49558926e-01 4.15368497e-01 -1.76063597e-01 -1.01545505e-01
-7.94991791e-01 7.28227675e-01 -4.63446714e-02 5.10301769e-01
-7.04983950e-01 -4.11686212e-01 1.66429847e-01 5.33411920e-01
4.57963973e-01 7.05135405e-01 -9.40109372e-01 -1.19604105e-02
-6.32737279e-01 1.92975312e-01 1.04366875e+00 9.09058034e-01
8.62638891e-01 -2.58536249e-01 2.76261508e-01 3.53767812e-01
8.20991695e-02 -2.82639742e-01 2.61944383e-01 -6.58067286e-01
2.47434482e-01 1.35109341e+00 2.37090170e-01 -1.34608459e+00
-9.05146539e-01 -2.62072563e-01 -8.22332621e-01 2.36985326e-01
1.76932290e-01 2.52855748e-01 -3.41106117e-01 1.02299809e+00
2.20356584e-01 -1.78768620e-01 -2.01130778e-01 6.14095211e-01
1.04867898e-01 4.54769552e-01 -2.70267315e-02 -8.41176331e-01
1.30731750e+00 -6.80443525e-01 -8.39569211e-01 -5.10168560e-02
8.42029631e-01 -7.57283449e-01 7.48772562e-01 1.04419065e+00
-8.33114624e-01 -5.28959930e-01 -7.33586788e-01 5.54410517e-01
-3.54653865e-01 -2.18524002e-02 8.46701682e-01 4.32757229e-01
-6.70023799e-01 8.61180663e-01 -5.39173782e-01 -7.30447531e-01
-1.32107705e-01 7.66843379e-01 -4.47471946e-01 -7.94368014e-02
-7.16342151e-01 1.11087513e+00 3.70888442e-01 -8.25011209e-02
9.49831866e-03 -5.56344569e-01 -2.32079595e-01 3.53854001e-01
5.96357405e-01 -7.18034387e-01 9.77856040e-01 -6.13189816e-01
-5.98842323e-01 4.15959358e-01 -3.08019131e-01 -5.32064676e-01
7.05540180e-02 2.53766924e-02 -1.01197124e+00 -1.02153361e-01
1.19955875e-01 -2.02724010e-01 4.66891766e-01 -1.23438942e+00
-9.79847729e-01 -5.55417538e-01 -8.38667229e-02 -4.85921055e-01
-3.10165733e-01 4.06785995e-01 1.34360254e-01 -2.19140843e-01
3.49369079e-01 -6.95805788e-01 -5.64224303e-01 -6.14997625e-01
-4.24210072e-01 -3.78189415e-01 7.26118684e-01 -6.33652568e-01
1.91006303e+00 -1.82439244e+00 -1.57613039e-01 5.25893092e-01
2.19365180e-01 -1.38758287e-01 4.58182871e-01 8.68559003e-01
-6.10808492e-01 2.66290843e-01 -6.71717376e-02 -2.53727287e-01
-6.77883849e-02 4.40200597e-01 -4.49797124e-01 1.57570448e-02
3.51247728e-01 1.68170705e-01 -4.44333702e-01 -4.46297377e-01
2.95321316e-01 -8.75504389e-02 -6.33241415e-01 3.06866020e-01
-6.83645383e-02 4.43055600e-01 -3.37926835e-01 5.84738433e-01
3.24046195e-01 -4.18231159e-01 4.75176185e-01 -1.39801335e-02
-3.31481785e-01 2.29931429e-01 -1.28748453e+00 7.74285138e-01
-3.02484006e-01 1.14589497e-01 -2.74096549e-01 -1.27425504e+00
1.24415958e+00 6.45785987e-01 6.97802365e-01 -3.51409346e-01
-8.56150780e-03 2.41065174e-01 1.40881240e-02 -6.73196375e-01
4.51554805e-01 -3.17741662e-01 -4.48527671e-02 7.44510055e-01
-3.02239031e-01 2.91738272e-01 2.91219592e-01 4.96985428e-02
1.41875005e+00 -4.15211976e-01 7.41800189e-01 -8.56277943e-02
4.81983602e-01 4.55937803e-01 7.05054581e-01 5.38018882e-01
1.42504394e-01 6.77662253e-01 6.35311365e-01 -1.30356503e+00
-8.85490179e-01 -3.23513150e-01 4.70527224e-02 7.01040387e-01
-1.85739487e-01 -8.30884874e-01 -4.62968856e-01 -7.85713196e-01
-1.13310046e-01 1.14874601e+00 -4.68897879e-01 -7.68484473e-02
-5.24887621e-01 -7.29166746e-01 -3.79454762e-01 3.50740552e-01
-1.72564775e-01 -1.09639096e+00 -6.86762154e-01 6.82438731e-01
-3.53058249e-01 -9.34380710e-01 2.71922380e-01 5.01861691e-01
-1.03402758e+00 -1.34004211e+00 3.90900731e-01 -1.98155850e-01
8.22160065e-01 2.04191148e-01 1.25643647e+00 6.27289951e-01
-5.45812428e-01 2.75471359e-01 -6.17630541e-01 -7.08562016e-01
-6.61727786e-01 -1.39504328e-01 4.69430059e-01 1.53803572e-01
7.42067635e-01 -7.16809630e-01 -2.96243072e-01 8.15243363e-01
-6.95437908e-01 1.52954469e-02 4.72726166e-01 3.33039999e-01
4.90150779e-01 6.00025594e-01 7.96001673e-01 -1.09249580e+00
5.48248589e-01 -7.54507601e-01 -3.28806937e-01 5.44744909e-01
-1.30043578e+00 -2.70275120e-02 7.09912777e-01 -2.77434792e-02
-8.87832403e-01 -4.54341620e-03 2.44137198e-02 -5.61388358e-02
-5.80948591e-01 5.68983734e-01 3.04558240e-02 2.76885986e-01
7.43591130e-01 -1.73260316e-01 -8.83991718e-02 -7.75462925e-01
-1.86366558e-01 8.73737931e-01 3.17441076e-01 -7.18476549e-02
6.17747247e-01 2.85291553e-01 2.70674955e-02 -3.22853029e-01
-6.40902579e-01 -8.36553693e-01 -6.66329503e-01 -3.26149017e-01
6.24570131e-01 -1.27225772e-01 -1.00740349e+00 -4.22722042e-01
-1.47189665e+00 3.55452538e-01 -3.18615645e-01 3.45155716e-01
-7.58132219e-01 1.84222370e-01 -7.21847564e-02 -1.09110582e+00
-2.11883336e-01 -7.09295869e-01 4.57305700e-01 -1.48753121e-01
-9.70058501e-01 -6.18290603e-01 -1.58012211e-01 6.36559188e-01
3.93001139e-01 1.79279119e-01 1.45638478e+00 -1.08576083e+00
-6.15628660e-01 -7.33864427e-01 5.19824971e-04 1.98531687e-01
2.87531197e-01 9.53578763e-03 -7.12278068e-01 1.74673527e-01
2.31971353e-01 3.97158891e-01 1.98145360e-01 8.72263238e-02
1.44287527e+00 -5.43308914e-01 -4.77847844e-01 -6.32004663e-02
1.19851589e+00 4.50102508e-01 5.63699782e-01 5.19600868e-01
7.52944946e-02 1.17642212e+00 1.19564092e+00 8.33972394e-01
2.34052420e-01 9.58802223e-01 6.98332608e-01 -3.66922617e-02
3.16937804e-01 2.06279933e-01 -8.79497230e-02 7.97935843e-01
-6.59069657e-01 -8.26567113e-02 -9.43378210e-01 2.87531257e-01
-2.16329193e+00 -1.16146684e+00 -6.77547634e-01 2.14343596e+00
2.71102965e-01 3.24407935e-01 1.83755934e-01 7.96628356e-01
5.64503491e-01 -6.98542118e-01 3.05853807e-03 -9.59925830e-01
2.71199107e-01 -4.06478345e-03 1.54759914e-01 2.43563324e-01
-5.97052813e-01 2.41821706e-01 6.09534597e+00 3.18866998e-01
-6.41218066e-01 -5.48647344e-03 5.83767474e-01 2.32334323e-02
-1.47908747e-01 1.82241574e-01 -7.69673467e-01 2.77151644e-01
1.25722075e+00 -1.81792602e-01 2.72626221e-01 1.28507042e+00
5.75748205e-01 -1.47770643e-01 -1.24482095e+00 6.08246386e-01
-3.63010317e-02 -1.50654364e+00 -3.02800667e-02 2.33542800e-01
3.76529217e-01 -6.00560129e-01 -4.03856248e-01 1.33217201e-01
-1.65990099e-01 -8.90172660e-01 3.72260988e-01 7.80207574e-01
1.81241468e-01 -8.85018945e-01 9.88577843e-01 6.67157233e-01
-8.41052592e-01 -7.12440252e-01 -3.52582902e-01 -6.44853771e-01
3.50102127e-01 6.26948059e-01 -1.38835025e+00 8.14063787e-01
9.05500114e-01 1.53986841e-01 -2.84806460e-01 1.00046265e+00
1.45206107e-02 5.16182601e-01 -1.11385606e-01 1.54087409e-01
-2.32648715e-01 -1.61579549e-01 2.57908314e-01 9.42198217e-01
5.55157959e-01 3.84480506e-01 -2.21769121e-02 6.38069451e-01
7.22318590e-01 1.46503627e-01 -6.62347376e-01 3.35189939e-01
2.51295418e-01 1.18776333e+00 -7.59748757e-01 -2.95664966e-01
-5.58686733e-01 6.51873469e-01 3.25325541e-02 -9.41795483e-02
-5.59603393e-01 -2.08760157e-01 6.42257690e-01 7.94132054e-01
-8.81250575e-02 -2.17349362e-03 -7.28715062e-01 -6.23232722e-01
4.20444682e-02 -1.03185403e+00 5.63287854e-01 -9.13039505e-01
-1.21604323e+00 9.05061066e-01 1.44109994e-01 -1.39604616e+00
-5.38035274e-01 -5.23683190e-01 -7.04420090e-01 7.41620481e-01
-1.04716623e+00 -8.07981670e-01 -2.64842719e-01 5.80614924e-01
5.03267288e-01 -2.37464726e-01 1.19281232e+00 2.77155668e-01
-3.85855615e-01 -5.45910448e-02 -5.09853482e-01 -6.42336428e-01
4.91139978e-01 -1.13490403e+00 4.51910384e-02 4.72924232e-01
1.53209522e-01 7.69757152e-01 1.08937693e+00 -6.43915057e-01
-8.84118676e-01 -8.09882879e-01 1.75318551e+00 -7.96504915e-01
5.70316195e-01 -3.68233062e-02 -1.04313385e+00 7.09593356e-01
-3.65052558e-02 -4.18944269e-01 1.02224231e+00 5.80111086e-01
3.00889224e-01 -3.58007312e-01 -1.11614680e+00 2.80257493e-01
8.15665960e-01 -1.76907554e-02 -7.37483740e-01 7.07060814e-01
7.01808393e-01 3.27064931e-01 -8.96989465e-01 4.61957842e-01
2.18887985e-01 -1.46661329e+00 5.40624797e-01 -9.92131293e-01
2.14385912e-01 -3.43069375e-01 -4.31874134e-02 -9.17601466e-01
-4.59423751e-01 -5.85166454e-01 -5.98110184e-02 1.01780999e+00
6.16265416e-01 -6.89629674e-01 7.61998355e-01 1.08983743e+00
-3.35875511e-01 -1.03373110e+00 -7.40158379e-01 -6.50408566e-01
-9.12536860e-01 -9.92520869e-01 9.94896472e-01 9.43044782e-01
3.98451030e-01 2.16356784e-01 -5.76995552e-01 5.08111417e-01
3.34707379e-01 3.78687024e-01 7.62558639e-01 -1.76962113e+00
-4.32464242e-01 -5.81392832e-02 -6.43187225e-01 -1.54007480e-01
-4.43895876e-01 -5.17539442e-01 -5.05746007e-01 -1.61731815e+00
7.72085786e-02 -5.25995731e-01 -3.42642367e-01 6.80282831e-01
2.35674277e-01 -3.67484801e-02 -3.85487243e-03 3.91206086e-01
-4.64651436e-01 -1.72759354e-01 5.12373626e-01 2.90841818e-01
-2.14369655e-01 5.71584105e-01 -8.68937790e-01 9.37928319e-01
8.83626461e-01 -8.08328331e-01 -2.13089362e-01 5.73814027e-02
5.26272655e-01 2.25086689e-01 4.29999679e-01 -1.02377522e+00
3.91876101e-01 -3.91524643e-01 1.79131195e-01 -7.17607081e-01
1.37378871e-01 -1.40061963e+00 6.68657064e-01 5.37209332e-01
1.94730889e-02 3.48519683e-01 -1.03350863e-01 3.94430250e-01
-3.77711713e-01 -4.94644374e-01 2.20542908e-01 -3.28667536e-02
-6.79197371e-01 2.72881165e-02 -5.06747127e-01 -9.22537506e-01
1.27440679e+00 -4.79071110e-01 -2.59344518e-01 -3.98419797e-01
-1.26834464e+00 -3.88788208e-02 1.79461628e-01 4.00764108e-01
5.33742189e-01 -8.55621338e-01 -3.32536250e-01 3.62175882e-01
3.36308777e-01 -3.96692932e-01 3.13400537e-01 1.14769864e+00
-3.54435861e-01 6.53612375e-01 -1.46326438e-01 -3.00362796e-01
-1.50818491e+00 1.01645601e+00 -2.10552052e-01 -4.88044739e-01
-4.73208755e-01 4.50216532e-01 -7.99216926e-02 -2.47482538e-01
5.33937756e-03 -3.37344974e-01 -4.88174111e-01 2.77727917e-02
7.79634595e-01 6.25547051e-01 4.95889962e-01 -2.24201590e-01
-3.60616654e-01 3.18457305e-01 -8.96283388e-02 5.13295233e-01
1.76143324e+00 -1.81851372e-01 -4.62456703e-01 4.38103676e-01
5.65231085e-01 -7.23135844e-02 -5.11388898e-01 1.74801067e-01
6.22718096e-01 -5.07295609e-01 -4.24778461e-01 -1.01460254e+00
-6.72374964e-01 6.01483047e-01 3.01888257e-01 1.18843997e+00
1.48021913e+00 2.16742605e-01 2.72053629e-01 4.40265298e-01
7.22746849e-01 -1.03506196e+00 -1.84090748e-01 3.27424370e-02
1.10142541e+00 -1.10916436e+00 9.33247954e-02 -7.63327897e-01
-7.49012053e-01 1.31472111e+00 4.23004568e-01 3.55075955e-01
5.81552267e-01 2.96084344e-01 -1.07372709e-01 -5.19633293e-01
-1.24763513e+00 6.77494109e-02 1.17463768e-01 6.09570920e-01
2.42938474e-01 2.02287048e-01 -5.17727137e-01 1.12234354e+00
-1.23778485e-01 6.73382878e-02 5.66574514e-01 8.46820474e-01
-7.33460784e-01 -1.37934482e+00 -6.66169286e-01 7.34701574e-01
-3.80878687e-01 7.75465816e-02 -5.15172303e-01 9.81538773e-01
4.37446356e-01 1.38594794e+00 -2.52498332e-02 -7.44214416e-01
8.11116576e-01 3.23795944e-01 -2.46010050e-02 -7.36559629e-01
-6.72082067e-01 -2.88060606e-01 3.64722788e-01 -7.18191803e-01
-9.33184847e-02 -6.37236834e-01 -1.23124194e+00 -5.58494568e-01
-4.43559200e-01 5.25731444e-01 8.56596768e-01 1.05778897e+00
3.76935869e-01 5.31290650e-01 9.49354112e-01 -4.34813738e-01
-4.61149544e-01 -8.78173172e-01 -8.56710434e-01 3.97435814e-01
-1.92396134e-01 -6.37200236e-01 -4.69418138e-01 1.85954407e-01] | [8.463934898376465, 5.869870185852051] |
11da22ec-a946-41f0-8b63-927adba81448 | wdr-face-the-first-database-for-studying-face | 2101.03826 | null | https://arxiv.org/abs/2101.03826v1 | https://arxiv.org/pdf/2101.03826v1.pdf | WDR FACE: The First Database for Studying Face Detection in Wide Dynamic Range | Currently, face detection approaches focus on facial information by varying specific parameters including pose, occlusion, lighting, background, race, and gender. These studies only utilized the information obtained from low dynamic range images, however, face detection in wide dynamic range (WDR) scenes has received little attention. To our knowledge, there is no publicly available WDR database for face detection research. To facilitate and support future face detection research in the WDR field, we propose the first WDR database for face detection, called WDR FACE, which contains a total of 398 16-bit megapixel grayscale wide dynamic range images collected from 29 subjects. These WDR images (WDRIs) were taken in eight specific WDR scenes. The dynamic range of 90% images surpasses 60,000:1, and that of 70% images exceeds 65,000:1. Furthermore, we show the effect of different face detection procedures on the WDRIs in our database. This is done with 25 different tone mapping operators and five different face detectors. We provide preliminary experimental results of face detection on this unique WDR database. | ['Orly Yadid-Pecht', 'Svetlana Yanushkevich', 'Kenneth Kam Fai Lai', 'Mengchen Lin', 'Jie Yang', 'Ziyi Liu'] | 2021-01-11 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 1.62443310e-01 -7.60589063e-01 -3.33018675e-02 -4.90577400e-01
-3.82166713e-01 -3.55062127e-01 2.22633064e-01 -8.37279856e-01
-4.48711336e-01 5.09480178e-01 -1.02842197e-01 -9.63164866e-02
1.47666216e-01 -6.72424495e-01 -1.59680814e-01 -6.16512656e-01
-1.60005361e-01 -8.07102025e-02 1.99140698e-01 -4.23459142e-01
1.24011308e-01 1.17438614e+00 -1.87217140e+00 3.05181354e-01
1.47727966e-01 8.10895801e-01 3.34479809e-02 7.69480944e-01
4.91257489e-01 3.28321546e-01 -7.75093734e-01 -3.55948746e-01
6.44295633e-01 -3.89671385e-01 -1.86678499e-01 2.22835869e-01
8.52333188e-01 -8.38638484e-01 -6.75207794e-01 1.17078936e+00
9.57590461e-01 9.73759405e-03 4.05443847e-01 -1.11127889e+00
-9.87708688e-01 3.06163132e-01 -1.28017867e+00 8.03464532e-01
6.45995557e-01 2.61809349e-01 2.86162853e-01 -1.26400840e+00
6.14096284e-01 1.74804306e+00 4.38619137e-01 9.17073190e-01
-1.16108322e+00 -1.40690756e+00 -2.89062411e-01 1.01204842e-01
-1.92933035e+00 -9.59575474e-01 7.43701756e-01 -2.41668537e-01
6.70068383e-01 2.42321342e-01 4.97035593e-01 9.27832901e-01
-6.80698082e-02 -1.62025720e-01 1.61176622e+00 -5.21322966e-01
-3.13086748e-01 2.42515638e-01 -4.56343666e-02 1.00200617e+00
4.57057774e-01 3.47740978e-01 -6.08862281e-01 6.95390403e-02
1.12866032e+00 -4.26491171e-01 -2.20754966e-01 3.86334300e-01
-5.10444641e-01 7.07750678e-01 -2.88183205e-02 3.02485049e-01
2.76176602e-01 -2.03920901e-01 7.94407502e-02 5.52472234e-01
4.35577601e-01 6.86715264e-03 -3.54744904e-02 2.49161407e-01
-6.49099171e-01 -1.27854794e-01 6.08244896e-01 8.37892890e-01
5.58655441e-01 2.48017415e-01 -6.86547011e-02 1.38183105e+00
4.86923397e-01 1.01676953e+00 3.10580999e-01 -8.48897934e-01
1.14603050e-01 1.67665318e-01 -1.08119428e-01 -1.32212806e+00
-3.08239400e-01 1.87162638e-01 -4.22585666e-01 4.38362509e-01
5.20258427e-01 -4.29285616e-01 -9.59247410e-01 1.51760495e+00
3.35724682e-01 -1.07974224e-02 -5.79652116e-02 1.10691988e+00
1.30282331e+00 4.66401905e-01 1.13146439e-01 -4.92013961e-01
1.68614507e+00 -8.36571753e-02 -7.84695506e-01 -1.57310352e-01
-2.34739244e-01 -1.22066116e+00 9.35473144e-01 3.79923433e-01
-8.47443044e-01 -7.15651333e-01 -1.13041508e+00 5.31431846e-02
-2.18066201e-01 5.99026561e-01 2.73002446e-01 1.60701048e+00
-1.26933312e+00 2.03664415e-02 -3.03386897e-01 -6.76289201e-01
3.40179026e-01 5.04822969e-01 -5.48223257e-01 -2.67058611e-01
-1.17656374e+00 9.76885915e-01 -1.08067982e-01 1.20673738e-01
-8.05822372e-01 -3.72266978e-01 -7.14858115e-01 -3.94739181e-01
2.46080965e-01 7.82365501e-02 6.82724357e-01 -7.64752984e-01
-1.51561952e+00 1.60514402e+00 -5.65235466e-02 3.81207885e-03
3.50644529e-01 1.49039432e-01 -1.07053697e+00 5.31101882e-01
-3.76876503e-01 5.30034423e-01 1.09947824e+00 -1.05303979e+00
-1.47036880e-01 -6.13295913e-01 -2.43671224e-01 5.14747202e-02
-5.08815110e-01 1.26190615e+00 -6.29203260e-01 -5.48877120e-01
-1.71322525e-01 -8.55490208e-01 3.85914266e-01 3.71103048e-01
-1.13335937e-01 6.46030232e-02 1.23702765e+00 -7.69450128e-01
1.05446398e+00 -2.26697636e+00 -7.49381125e-01 2.17490464e-01
-2.04533469e-02 5.52207887e-01 -4.23990965e-01 -1.93579912e-01
-2.96523720e-01 -5.88703714e-02 4.12759453e-01 1.54802963e-01
-2.16188401e-01 -2.08784983e-01 -2.92500556e-01 9.04121637e-01
4.60877046e-02 3.49343568e-01 -2.74275184e-01 -8.23246896e-01
3.18210512e-01 8.23396564e-01 -3.43025029e-01 1.84741914e-02
5.05600393e-01 7.16968114e-03 -2.46142030e-01 1.20244491e+00
1.24817169e+00 3.73846382e-01 -7.88511988e-03 -4.58948493e-01
-1.30894303e-01 -5.62114418e-01 -1.25405896e+00 7.61516392e-01
-1.92018986e-01 1.06461132e+00 2.73663163e-01 -3.43985081e-01
1.54559338e+00 1.31560534e-01 4.24435765e-01 -7.84402668e-01
3.12234849e-01 9.21039730e-02 1.89040631e-01 -5.54934442e-01
5.96684158e-01 -2.35170275e-01 2.95310140e-01 2.94078797e-01
-1.78582624e-01 9.07541066e-02 3.74738038e-01 -1.04839407e-01
6.32807493e-01 -3.22168380e-01 2.39692956e-01 -2.26799712e-01
6.29524171e-01 -4.43401873e-01 4.93656486e-01 4.45136368e-01
-6.94875240e-01 6.15719259e-01 2.64163554e-01 -1.35483518e-01
-7.15131044e-01 -1.27691770e+00 -7.70164371e-01 1.28030324e+00
3.43325853e-01 -8.12993571e-02 -6.02417946e-01 -2.60078683e-02
2.31565051e-02 2.17136052e-02 -5.48032701e-01 -9.25420821e-02
-7.36152589e-01 -1.27792370e+00 9.38921630e-01 2.88306624e-01
9.37687635e-01 -9.85318959e-01 -5.83023250e-01 -4.17772561e-01
-5.47721377e-03 -1.23845279e+00 -7.97372758e-01 -5.62091470e-01
-4.44188327e-01 -1.24627602e+00 -6.89717650e-01 -8.87293458e-01
6.48163497e-01 5.00006795e-01 9.55970466e-01 1.09140232e-01
-1.30050647e+00 4.39408630e-01 -2.31289819e-01 -2.35197648e-01
-3.87352496e-01 -7.18783021e-01 2.86021769e-01 9.98481885e-02
7.82301545e-01 -1.32510746e-02 -5.96343577e-01 9.07884061e-01
-3.15439939e-01 -5.63141227e-01 4.35951799e-01 4.71445620e-01
4.24629271e-01 2.94264048e-01 6.01380765e-01 -5.64424813e-01
4.63329732e-01 4.61487174e-02 -8.59619737e-01 2.32058495e-01
-1.73673540e-01 -6.01674557e-01 1.09946236e-01 -8.02854419e-01
-1.56645525e+00 -9.55798328e-02 -8.18464756e-02 -4.28282797e-01
-1.70149907e-01 -5.44226289e-01 -1.65792480e-01 -6.44038141e-01
8.79663467e-01 5.77614568e-02 1.67880446e-01 -1.47850245e-01
6.07387051e-02 1.06501830e+00 7.34011054e-01 -4.13450718e-01
9.51646924e-01 4.93076950e-01 -2.47037187e-01 -1.28422666e+00
-1.87783480e-01 -2.97986776e-01 -1.76078677e-01 -5.45958817e-01
8.08906496e-01 -1.11846364e+00 -8.41685772e-01 8.03825796e-01
-5.62722027e-01 -4.98456359e-02 1.89755931e-01 6.36289418e-01
-5.15033305e-02 1.29914895e-01 -8.56177568e-01 -9.84156966e-01
-3.10331821e-01 -1.06631207e+00 9.90926504e-01 4.04844522e-01
2.14942634e-01 -4.44166869e-01 -2.71747738e-01 3.23457062e-01
5.93984246e-01 9.01527554e-02 4.44140375e-01 -7.98118650e-04
-3.40498596e-01 -7.28473663e-02 -6.24066591e-01 2.03529149e-01
4.96392429e-01 5.47302186e-01 -1.10315883e+00 -5.14934361e-01
-6.15650415e-02 -2.32549369e-01 8.20509255e-01 4.24233884e-01
1.14647090e+00 2.03733087e-01 -2.20414296e-01 4.90460604e-01
1.43522429e+00 6.55762672e-01 9.49522793e-01 -3.31664644e-02
3.65700215e-01 6.16984189e-01 6.82481229e-01 5.26675165e-01
-2.19882146e-01 8.51078987e-01 1.02186039e-01 -2.83317119e-01
-7.42765844e-01 1.30462289e-01 6.29083157e-01 -4.12057247e-03
-2.52420902e-01 6.67115301e-02 -6.21710896e-01 -1.12422453e-02
-6.26395106e-01 -1.23190498e+00 1.69362828e-01 2.01204801e+00
8.50261688e-01 -3.37746501e-01 2.58099437e-01 -3.70644480e-02
1.34974480e+00 1.45791933e-01 -3.80011559e-01 -3.35772753e-01
-2.38127232e-01 3.37977916e-01 4.45679724e-01 1.75336912e-01
-1.19622028e+00 9.42446828e-01 7.35662746e+00 8.25562358e-01
-1.41873741e+00 -5.94596192e-02 8.51803184e-01 -3.61131281e-01
2.74429649e-01 -7.57930100e-01 -1.34338570e+00 3.60572726e-01
5.46894312e-01 -8.30947310e-02 4.92322445e-01 8.87520015e-01
2.19126686e-01 -3.21597695e-01 -6.62333727e-01 1.59843683e+00
6.47385955e-01 -5.53932309e-01 -3.88011098e-01 1.65543124e-01
5.33516288e-01 -5.12974083e-01 7.50726700e-01 1.16099253e-01
1.69930696e-01 -1.21650183e+00 1.95618242e-01 1.47468328e-01
1.46672559e+00 -9.29568052e-01 3.53514194e-01 -4.78286982e-01
-1.25263071e+00 -2.00715423e-01 -6.79114878e-01 4.18967634e-01
-3.87081355e-01 3.03049326e-01 -7.34174132e-01 -2.19379887e-01
8.92770290e-01 3.58806461e-01 -8.74210477e-01 6.31505668e-01
2.09492669e-01 3.47050667e-01 -3.01301062e-01 6.30534291e-02
-6.76869512e-01 -1.46390244e-01 4.89826262e-01 1.23775089e+00
3.39295924e-01 4.02563512e-01 -5.71323521e-02 6.24568760e-01
-4.08567905e-01 1.65981859e-01 -6.85238063e-01 1.37345120e-01
6.93076551e-01 1.47104383e+00 -9.08951461e-01 7.11129457e-02
-5.54304540e-01 5.57506800e-01 -3.96142840e-01 2.52836823e-01
-9.11624134e-01 -5.25406122e-01 8.61349642e-01 1.73376903e-01
1.22312687e-01 -2.26300266e-02 9.23215412e-03 -8.73740077e-01
-2.38944501e-01 -1.14632034e+00 5.51437795e-01 -6.77637875e-01
-1.20706475e+00 7.46237576e-01 2.92704612e-01 -8.57694507e-01
2.91338824e-02 -8.76344204e-01 -3.28294814e-01 9.26878691e-01
-1.33573139e+00 -7.54968941e-01 -5.70175231e-01 9.36788321e-01
4.15056348e-01 -6.72258317e-01 5.12740195e-01 6.40555978e-01
-8.83776844e-01 1.08520222e+00 -3.56196016e-01 4.70125407e-01
1.10043406e+00 -6.78539753e-01 6.50783405e-02 1.02339959e+00
-1.94455877e-01 8.26610982e-01 5.65403044e-01 -5.87454975e-01
-1.59326530e+00 -1.04667187e+00 2.91518658e-01 -1.12092778e-01
2.60800332e-01 -4.33290333e-01 -6.80167198e-01 4.80668992e-01
-1.35551482e-01 4.17110026e-01 6.49519563e-01 -1.63191929e-01
-5.40512741e-01 -5.68898380e-01 -1.81590211e+00 6.83596730e-01
1.11554503e+00 -5.83246529e-01 -1.31072491e-01 1.21878140e-01
1.31204531e-01 -4.10739005e-01 -8.76389027e-01 4.13729340e-01
8.05960953e-01 -9.60472643e-01 1.32386661e+00 1.33031607e-01
-1.25380009e-01 -3.58255655e-01 -4.00807142e-01 -7.41219997e-01
2.19159722e-02 -3.93503547e-01 3.10745627e-01 1.42574441e+00
-7.80795962e-02 -8.13670039e-01 5.77470839e-01 5.89267433e-01
4.06769514e-01 -2.96703070e-01 -8.18203688e-01 -7.54025519e-01
-2.40332901e-01 -4.02095281e-02 4.47247475e-01 6.52817190e-01
-5.00770152e-01 -1.87562019e-01 -5.30180275e-01 1.69397339e-01
7.64841020e-01 3.30403537e-01 5.16720176e-01 -8.55799019e-01
-6.54149801e-02 -1.76363274e-01 -6.07111752e-01 -3.48055273e-01
9.13006663e-02 -4.63134438e-01 -1.45192176e-01 -8.45624566e-01
4.32289660e-01 -2.78272182e-01 1.92276984e-01 4.44142401e-01
7.32909963e-02 1.23283362e+00 1.98231399e-01 -1.44260392e-01
-3.00263446e-02 5.48830107e-02 1.22400093e+00 -8.78600031e-02
-1.22410275e-01 -3.66162747e-01 -7.77004719e-01 7.78017700e-01
7.09341407e-01 -8.02997649e-02 -3.04046422e-01 -4.99688201e-02
-5.70045292e-01 -6.35814667e-02 2.39756599e-01 -9.71690536e-01
5.71014769e-02 -2.83839852e-01 1.16747940e+00 -4.07056272e-01
6.44789636e-01 -5.70115268e-01 3.01742405e-01 5.94041467e-01
-6.70354813e-02 -3.93485129e-02 2.68843234e-01 6.63140565e-02
4.82502915e-02 6.63232952e-02 1.72223115e+00 -6.55183941e-02
-1.08325136e+00 4.16779399e-01 -3.64694268e-01 -4.35656756e-02
1.34889627e+00 -5.34815252e-01 -5.66569865e-01 8.78998451e-03
-5.24510086e-01 -3.30527455e-01 3.55752617e-01 5.93077481e-01
9.61218178e-01 -1.27309406e+00 -9.63035762e-01 7.65655220e-01
-2.79182941e-02 -9.79277551e-01 2.41632789e-01 4.35749710e-01
-5.35230756e-01 -5.20876050e-02 -7.53600001e-01 -3.88049275e-01
-2.28340030e+00 3.96308213e-01 5.60274720e-01 5.83244681e-01
-5.93293548e-01 7.79623330e-01 8.87685642e-02 2.36520216e-01
2.47064587e-02 3.72843295e-01 -5.63020229e-01 1.86164990e-01
1.07284367e+00 7.34257281e-01 6.88927099e-02 -1.14810956e+00
-5.55129468e-01 1.07867146e+00 -1.90092474e-01 -5.01595698e-02
7.68451750e-01 -3.11106622e-01 -1.16668157e-01 -1.35481760e-01
1.31467986e+00 2.38274321e-01 -9.57353652e-01 -4.79964800e-02
-6.21856689e-01 -1.27258384e+00 -1.39581054e-01 -6.11216724e-01
-1.51712894e+00 5.76960862e-01 1.39185488e+00 -2.11703405e-01
1.57696450e+00 6.82129487e-02 2.77176082e-01 2.55144596e-01
5.85299194e-01 -1.10259199e+00 5.59985340e-01 1.54825851e-01
1.02132702e+00 -1.19007027e+00 2.72091548e-03 -8.12494755e-01
-3.25666428e-01 1.22164989e+00 9.80811238e-01 5.53194918e-02
5.96425176e-01 5.03839254e-01 3.71481299e-01 3.88741568e-02
-4.65915591e-01 -3.74824613e-01 -3.28672193e-02 9.71450031e-01
4.21108782e-01 -1.83174998e-01 -7.25644454e-02 -3.63921076e-02
-3.22732389e-01 -3.54849845e-02 5.51921844e-01 4.74129856e-01
-7.16537178e-01 -7.38773584e-01 -9.95896816e-01 5.55315197e-01
-8.82192612e-01 2.74615377e-01 -2.29952008e-01 9.86318946e-01
5.99841364e-02 1.27006972e+00 1.49160191e-01 -2.58317679e-01
3.05103958e-01 -3.23276371e-01 8.24013948e-01 -4.29628581e-01
-1.16356656e-01 4.74400930e-02 2.77555212e-02 -4.42102432e-01
-4.27508533e-01 -5.63841164e-01 -9.31137145e-01 -7.20030665e-01
-2.06428379e-01 -3.94233495e-01 5.61792314e-01 2.16159359e-01
-1.62055477e-01 1.84644714e-01 8.88538539e-01 -5.57094038e-01
-3.41741264e-01 -9.68956888e-01 -1.16653240e+00 2.31143758e-01
2.94086576e-01 -9.72486734e-01 -3.10416132e-01 2.05088496e-01] | [13.23963737487793, 0.7978472709655762] |
e063e00d-04d4-4d9d-808c-fb875caba203 | face-recognition-using-synthetic-face-data | 2305.10079 | null | https://arxiv.org/abs/2305.10079v1 | https://arxiv.org/pdf/2305.10079v1.pdf | Face Recognition Using Synthetic Face Data | In the field of deep learning applied to face recognition, securing large-scale, high-quality datasets is vital for attaining precise and reliable results. However, amassing significant volumes of high-quality real data faces hurdles such as time limitations, financial burdens, and privacy issues. Furthermore, prevalent datasets are often impaired by racial biases and annotation inaccuracies. In this paper, we underscore the promising application of synthetic data, generated through rendering digital faces via our computer graphics pipeline, in achieving competitive results with the state-of-the-art on synthetic data across multiple benchmark datasets. By finetuning the model,we obtain results that rival those achieved when training with hundreds of thousands of real images (98.7% on LFW [1]). We further investigate the contribution of adding intra-class variance factors (e.g., makeup, accessories, haircuts) on model performance. Finally, we reveal the sensitivity of pre-trained face recognition models to alternating specific parts of the face by leveraging the granular control capability in our platform. | ['Orly Zvitia', 'Max Kogan', 'Vladimir Loginov', 'Alexey Gruzdev', 'Omer Granoviter'] | 2023-05-17 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [ 2.67443001e-01 -8.77811201e-03 2.35869467e-01 -8.46733689e-01
-7.87515581e-01 -4.85571474e-01 7.06345975e-01 -3.99768084e-01
-4.09898579e-01 7.22833693e-01 3.83455269e-02 -1.45536378e-01
3.08087338e-02 -7.79898167e-01 -9.61677432e-01 -3.82230192e-01
1.18938342e-01 2.46853799e-01 -5.34683526e-01 1.66523992e-03
1.56443521e-01 8.84480655e-01 -1.84603941e+00 3.87164563e-01
6.88891709e-01 1.16871488e+00 -6.08126640e-01 8.27474371e-02
2.64358614e-02 3.68770152e-01 -9.59145606e-01 -8.41468215e-01
8.16906631e-01 -4.14222740e-02 -1.77467346e-01 2.62924582e-01
1.44201088e+00 -8.37926507e-01 -2.00850099e-01 8.03076863e-01
6.63374484e-01 -2.54299909e-01 6.27994895e-01 -1.49105287e+00
-7.00351119e-01 1.01301074e-01 -7.79423058e-01 -1.15430400e-01
8.80135223e-02 5.13405502e-01 5.40800393e-01 -1.12621748e+00
6.03246450e-01 1.42790616e+00 7.88202167e-01 8.06063414e-01
-1.60721397e+00 -1.38334286e+00 -6.53897971e-02 -2.50168234e-01
-1.44209909e+00 -1.33972347e+00 5.28372467e-01 -6.48982704e-01
5.00205040e-01 2.21518278e-01 2.92347878e-01 1.67644906e+00
-1.99618470e-02 3.18423510e-01 1.25604880e+00 -1.47888079e-01
1.91265553e-01 2.30054632e-01 -1.47273928e-01 5.51545024e-01
5.96978128e-01 1.77461222e-01 -7.41511881e-01 -5.67405760e-01
6.59435987e-01 -3.65545340e-02 -1.87695891e-01 -3.50435048e-01
-7.11794376e-01 6.72884762e-01 1.26179725e-01 -2.13085741e-01
-1.72816172e-01 6.18776446e-03 2.85400659e-01 1.29426420e-01
5.72495699e-01 5.65920532e-01 -4.38674301e-01 1.22075908e-01
-1.25486302e+00 3.09195399e-01 6.08890295e-01 8.78204942e-01
6.08829260e-01 3.40069473e-01 -2.22653195e-01 7.45065153e-01
1.69369996e-01 5.39471626e-01 1.49725810e-01 -1.12573361e+00
3.24042410e-01 5.19821882e-01 2.72940606e-01 -1.07894647e+00
-1.62684530e-01 -5.18381178e-01 -7.29212224e-01 4.56132501e-01
8.15192759e-01 -1.87382713e-01 -9.61087883e-01 1.79032922e+00
3.43240619e-01 1.54584765e-01 -3.18591416e-01 6.96124315e-01
6.09817684e-01 -4.35948744e-02 2.64432073e-01 4.25349083e-03
1.31172192e+00 -4.37477171e-01 -4.11163867e-01 -2.14598984e-01
1.75923213e-01 -8.36570680e-01 1.28349745e+00 5.53825259e-01
-7.49108791e-01 -5.00633657e-01 -9.42653120e-01 1.69089839e-01
-1.94961101e-01 3.58841300e-01 5.70067763e-01 1.25003433e+00
-1.10125399e+00 5.76280415e-01 -5.41163146e-01 -2.93901801e-01
1.32949865e+00 5.13653159e-01 -7.31219828e-01 -4.88970637e-01
-7.03853250e-01 4.50664520e-01 -2.99019545e-01 1.24468274e-01
-9.50832069e-01 -1.25890136e+00 -5.93062520e-01 -4.49633226e-02
3.99910331e-01 -3.66328537e-01 9.59881544e-01 -8.22069168e-01
-1.04999995e+00 9.85797822e-01 -3.27761397e-02 -8.30854103e-02
9.50370908e-01 -3.04605424e-01 -5.08068800e-01 -1.65282235e-01
-1.00774825e-01 6.22837663e-01 1.20510697e+00 -1.29015577e+00
-1.23654149e-01 -7.64417171e-01 -3.30396265e-01 -4.43197966e-01
-6.82339668e-01 2.46966466e-01 -7.19763190e-02 -5.05600691e-01
-4.08488333e-01 -7.49131620e-01 4.47576679e-02 4.73566383e-01
-3.03436637e-01 1.42564848e-01 9.02820528e-01 -8.89435351e-01
7.15743184e-01 -2.41727805e+00 -4.20291513e-01 1.82172447e-01
3.97442043e-01 5.11693060e-01 -4.02805775e-01 -4.11910079e-02
-9.12571233e-03 4.69367504e-01 -3.85412537e-02 -5.13190031e-01
1.14583656e-01 -1.59379810e-01 -2.56477386e-01 6.04544818e-01
5.85361242e-01 8.12508821e-01 -3.96222502e-01 -1.49676472e-01
-6.43195435e-02 7.71669745e-01 -6.46271110e-01 2.52592444e-01
-9.88477990e-02 3.29222679e-01 -7.46444017e-02 9.93109047e-01
1.00658500e+00 -6.44632950e-02 1.93716615e-01 -3.22905064e-01
2.18706533e-01 -7.99567252e-02 -9.28844213e-01 1.30622518e+00
-3.52500945e-01 4.71187413e-01 3.59015703e-01 -3.33780229e-01
1.02684355e+00 -3.26149128e-02 2.34750032e-01 -7.70504653e-01
1.60855711e-01 1.92142457e-01 -3.87691264e-03 -3.23814124e-01
4.31662440e-01 -1.54286712e-01 2.44495884e-01 5.20951271e-01
4.68484350e-02 2.98211761e-02 -2.05701739e-01 1.58620365e-02
9.26189125e-01 -4.81352173e-02 -3.12518090e-01 -4.78048235e-01
-1.57478645e-01 -3.94402266e-01 4.92887080e-01 5.61840415e-01
-3.85081440e-01 7.63979554e-01 7.33188510e-01 -5.14768660e-01
-1.31240761e+00 -9.27075386e-01 -2.92397857e-01 8.66944075e-01
-7.84894109e-01 -2.12737933e-01 -9.54002798e-01 -6.48817182e-01
5.10556757e-01 5.38724959e-01 -8.65750551e-01 -2.45924234e-01
-3.26240599e-01 -1.02062476e+00 9.24245059e-01 4.37987655e-01
3.61745745e-01 -5.17856419e-01 -6.13869965e-01 -8.22801590e-02
3.18438292e-01 -1.27864194e+00 -3.27393264e-01 -5.62756360e-01
-5.73031127e-01 -1.18258595e+00 -6.05924428e-01 -1.75817147e-01
7.13577569e-01 3.23488563e-02 1.30487692e+00 1.12249583e-01
-8.16104949e-01 1.83782160e-01 1.24470882e-01 -5.30066252e-01
-6.48299754e-02 4.45176475e-03 3.25696737e-01 3.60457331e-01
5.34117758e-01 -4.22380120e-01 -7.46824086e-01 4.60172534e-01
-7.29391336e-01 -2.63575405e-01 4.52497214e-01 7.87970126e-01
1.92573622e-01 -1.80057600e-01 6.98772967e-01 -1.18515754e+00
5.74916899e-01 -3.40596229e-01 -8.22483361e-01 1.72162697e-01
-6.11395836e-01 -2.10542306e-01 4.29324120e-01 -4.84797150e-01
-1.11689985e+00 -1.83169052e-01 1.82680756e-01 -5.60286880e-01
-3.20844680e-01 -1.67871565e-01 -4.11918819e-01 -3.51359069e-01
8.92339647e-01 -2.77588874e-01 3.07733953e-01 -5.19856453e-01
2.24253818e-01 7.17581391e-01 3.92083973e-01 -8.40620995e-01
6.99241817e-01 5.76497912e-01 -4.08189967e-02 -7.04296887e-01
-7.32465982e-01 2.47779474e-01 -4.29589123e-01 -2.28674009e-01
4.20593709e-01 -1.05924165e+00 -7.60721207e-01 7.72139549e-01
-7.90699184e-01 -3.66275758e-01 -1.83151215e-01 1.19071506e-01
-2.32207356e-03 -1.17977530e-01 -5.33989429e-01 -8.96566033e-01
-2.25456357e-01 -1.07319295e+00 1.23019111e+00 1.98359519e-01
-2.51631975e-01 -3.24504346e-01 -4.30177510e-01 7.74110377e-01
7.58178234e-01 6.67379677e-01 8.29780102e-01 -4.49455529e-01
-5.19455373e-01 -2.56361753e-01 -5.30039728e-01 4.35263425e-01
2.28675887e-01 4.12151366e-01 -1.48216975e+00 -4.76139545e-01
-7.18941316e-02 -8.32920671e-01 5.21681845e-01 -1.75090618e-02
1.42678440e+00 -3.99320662e-01 -5.53971380e-02 6.97966158e-01
1.27361631e+00 -1.09005928e-01 6.17006302e-01 -5.33452667e-02
7.85484195e-01 1.02403402e+00 2.95846820e-01 7.00247884e-01
5.40660098e-02 5.86070716e-01 2.50832409e-01 -1.13587357e-01
-2.11208001e-01 -2.88584262e-01 -9.27705131e-03 -5.85441850e-02
1.92533568e-01 2.93115713e-02 -1.03530192e+00 3.77661794e-01
-1.25998390e+00 -7.16798782e-01 8.48267823e-02 2.34042120e+00
7.28394628e-01 7.42373895e-03 1.32381365e-01 -9.36110243e-02
5.78136802e-01 1.42267928e-01 -8.99320841e-01 -5.13645634e-02
-2.50765651e-01 3.85218382e-01 4.70388949e-01 1.59581125e-01
-8.22038114e-01 8.18437159e-01 6.68258858e+00 7.29407728e-01
-1.26620305e+00 -2.92337947e-02 1.42648673e+00 -7.56942570e-01
-3.63823891e-01 -7.07528532e-01 -7.86104500e-01 3.73838544e-01
1.02179289e+00 4.17067818e-02 7.44758427e-01 8.03835332e-01
1.46894991e-01 8.28868970e-02 -1.27928686e+00 1.03405869e+00
3.67198765e-01 -1.17257237e+00 -1.16362768e-04 4.71478015e-01
8.09228837e-01 3.52326259e-02 5.63187480e-01 9.47650895e-02
2.88576901e-01 -1.70206702e+00 7.69705653e-01 3.66289407e-01
1.43187630e+00 -7.44122326e-01 4.57648605e-01 -1.45576447e-01
-4.86262679e-01 3.26714329e-02 -3.91653001e-01 3.60867567e-02
-3.47993761e-01 8.23968291e-01 -8.34487081e-01 1.04744889e-01
6.17626429e-01 2.34912947e-01 -8.37527454e-01 6.37359321e-01
8.01721141e-02 5.35985708e-01 -3.61186951e-01 3.65052074e-01
-2.45071039e-01 -8.38362426e-02 -2.34809086e-01 7.72137702e-01
3.20551634e-01 -1.72330718e-02 -3.40119570e-01 1.08511913e+00
-6.40352070e-01 -6.01619668e-02 -7.54945457e-01 -2.87768066e-01
5.46050966e-01 1.33340466e+00 -3.04305345e-01 -7.11696744e-02
-3.90176922e-01 5.56044936e-01 4.66366738e-01 4.76011574e-01
-4.99605268e-01 1.33997202e-01 1.17131281e+00 4.95494187e-01
8.60456377e-02 -1.20962389e-01 -5.60808957e-01 -1.05222261e+00
3.56365710e-01 -1.42921841e+00 6.50922582e-02 -5.21723270e-01
-1.50333989e+00 6.15120709e-01 -3.80832016e-01 -6.48235142e-01
-1.59942508e-02 -8.66511524e-01 -2.05396399e-01 1.04410148e+00
-1.28227496e+00 -1.15800953e+00 -5.11166930e-01 5.26137471e-01
1.10999666e-01 -4.41215813e-01 9.12552655e-01 5.94309270e-01
-7.84227252e-01 1.18942618e+00 7.01544285e-02 2.35118419e-01
1.02806342e+00 -5.54200411e-01 8.64799559e-01 6.38639927e-01
3.50373425e-02 8.01129162e-01 4.65523124e-01 -4.57568526e-01
-1.55361772e+00 -1.15022838e+00 4.54561949e-01 -8.08742642e-01
4.51003551e-01 -9.90496993e-01 -9.91310060e-01 5.65838516e-01
-3.04766923e-01 3.07572275e-01 9.90741432e-01 2.89823890e-01
-1.02545249e+00 -3.99950594e-01 -1.63621914e+00 5.80189884e-01
1.21614254e+00 -6.92633271e-01 7.33253583e-02 1.93229720e-01
3.98659348e-01 -2.03835949e-01 -8.37072968e-01 4.29602683e-01
1.09450758e+00 -1.18955815e+00 8.77208650e-01 -9.25590694e-01
6.47598445e-01 1.24314621e-01 -3.04917514e-01 -1.14005613e+00
-5.49073778e-02 -5.38221359e-01 5.24820797e-02 1.47761178e+00
4.39926088e-01 -6.26663208e-01 1.16500890e+00 1.41178954e+00
3.58754039e-01 -8.29317510e-01 -8.44645739e-01 -7.00440824e-01
1.03696167e-01 -3.12884897e-01 1.14725173e+00 1.08764768e+00
-6.15736008e-01 -6.28774688e-02 -4.82746303e-01 -4.91542891e-02
8.91190827e-01 -2.50349313e-01 1.05754018e+00 -1.26673734e+00
-1.53666899e-01 -2.95595378e-01 -2.46563137e-01 -1.16989449e-01
4.50008154e-01 -3.42773855e-01 -3.13271075e-01 -8.92894685e-01
2.97619730e-01 -7.06218183e-01 -1.19209945e-01 5.11752486e-01
-1.16496287e-01 6.96543694e-01 3.89979035e-01 5.73195843e-03
-8.18030015e-02 4.51268971e-01 1.05427134e+00 -1.22276910e-01
3.33725452e-01 -3.66040468e-01 -1.13566744e+00 5.41383028e-01
7.42734909e-01 -3.81748915e-01 -1.39655873e-01 -7.44191945e-01
-1.05224460e-01 -3.26706350e-01 4.43461031e-01 -9.63753223e-01
-1.04446240e-01 -1.69145271e-01 7.88942099e-01 1.32837966e-01
5.47214329e-01 -8.22668135e-01 3.34728509e-01 2.69123644e-01
-2.89768398e-01 -2.04625651e-02 4.81297672e-01 3.82893264e-01
2.13754535e-01 3.15280735e-01 8.14301252e-01 -2.90305982e-03
-1.92794695e-01 4.70377803e-01 6.45118952e-02 2.01658577e-01
9.08331573e-01 -3.76643566e-03 -7.12226570e-01 -1.64926410e-01
-5.11430949e-02 -4.46483158e-02 8.35230529e-01 6.04219675e-01
3.35077852e-01 -1.26935458e+00 -9.01036382e-01 8.08768868e-01
2.84574807e-01 -1.12837680e-01 2.48406380e-01 2.37761512e-01
-3.66689861e-01 4.74798195e-02 -5.28503597e-01 -3.07710290e-01
-1.32658398e+00 2.46342033e-01 2.16843218e-01 2.74580389e-01
-1.76738709e-01 1.10364866e+00 1.19884506e-01 -4.35855627e-01
3.24566603e-01 9.02359113e-02 4.62206483e-01 1.64273754e-01
8.45626295e-01 4.44999903e-01 3.13448250e-01 -5.53355455e-01
-4.23735410e-01 2.68977225e-01 -2.63697922e-01 -1.16942503e-01
1.26685309e+00 4.45163220e-01 -2.95899175e-02 4.56913970e-02
1.10853338e+00 3.81518528e-02 -1.81288147e+00 5.00183478e-02
-2.70407796e-01 -1.06854272e+00 -1.54639229e-01 -9.69885588e-01
-1.40218699e+00 9.31191385e-01 7.56856859e-01 -2.71036685e-01
8.30766499e-01 -3.38293701e-01 5.88622987e-01 1.12815514e-01
7.01121032e-01 -8.72017682e-01 1.15542449e-01 -1.04848020e-01
9.41501498e-01 -1.47468328e+00 1.39240980e-01 -3.73565465e-01
-3.69366586e-01 6.92596555e-01 8.86705160e-01 1.53012648e-01
5.02717793e-01 5.21226823e-01 3.78523380e-01 -1.31398752e-01
-7.45779753e-01 6.30778432e-01 4.25805189e-02 6.26979530e-01
3.94637346e-01 2.69380182e-01 1.27225399e-01 3.75045180e-01
-3.67436320e-01 1.75573170e-01 4.61265087e-01 6.86435342e-01
6.99449778e-02 -1.03849876e+00 -5.73113918e-01 1.03512812e+00
-6.17740393e-01 -5.97088747e-02 -5.65141678e-01 7.25723267e-01
1.96946055e-01 9.19275284e-01 3.09678555e-01 -4.97764945e-01
4.45362061e-01 1.59204781e-01 5.32178164e-01 -5.70000172e-01
-6.69057190e-01 -2.36513942e-01 1.78640217e-01 -7.66613126e-01
-1.95602253e-02 -8.30365360e-01 -3.07565421e-01 -7.79569030e-01
3.32783870e-02 -3.48452538e-01 9.21863139e-01 6.19514287e-01
9.20159400e-01 2.22913623e-01 4.61035639e-01 -7.80635536e-01
-9.23836529e-01 -9.16089177e-01 -6.42948687e-01 6.77292228e-01
1.82471588e-01 -6.68219566e-01 -4.47314918e-01 -1.39846712e-01] | [12.902861595153809, 0.8113130331039429] |
1dbaedc7-6c52-4e69-a19c-511190a6a3e0 | word-embeddings-for-banking-industry | 2306.01807 | null | https://arxiv.org/abs/2306.01807v1 | https://arxiv.org/pdf/2306.01807v1.pdf | Word Embeddings for Banking Industry | Applications of Natural Language Processing (NLP) are plentiful, from sentiment analysis to text classification. Practitioners rely on static word embeddings (e.g. Word2Vec or GloVe) or static word representation from contextual models (e.g. BERT or ELMo) to perform many of these NLP tasks. These widely available word embeddings are built from large amount of text, so they are likely to have captured most of the vocabulary in different context. However, how well would they capture domain-specific semantics and word relatedness? This paper explores this idea by creating a bank-specific word embeddings and evaluates them against other sources of word embeddings such as GloVe and BERT. Not surprising that embeddings built from bank-specific corpora does a better job of capturing the bank-specific semantics and word relatedness. This finding suggests that bank-specific word embeddings could be a good stand-alone source or a complement to other widely available embeddings when performing NLP tasks specific to the banking industry. | ['Avnish Patel'] | 2023-06-02 | null | null | null | null | ['word-embeddings', 'sentiment-analysis'] | ['methodology', 'natural-language-processing'] | [-5.82421601e-01 -7.22870976e-02 -3.93419832e-01 -5.67952216e-01
-4.30916339e-01 -8.09412420e-01 8.26373816e-01 7.98631072e-01
-8.37555408e-01 4.82295394e-01 9.47592199e-01 -5.80382884e-01
-3.24400561e-03 -9.71416712e-01 -3.04158814e-02 -5.25051236e-01
1.82368606e-01 5.31123519e-01 6.87041581e-02 -8.36975574e-01
5.50907135e-01 4.83247817e-01 -1.13994157e+00 3.47763970e-02
3.57033640e-01 5.93493283e-01 1.48186952e-01 5.69858313e-01
-8.76757503e-01 3.10103416e-01 -5.31035006e-01 -4.99042422e-01
1.70794681e-01 8.35810006e-02 -7.89938927e-01 -3.36092353e-01
-3.67012247e-02 5.25797904e-02 -1.34346291e-01 7.62199223e-01
6.18366897e-01 3.02816093e-01 5.79405427e-01 -9.13281024e-01
-1.28784120e+00 5.61705351e-01 -3.07456374e-01 5.40705681e-01
4.09870356e-01 1.27093717e-01 1.57621694e+00 -1.06046605e+00
5.36708415e-01 1.25581074e+00 7.01004505e-01 2.92660475e-01
-9.77010071e-01 -4.46694791e-01 1.57003887e-02 -4.08646800e-02
-1.12107158e+00 -4.72630002e-02 5.81799448e-01 -5.40001988e-01
1.53534782e+00 3.38477716e-02 6.13858223e-01 1.39184630e+00
3.72698992e-01 3.96717519e-01 9.42752182e-01 -6.61838770e-01
1.13057092e-01 7.47965813e-01 8.07067394e-01 1.23575635e-01
7.73727119e-01 -5.01457155e-02 -5.13419211e-01 -4.11912233e-01
5.92331827e-01 3.05480182e-01 -2.02877417e-01 -2.76988387e-01
-1.20235252e+00 1.33686614e+00 2.56560475e-01 8.49330366e-01
-3.13804388e-01 8.57108757e-02 6.47801995e-01 3.47057670e-01
5.64697742e-01 9.97612655e-01 -8.00133646e-01 -3.84198159e-01
-7.51490057e-01 3.84409547e-01 9.60120499e-01 7.79117107e-01
9.28159654e-01 3.63672748e-02 -5.53179495e-02 9.62180674e-01
4.85941380e-01 2.27994546e-01 1.23382175e+00 -1.83732048e-01
3.40611011e-01 7.44924188e-01 1.40305921e-01 -1.36147690e+00
-3.97562355e-01 5.22347689e-02 -6.72845766e-02 -4.02635038e-02
3.72156441e-01 -3.03088307e-01 -8.73494089e-01 1.41514289e+00
3.78347635e-02 -1.26185387e-01 2.12137863e-01 8.27052355e-01
7.60381877e-01 7.21322119e-01 2.30345011e-01 1.73004419e-01
1.76685083e+00 -7.98698068e-01 -7.04697251e-01 -7.29217052e-01
8.46950829e-01 -8.30713809e-01 1.41800642e+00 9.36082751e-02
-4.94369507e-01 -4.90102887e-01 -1.04894030e+00 -2.00496510e-01
-1.35731530e+00 -6.73696041e-01 7.12439716e-01 1.11955893e+00
-8.96931052e-01 4.60756838e-01 -5.25053680e-01 -9.23463404e-01
4.93436232e-02 -1.76498871e-02 -6.15706980e-01 -2.10789904e-01
-1.51480806e+00 1.45176518e+00 3.99690568e-01 -3.89328688e-01
-3.62211317e-01 -5.77682316e-01 -1.22578502e+00 4.74295877e-02
-2.55418848e-02 -4.06912416e-01 7.95230567e-01 -8.22224081e-01
-8.39291990e-01 9.45720375e-01 -1.70643598e-01 -5.26120007e-01
-2.51668036e-01 -4.56514657e-01 -5.54349124e-01 -1.86864674e-01
2.50003517e-01 2.90166855e-01 5.68960607e-01 -9.90868688e-01
-2.29548395e-01 -3.85726333e-01 9.87843722e-02 -7.11668134e-02
-9.86359119e-01 4.06501263e-01 1.03678375e-01 -8.57499778e-01
-3.14743519e-01 -6.82869554e-01 -2.61728823e-01 -2.88315326e-01
3.66938770e-01 -4.77161050e-01 6.53996408e-01 -7.41637886e-01
1.53767943e+00 -2.01242328e+00 -3.87036830e-01 -2.88562104e-02
8.33466575e-02 5.25643587e-01 -3.48630041e-01 1.16201067e+00
-3.35050732e-01 6.92317903e-01 4.68818210e-02 -7.81564862e-02
1.61753908e-01 6.59713984e-01 -3.47667485e-01 3.65741372e-01
4.67190653e-01 1.10435903e+00 -1.05888522e+00 -3.63280714e-01
3.55444312e-01 5.47693789e-01 -4.00284559e-01 -9.98793468e-02
3.18336561e-02 -3.31949651e-01 -3.89293075e-01 4.45837975e-01
3.00389111e-01 1.37922481e-01 2.07267702e-01 6.04651049e-02
4.27698642e-02 4.44409579e-01 -9.47043777e-01 1.44417119e+00
-7.71106243e-01 8.85282099e-01 -3.73361707e-01 -1.20715630e+00
1.12764180e+00 4.64198261e-01 2.48491511e-01 -3.83888811e-01
4.27492768e-01 3.24133970e-02 4.51138243e-02 -6.28209710e-01
1.00316572e+00 -8.39618444e-01 -4.07194108e-01 6.95574105e-01
3.83958608e-01 -9.50937495e-02 -3.91129311e-03 1.74625769e-01
1.24860013e+00 -3.40119630e-01 8.49196017e-01 -5.89781165e-01
3.28962415e-01 1.91918209e-01 3.90687108e-01 4.04838681e-01
-5.26212990e-01 8.57633531e-01 3.36978734e-01 -5.56047618e-01
-9.76067781e-01 -9.12068129e-01 -1.91379681e-01 1.11653435e+00
-2.01566339e-01 -8.49501252e-01 -1.34426206e-01 -5.77910900e-01
3.33602518e-01 1.01975060e+00 -7.58083820e-01 -2.10329995e-01
-4.42557216e-01 -6.10471547e-01 4.16581273e-01 8.36813688e-01
-3.79420221e-01 -1.08770657e+00 -5.96045494e-01 6.19014084e-01
2.91395664e-01 -9.82305706e-01 -2.28437364e-01 4.92091805e-01
-7.34406531e-01 -8.78692329e-01 -5.38087547e-01 -7.64720023e-01
1.13800369e-01 4.44400311e-01 1.37898767e+00 -1.68341458e-01
-2.44953245e-01 5.09455383e-01 -1.01006567e+00 -7.21134186e-01
-1.04613751e-01 -8.73991922e-02 3.25138181e-01 -2.05469251e-01
1.33668935e+00 -4.66809392e-01 -2.67316788e-01 1.01261646e-01
-1.17693329e+00 -8.71081591e-01 2.80390769e-01 9.79758501e-01
2.60531697e-02 -1.26158953e-01 7.24564612e-01 -1.04039979e+00
1.32928073e+00 -7.82529771e-01 1.49103105e-01 1.13011494e-01
-7.09954381e-01 3.94721963e-02 5.35218298e-01 -5.51476300e-01
-6.40677392e-01 -5.71482539e-01 -1.59641668e-01 -2.42583886e-01
-2.18556717e-01 8.42589855e-01 6.61524907e-02 4.01377827e-01
9.46633339e-01 -1.29538178e-01 -1.13451354e-01 -5.19722044e-01
6.61110044e-01 8.48192036e-01 -7.92403519e-02 -6.60600126e-01
8.28444898e-01 2.56350428e-01 -7.73397446e-01 -1.07164431e+00
-7.34329998e-01 -1.02605772e+00 -6.04271531e-01 2.84701824e-01
1.25775933e+00 -6.03699744e-01 1.50425017e-01 -1.99382305e-01
-1.16557324e+00 1.55930743e-02 -6.16163731e-01 5.80414832e-01
1.02114081e-02 3.29056531e-01 -3.40112329e-01 -7.48834729e-01
-2.05917850e-01 -7.30842650e-01 8.21483731e-01 1.34999558e-01
-8.18142235e-01 -1.58328748e+00 4.95815068e-01 1.57718748e-01
5.94744086e-01 2.10574821e-01 1.02135372e+00 -1.48912513e+00
3.99261385e-01 -6.96392298e-01 -1.77914515e-01 7.00841963e-01
6.22160554e-01 -7.50352964e-02 -9.29624021e-01 -2.21153960e-01
5.12139238e-02 -1.07323870e-01 6.68962419e-01 -4.04296145e-02
3.35882425e-01 -3.14525217e-01 -1.80961594e-01 1.47127375e-01
1.72870576e+00 -2.00701896e-02 4.12092954e-01 6.47460043e-01
4.40256834e-01 7.97950685e-01 5.72182000e-01 4.90278512e-01
2.06391692e-01 3.31517607e-01 1.12477988e-01 3.09786946e-01
2.44615912e-01 -2.25191846e-01 5.64389288e-01 1.02780604e+00
3.51724505e-01 -2.39923149e-01 -1.46365750e+00 1.09144902e+00
-1.49663877e+00 -7.07644105e-01 -6.75705224e-02 1.83068478e+00
6.34417653e-01 2.31057599e-01 -8.28883871e-02 2.01013342e-01
4.84968811e-01 6.12661779e-01 6.39907792e-02 -1.23420811e+00
-1.11497387e-01 7.43556976e-01 5.10883451e-01 3.22262198e-01
-6.97983921e-01 1.02151155e+00 6.15889406e+00 5.62221110e-01
-8.78184617e-01 3.60903829e-01 1.21216886e-01 -1.10273711e-01
-5.19689679e-01 1.80478469e-01 -7.84771979e-01 4.59740549e-01
1.32086968e+00 -4.59572405e-01 -1.67168885e-01 9.99707878e-01
9.51465592e-02 1.56294078e-01 -1.15839159e+00 9.24103081e-01
2.90606856e-01 -1.20814800e+00 2.40655020e-01 8.93129781e-02
5.72144330e-01 1.19885959e-01 -1.26583040e-01 5.34287810e-01
6.53375149e-01 -1.37833047e+00 2.13613942e-01 1.08856723e-01
4.63440388e-01 -6.12578154e-01 1.32909715e+00 1.64171621e-01
-9.15911913e-01 -3.97432372e-02 -8.42585802e-01 -3.57712954e-01
4.76055652e-01 5.88428617e-01 -7.58130133e-01 4.53279942e-01
5.86867094e-01 8.02780449e-01 -6.77068889e-01 4.85837370e-01
-1.97198480e-01 6.69236541e-01 -8.17269459e-02 -2.71076828e-01
7.53573596e-01 -1.84707090e-01 4.15827483e-01 1.69774795e+00
1.28184468e-01 1.56765226e-02 -1.24203250e-01 3.43118787e-01
1.79804921e-01 4.14646059e-01 -9.65741813e-01 -7.09128618e-01
4.68047142e-01 1.28885615e+00 -5.02247095e-01 -2.71438897e-01
-9.58802164e-01 6.83493137e-01 2.55447090e-01 1.96600437e-01
-5.93452036e-01 -6.31337643e-01 1.37881792e+00 1.84130311e-01
2.50840247e-01 -4.84428287e-01 -4.71672922e-01 -1.12005174e+00
-1.25900477e-01 -7.41445839e-01 3.49336207e-01 -7.92282403e-01
-1.84273827e+00 5.27903378e-01 -1.64001256e-01 -9.34907198e-01
-5.31328171e-02 -1.16906416e+00 -9.17289972e-01 9.84778881e-01
-1.48616076e+00 -9.50348079e-01 2.24009287e-02 3.01184595e-01
5.72732091e-01 -3.83630186e-01 1.03059411e+00 1.13979466e-02
-2.82977343e-01 4.13544536e-01 7.81238079e-02 3.60271543e-01
1.13355172e+00 -1.36436045e+00 3.59698325e-01 5.73404491e-01
5.35455048e-01 1.18383610e+00 8.32892597e-01 -3.45490366e-01
-1.26611173e+00 -7.59541750e-01 1.49288273e+00 -1.07573485e+00
1.27325547e+00 -3.58936608e-01 -1.08688760e+00 8.07877064e-01
5.35768926e-01 -1.10009983e-01 1.19464457e+00 3.71016234e-01
-7.38790989e-01 -1.32594202e-02 -1.07015920e+00 6.69524848e-01
5.70085049e-01 -7.74733126e-01 -1.50127840e+00 2.05865040e-01
1.03541863e+00 4.17725146e-01 -8.65231574e-01 -1.16205841e-01
4.42707509e-01 -6.69010460e-01 1.11990750e+00 -1.17067051e+00
7.14413166e-01 1.65561765e-01 -4.48913723e-01 -1.52210069e+00
-3.67833912e-01 -2.90910512e-01 3.27016503e-01 1.63576329e+00
5.29545128e-01 -9.89867508e-01 5.53358674e-01 8.68865430e-01
5.70354350e-02 -7.32451797e-01 -7.23229885e-01 -9.32769179e-01
7.84006894e-01 -7.92851210e-01 6.50566280e-01 1.49189544e+00
3.96359354e-01 6.34622574e-01 1.83177963e-02 -1.30609989e-01
-1.27792299e-01 -2.81815082e-01 6.64449215e-01 -1.14275169e+00
-1.88807230e-02 -5.21864414e-01 -8.54668438e-01 -5.31154633e-01
3.59146267e-01 -1.08028460e+00 -3.07167917e-01 -1.67121542e+00
-2.18184933e-01 -2.20577374e-01 -6.57899320e-01 3.22063833e-01
-2.36193225e-01 -2.32087225e-02 1.82295352e-01 -2.20889464e-01
8.92908052e-02 3.10783982e-01 6.63737118e-01 -4.15119529e-02
-4.64158691e-02 -6.61053479e-01 -1.25990450e+00 5.97538054e-01
1.07593060e+00 -5.78655541e-01 -2.85922408e-01 -5.23064196e-01
5.18225729e-01 -4.65658396e-01 7.53001273e-02 -5.02219379e-01
-2.79401662e-03 -2.58969724e-01 1.10618167e-01 9.06035304e-03
2.36404851e-01 -9.31420803e-01 -4.15747583e-01 6.48600189e-03
-1.63665503e-01 4.90054011e-01 2.27355048e-01 6.44777119e-01
-5.94312549e-01 -6.25959635e-01 5.15736580e-01 -4.43409443e-01
-9.22111869e-01 -4.30666357e-02 -5.40476978e-01 4.06569064e-01
8.86539221e-01 -6.20008230e-01 -1.76652715e-01 -2.72120684e-01
-5.41208327e-01 1.12839811e-01 5.03630996e-01 9.52919781e-01
5.52560210e-01 -1.31185985e+00 -6.06027663e-01 1.37056202e-01
4.71918970e-01 -3.19141716e-01 -3.46625715e-01 4.22989488e-01
-5.91160774e-01 6.55643284e-01 -1.46936297e-01 5.70980087e-02
-9.21775937e-01 8.19838464e-01 -1.53316587e-01 -4.06369269e-01
-4.55358624e-01 1.02792871e+00 8.82487595e-02 -5.87785959e-01
-2.61457801e-01 -3.47372949e-01 -4.17982519e-01 7.31580675e-01
5.49239039e-01 9.25635267e-03 -3.74051221e-02 -8.19023311e-01
-7.07086086e-01 5.72700977e-01 -1.09968549e-02 -2.79349625e-01
1.70072639e+00 2.81610247e-02 4.75484394e-02 7.89362192e-01
1.48325455e+00 1.28333271e-01 -3.20481747e-01 -1.94862798e-01
4.53948647e-01 -6.71337664e-01 4.75740209e-02 -4.23107684e-01
-8.04533243e-01 1.11737120e+00 1.77900746e-01 3.28514606e-01
4.40951437e-01 3.52969468e-02 9.11225438e-01 2.19159827e-01
3.99034530e-01 -1.24898887e+00 1.41650185e-01 5.96365154e-01
7.86037087e-01 -1.16714990e+00 -2.45835409e-02 1.99440911e-01
-9.47838247e-01 1.15282226e+00 3.51724356e-01 -5.56038320e-01
1.05882084e+00 1.17485218e-01 2.76993394e-01 -2.82046765e-01
-7.32124925e-01 -4.41714078e-01 -2.14947797e-02 7.87880182e-01
8.95628572e-01 3.56166810e-02 -7.28121340e-01 9.00769591e-01
-4.35880214e-01 -4.30364490e-01 6.21945977e-01 1.11188889e+00
-5.10066807e-01 -1.39197016e+00 -2.30025768e-01 5.92542708e-01
-5.23152888e-01 -3.96540016e-01 -5.67325473e-01 9.14268255e-01
-1.62625071e-02 1.13196778e+00 9.23068672e-02 -3.44801873e-01
3.64580214e-01 6.87552154e-01 -1.42689794e-01 -1.24947834e+00
-1.03466475e+00 -5.27478933e-01 1.74664840e-01 -3.47833961e-01
-2.88399547e-01 -4.80687320e-01 -9.46997702e-01 -3.24050069e-01
-3.57150942e-01 2.19720691e-01 5.07669806e-01 7.73014963e-01
1.54005006e-01 2.07581416e-01 1.63048014e-01 -5.36061943e-01
-5.09368360e-01 -1.18588912e+00 -7.48190522e-01 7.35462785e-01
2.44960845e-01 -6.66786253e-01 -4.22001392e-01 -1.87156186e-01] | [10.461445808410645, 8.719318389892578] |
72a5962e-9eb4-44e0-8c75-13238b89130e | asdot-any-shot-data-to-text-generation-with | 2210.04325 | null | https://arxiv.org/abs/2210.04325v3 | https://arxiv.org/pdf/2210.04325v3.pdf | ASDOT: Any-Shot Data-to-Text Generation with Pretrained Language Models | Data-to-text generation is challenging due to the great variety of the input data in terms of domains (e.g., finance vs sports) or schemata (e.g., diverse predicates). Recent end-to-end neural methods thus require substantial training examples to learn to disambiguate and describe the data. Yet, real-world data-to-text problems often suffer from various data-scarce issues: one may have access to only a handful of or no training examples, and/or have to rely on examples in a different domain or schema. To fill this gap, we propose Any-Shot Data-to-Text (ASDOT), a new approach flexibly applicable to diverse settings by making efficient use of any given (or no) examples. ASDOT consists of two steps, data disambiguation and sentence fusion, both of which are amenable to be solved with off-the-shelf pretrained language models (LMs) with optional finetuning. In the data disambiguation stage, we employ the prompted GPT-3 model to understand possibly ambiguous triples from the input data and convert each into a short sentence with reduced ambiguity. The sentence fusion stage then uses an LM like T5 to fuse all the resulting sentences into a coherent paragraph as the final description. We evaluate extensively on various datasets in different scenarios, including the zero-/few-/full-shot settings, and generalization to unseen predicates and out-of-domain data. Experimental results show that ASDOT consistently achieves significant improvement over baselines, e.g., a 30.81 BLEU gain on the DART dataset under the zero-shot setting. | ['Zhiting Hu', 'Eric P. Xing', 'Yucheng Zhou', 'Zhengzhong Liu', 'Jiannan Xiang'] | 2022-10-09 | null | null | null | null | ['data-to-text-generation'] | ['natural-language-processing'] | [ 5.31163216e-01 1.95165589e-01 -1.81257412e-01 -5.63623905e-01
-1.33051550e+00 -7.76945591e-01 6.96105242e-01 4.92009163e-01
-4.92520630e-01 1.04044044e+00 2.60787189e-01 -3.51417512e-01
-1.26459211e-01 -7.24859715e-01 -7.43835568e-01 -2.97405243e-01
4.44079280e-01 1.02879536e+00 8.40392411e-02 -7.05197871e-01
-4.95601296e-02 -1.26185313e-01 -1.58146954e+00 4.95771438e-01
1.10176075e+00 9.50402558e-01 4.66315717e-01 5.32726347e-01
-6.97524965e-01 3.31880033e-01 -8.72881949e-01 -6.09322846e-01
3.25068891e-01 -3.44589829e-01 -8.66546154e-01 1.02916226e-01
4.85673964e-01 7.77257513e-03 6.99053258e-02 8.85182798e-01
8.25699627e-01 2.94676036e-01 3.27868134e-01 -1.01812851e+00
-6.22288465e-01 1.03062665e+00 -4.14785743e-01 1.58615947e-01
5.99413097e-01 1.02187142e-01 1.15934050e+00 -1.16230881e+00
7.45149612e-01 1.47289526e+00 3.25138599e-01 6.82549059e-01
-1.40824091e+00 -4.91339296e-01 2.12114811e-01 -6.06653653e-02
-1.02068269e+00 -6.68501437e-01 4.62137550e-01 -2.06783324e-01
1.12102258e+00 3.59064370e-01 5.79504669e-02 1.40519738e+00
-3.87865826e-02 7.15488970e-01 8.61661196e-01 -6.04306579e-01
2.20541000e-01 9.40711796e-02 3.07606667e-01 2.21881568e-01
3.76933277e-01 -2.66968459e-01 -6.88389957e-01 -8.38121474e-02
3.31566781e-02 -1.93426624e-01 -1.52186513e-01 1.67965293e-01
-1.20130551e+00 6.59532845e-01 2.11616114e-01 1.88766375e-01
-2.23140553e-01 -3.35255504e-01 5.45029461e-01 4.06914800e-01
3.95428598e-01 6.96093440e-01 -7.09298670e-01 -1.46023691e-01
-9.78903830e-01 6.55330122e-01 9.87020016e-01 1.20200181e+00
8.35918784e-01 -1.74642012e-01 -5.95481157e-01 1.07663965e+00
-8.68779719e-02 4.29363996e-01 7.62678385e-01 -4.59093064e-01
1.42199504e+00 4.94405478e-01 1.23616032e-01 -5.76112390e-01
-2.93232322e-01 -2.07608804e-01 -8.21385860e-01 -2.19370708e-01
3.04782093e-01 -5.97297549e-01 -1.32853401e+00 1.87390459e+00
3.20878774e-01 -2.16812760e-01 5.35428286e-01 8.30346763e-01
1.19391429e+00 6.10101163e-01 1.01943329e-01 -1.80082902e-01
1.58893180e+00 -8.27649951e-01 -6.10632062e-01 -8.47384214e-01
6.97893798e-01 -8.58469486e-01 1.28070569e+00 1.05834253e-01
-1.21287167e+00 -5.64825714e-01 -9.58067954e-01 -4.80537891e-01
-6.29814565e-01 7.54471049e-02 7.89816529e-02 3.28898519e-01
-7.34990954e-01 6.11459851e-01 -6.15185678e-01 -6.00056767e-01
1.75744370e-01 3.08257163e-01 -3.20418328e-01 -1.85158253e-01
-1.58080482e+00 8.00359786e-01 1.10175312e+00 -2.73227364e-01
-3.29863042e-01 -8.19326818e-01 -1.05521488e+00 1.72924981e-01
9.22515333e-01 -1.17502308e+00 1.52336121e+00 -7.63054848e-01
-1.31316757e+00 6.57780409e-01 -3.21632773e-01 -6.78529382e-01
4.95636433e-01 -3.69788021e-01 -3.74091327e-01 -2.52977133e-01
5.50872326e-01 6.92646980e-01 5.63684583e-01 -9.93649483e-01
-7.75674760e-01 -2.12420076e-01 1.89669788e-01 4.77624416e-01
-1.14843160e-01 6.18853495e-02 -4.64081258e-01 -7.39119530e-01
-1.97850801e-02 -7.95962453e-01 -1.04109883e-01 -5.86983204e-01
-8.48662138e-01 -3.46221149e-01 6.38163686e-01 -5.68725884e-01
1.34691560e+00 -2.00995493e+00 -4.98109311e-02 -2.12628409e-01
-1.05237119e-01 4.04276252e-01 -2.58486807e-01 7.17770576e-01
-1.90144703e-01 2.79911876e-01 -3.10480535e-01 -6.75238967e-01
1.84463724e-01 3.60691637e-01 -5.00902653e-01 -4.47234333e-01
5.74464977e-01 7.94854462e-01 -9.52878714e-01 -5.91540992e-01
-3.52512337e-02 5.42510934e-02 -3.09957057e-01 1.34882271e-01
-5.82524478e-01 1.90145537e-01 -4.97222096e-01 3.80963176e-01
4.10573810e-01 -2.40866736e-01 7.33023360e-02 -5.52799478e-02
4.12601046e-02 7.48254836e-01 -1.33387864e+00 2.00243807e+00
-4.90290731e-01 3.00772130e-01 -2.69433498e-01 -7.89843976e-01
8.10774803e-01 3.07774752e-01 6.14887401e-02 -5.96192002e-01
1.69993311e-01 4.05897081e-01 -6.28258809e-02 -5.61732829e-01
9.42031205e-01 -3.30991387e-01 -5.45489073e-01 2.77593732e-01
3.47427368e-01 -1.53193966e-01 8.53788018e-01 4.06620532e-01
1.08682299e+00 -9.70776677e-02 3.87847185e-01 1.32203903e-02
1.76768064e-01 2.27869585e-01 7.82009542e-01 8.89943242e-01
3.53808731e-01 9.30390120e-01 5.08321285e-01 -1.47605091e-01
-9.89318669e-01 -8.33204627e-01 8.14044401e-02 1.05586731e+00
3.92036662e-02 -5.98550022e-01 -5.99789023e-01 -7.03570187e-01
-1.96599253e-02 1.16427696e+00 -4.00548577e-01 -5.44257946e-02
-5.12339532e-01 -7.06967890e-01 3.90139967e-01 4.66028661e-01
3.91971380e-01 -9.69899774e-01 -3.37212771e-01 5.02044201e-01
-4.12000984e-01 -1.31368148e+00 -4.74090338e-01 4.61351067e-01
-4.99099791e-01 -6.65199220e-01 -3.55459839e-01 -6.35689080e-01
5.02959967e-01 2.51160175e-01 1.34794486e+00 -3.46798211e-01
-9.68176220e-03 -8.52494016e-02 -5.14743745e-01 -5.68051577e-01
-5.87198555e-01 2.30111897e-01 3.05424966e-02 -1.75070569e-01
3.89141560e-01 -3.31456453e-01 -3.40122581e-01 2.97922287e-02
-1.05405033e+00 1.27244100e-01 6.37863278e-01 1.11736488e+00
6.48718655e-01 2.09644400e-02 7.01380312e-01 -1.14087987e+00
9.79216337e-01 -6.30537033e-01 -2.10988134e-01 5.20619631e-01
-4.85456079e-01 3.44692171e-01 7.67061412e-01 -4.34004515e-01
-1.10985661e+00 -2.74941660e-02 -4.27617729e-02 -3.81177455e-01
-2.15886727e-01 8.86683464e-01 -5.11645377e-01 7.05817819e-01
7.95962334e-01 6.39083087e-02 -2.36427009e-01 -4.78068650e-01
5.87957084e-01 7.51330078e-01 7.61970818e-01 -7.83072114e-01
7.01400459e-01 -7.29266331e-02 -4.32256430e-01 -5.39390802e-01
-1.20278382e+00 -3.15349221e-01 -5.17168224e-01 3.85202944e-01
7.87155807e-01 -9.11965132e-01 2.32665613e-02 1.49264023e-01
-1.25358248e+00 -5.68693392e-02 -4.52506989e-01 8.99565592e-02
-3.30724627e-01 2.21896738e-01 -3.03151190e-01 -4.95881557e-01
-7.25036263e-01 -1.06189251e+00 1.24670863e+00 4.19181615e-01
-5.29345930e-01 -8.02013636e-01 -1.57718197e-01 4.93224472e-01
1.34326905e-01 2.05811575e-01 9.53663170e-01 -1.17400420e+00
-1.88512072e-01 -1.04077406e-01 -5.54757565e-02 2.38695353e-01
3.13194841e-01 -1.16202049e-01 -9.01156723e-01 -2.63314664e-01
-1.99132010e-01 -5.27337074e-01 9.18578565e-01 5.81480702e-03
1.01161110e+00 -4.79742795e-01 -2.04970822e-01 2.78333277e-01
1.27502108e+00 1.15879690e-02 2.59983301e-01 2.49139234e-01
5.34383476e-01 6.33145630e-01 9.52657938e-01 4.07589525e-01
5.30045152e-01 6.83933973e-01 5.02723157e-02 3.25345956e-02
-1.98999316e-01 -4.78460222e-01 2.46236339e-01 5.57871342e-01
4.28680778e-01 -6.84576988e-01 -9.76571143e-01 5.87110877e-01
-1.89135647e+00 -9.51944351e-01 1.75638050e-01 2.18051481e+00
1.30587876e+00 4.38257635e-01 -3.48808281e-02 5.75168617e-02
8.07622254e-01 1.09757446e-01 -6.78484917e-01 -3.59534502e-01
-3.63710791e-01 2.04681262e-01 2.88918585e-01 3.10035497e-01
-1.08900833e+00 1.01775515e+00 5.02353668e+00 1.07976520e+00
-1.17869270e+00 -5.56549989e-02 6.09457076e-01 -2.80338526e-01
-3.52402538e-01 6.46039797e-03 -1.19524264e+00 7.14487374e-01
1.00981593e+00 -5.57043374e-01 2.58408040e-01 5.79610705e-01
2.04383284e-01 -9.93444398e-02 -1.26399028e+00 1.00536168e+00
1.09951481e-01 -1.24046934e+00 2.72876740e-01 -3.61901700e-01
6.38081551e-01 8.43492374e-02 -1.21083587e-01 6.09248817e-01
5.58770478e-01 -8.06734324e-01 9.12320018e-01 1.93232626e-01
9.74317133e-01 -5.34245253e-01 6.89744949e-01 6.99115038e-01
-8.75691891e-01 -5.10961711e-02 -2.86601901e-01 6.24153251e-03
3.43425572e-01 7.03598261e-01 -1.21330583e+00 9.50882018e-01
5.95196843e-01 4.45474416e-01 -4.03987050e-01 6.69441879e-01
-1.44758910e-01 2.46556640e-01 -5.24774909e-01 -6.67237937e-02
2.19518572e-01 1.87028646e-01 6.48409426e-01 1.22416651e+00
6.46509767e-01 9.47194621e-02 4.48168546e-01 7.96383560e-01
-4.11779672e-01 6.63157273e-03 -4.52972561e-01 -1.81277812e-01
7.80573308e-01 1.22882223e+00 -3.72627854e-01 -5.93117714e-01
-4.20010060e-01 8.35517347e-01 4.34272528e-01 3.73436213e-01
-4.39519823e-01 -6.42282248e-01 6.70720875e-01 8.16460401e-02
3.01108718e-01 7.92397782e-02 -4.23358619e-01 -1.29714906e+00
4.06083286e-01 -1.01104271e+00 5.93321502e-01 -7.04935133e-01
-1.55266702e+00 7.72115588e-01 1.71713725e-01 -1.26155198e+00
-6.53167605e-01 -3.12644124e-01 -6.49145782e-01 1.11451423e+00
-1.40616608e+00 -9.59725440e-01 -1.38436958e-01 3.84525001e-01
9.50610340e-01 -1.22218624e-01 7.35141933e-01 2.19548166e-01
-6.82322383e-01 6.28312528e-01 2.80263983e-02 2.72071093e-01
1.01309514e+00 -1.45720112e+00 8.07027698e-01 1.04151309e+00
1.43604293e-01 6.26544237e-01 8.80078495e-01 -7.62458742e-01
-1.26461887e+00 -1.31188810e+00 1.37603295e+00 -5.11694729e-01
8.45941246e-01 -5.86071908e-01 -9.85736132e-01 6.18641257e-01
2.99266130e-01 -1.51521593e-01 6.22412741e-01 2.11521432e-01
-1.95540667e-01 -1.44112229e-01 -1.03376997e+00 6.71396732e-01
9.15530920e-01 -3.33458722e-01 -9.84148145e-01 3.30601811e-01
1.08072615e+00 -8.48560154e-01 -6.70053601e-01 3.17690700e-01
1.70705002e-02 -5.65068543e-01 6.87445343e-01 -9.02732730e-01
5.74662805e-01 -2.60379344e-01 -1.78931177e-01 -1.55323708e+00
-5.37470318e-02 -8.75313818e-01 6.65674433e-02 1.65164804e+00
9.63398099e-01 -4.42732364e-01 3.89671475e-01 9.86945510e-01
-4.36239153e-01 -8.13477755e-01 -9.73432243e-01 -7.56199956e-01
7.39533156e-02 -4.12972867e-01 9.37811732e-01 8.07728112e-01
8.30426291e-02 9.12501395e-01 -1.69001043e-01 -4.39870544e-02
1.89713046e-01 2.73874342e-01 7.53631115e-01 -1.06386638e+00
-3.57937276e-01 -2.28025675e-01 -6.70359877e-04 -1.00990474e+00
-4.21410380e-03 -9.83604074e-01 2.61359721e-01 -1.72358727e+00
4.31995653e-02 -4.61528212e-01 -1.01919785e-01 7.74245024e-01
-5.99010289e-01 -3.22386920e-01 4.32950526e-01 7.50444606e-02
-6.10244751e-01 4.90101010e-01 1.03974175e+00 -2.24664330e-01
-3.66384804e-01 -2.87053213e-02 -1.15895891e+00 3.82524878e-01
8.03238690e-01 -4.86965746e-01 -5.10627925e-01 -7.72636414e-01
2.29251340e-01 2.82818854e-01 1.31612971e-01 -9.01620328e-01
2.29432911e-01 -2.15934396e-01 1.53731763e-01 -6.22087359e-01
3.71621758e-01 -4.67166960e-01 -1.18757680e-01 -3.51162665e-02
-4.98216480e-01 1.48597598e-01 2.89609015e-01 5.16583264e-01
-3.20581675e-01 -2.73198396e-01 5.29867053e-01 -1.75391987e-01
-6.63102984e-01 1.53736562e-01 8.89479369e-02 6.08852684e-01
7.58729815e-01 -1.47965625e-01 -7.46331155e-01 -1.14341609e-01
-7.25908279e-01 5.84286749e-01 2.36427873e-01 7.91943729e-01
3.52708787e-01 -1.24872637e+00 -9.26968455e-01 1.43249184e-01
2.76048362e-01 5.56227386e-01 1.87597811e-01 4.97551620e-01
3.26605365e-02 4.96510476e-01 9.79902521e-02 -4.46295828e-01
-1.22998321e+00 5.09863436e-01 7.57614449e-02 -5.33631265e-01
-3.75579029e-01 7.48920798e-01 -4.93454821e-02 -4.77159262e-01
1.21298552e-01 -5.39961755e-01 -3.65943722e-02 2.64701366e-01
5.26890397e-01 2.58440115e-02 3.89767587e-01 -4.13549900e-01
-1.87682346e-01 2.24370435e-01 -4.17748749e-01 -1.90931156e-01
1.06856036e+00 -1.47625491e-01 2.59022474e-01 4.49776262e-01
9.22503829e-01 -1.79384947e-01 -1.02962649e+00 -6.59639716e-01
1.57746702e-01 -1.98246285e-01 -3.13005477e-01 -1.20784366e+00
-5.73447168e-01 7.60878861e-01 9.98287797e-02 3.23990822e-01
1.12309480e+00 1.16000175e-01 1.10157406e+00 5.54094493e-01
1.44088358e-01 -1.26833248e+00 -1.32385567e-01 8.00972879e-01
9.58289742e-01 -1.48239243e+00 -2.75624514e-01 -3.32766622e-01
-8.71267915e-01 9.78233516e-01 6.75297081e-01 2.97675908e-01
1.90650687e-01 2.15675622e-01 1.40589982e-01 5.82689084e-02
-1.14877784e+00 -3.50576133e-01 3.19104791e-01 3.13895911e-01
4.02105510e-01 -1.88293420e-02 -3.46592903e-01 9.28327978e-01
-6.45476580e-01 -1.67460740e-01 4.35091466e-01 9.48687613e-01
-4.58634734e-01 -1.41499877e+00 -2.69721925e-01 7.24968851e-01
-4.28358585e-01 -2.87159622e-01 -5.82552195e-01 6.97212994e-01
2.23120570e-01 1.22202885e+00 -3.55589725e-02 -3.35368872e-01
6.53460324e-01 3.40282381e-01 5.09121045e-02 -1.10917115e+00
-7.13922381e-01 4.42257710e-02 4.81726408e-01 -2.70224571e-01
-1.19508550e-01 -8.20695698e-01 -1.49824476e+00 -1.10824272e-01
-2.66558319e-01 1.75240189e-01 5.61274648e-01 1.18135762e+00
7.35928893e-01 5.32731116e-01 3.15726489e-01 -6.07010663e-01
-7.37333000e-01 -1.18156481e+00 -2.69752532e-01 6.02700293e-01
3.36243331e-01 -4.99301434e-01 -1.11658126e-01 2.79082246e-02] | [11.589054107666016, 8.82402229309082] |
36a4b8d6-763a-42e8-af36-6ced43d0fe43 | efficient-vertical-federated-learning-method | null | null | https://ieeexplore.ieee.org/abstract/document/9930870 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9930870 | Efficient Vertical Federated Learning Method for Ridge Regression of Large-Scale Samples via Least-Squares Solution | Integrating data from multiple parties to achieve cross-institutional machine learning is an important trend in Industry 4.0 era. However, the privacy risks from sharing data pose a significant challenge to data integration. To integrate data without sharing data and meet large-scale samples' modeling needs, we propose two vertical federation learning algorithms for ridge regression via least-squares solution for two-party and multi-party scenarios, respectively. Compared with the state-of-the-art algorithms, our algorithms only need one round of calculation for the optimization instead of iteration. Furthermore, our algorithms can effectively handle large-scale samples due to the number of cryptographic operations in our algorithms being independent of the number of samples. Through our proposed the matrix secure agent computing theory and $\delta$ -data indistinguishability theory, we provide quantitative theoretical guarantees for the security of our algorithms. Our algorithms satisfy complete data indistinguishability under the “semi-honest” assumption and the quantitative security under the “malicious” assumption. The experiments show that our proposed algorithm takes only about 400 seconds to handle up to 9.6 million large-scale samples, while the state-of-the-art algorithms take close to 1000 seconds to handle every 1000 samples, which embodies the advantage of our algorithms in handling large-scale samples.δ -data indistinguishability theory, we provide quantitative theoretical guarantees for the security of our algorithms. Our algorithms satisfy complete data indistinguishability under the “semi-honest” assumption and the quantitative security under the “malicious” assumption. The experiments show that our proposed algorithm takes only about 400 seconds to handle up to 9.6 million large-scale samples, while the state-of-the-art algorithms take close to 1000 seconds to handle every 1000 samples, which embodies the advantage of our algorithms in handling large-scale samples. | ['Jiayin Li', 'Kun Guo', 'Zhiyong Yu', 'Ximeng Liu', 'Jianping Cai'] | 2022-10-26 | null | null | null | ieee-transactions-on-emerging-topics-in | ['data-integration'] | ['knowledge-base'] | [-3.15218657e-01 -1.99056625e-01 -4.08944279e-01 -2.96772450e-01
-1.12120461e+00 -9.20162678e-01 1.90237120e-01 9.96019468e-02
-5.90087056e-01 5.25849998e-01 -4.54597712e-01 -5.18234313e-01
-2.17082694e-01 -7.48569012e-01 -9.46960926e-01 -7.99512565e-01
-2.96904713e-01 4.73300308e-01 -1.99575260e-01 2.00137449e-03
-8.28254372e-02 3.18365425e-01 -1.04016376e+00 1.91757515e-01
5.19589961e-01 1.38407254e+00 -6.75219834e-01 2.79394746e-01
4.87179831e-02 5.36282837e-01 -2.76899070e-01 -7.08673596e-01
1.19594491e+00 -4.62855212e-02 -7.00409591e-01 -1.64857700e-01
2.40388975e-01 -4.17102665e-01 -2.52583206e-01 1.40995026e+00
1.96292117e-01 -5.42909384e-01 1.43091232e-01 -2.01115227e+00
-6.95899963e-01 8.43526363e-01 -8.49817395e-01 -5.93337476e-01
1.43249994e-02 2.67283529e-01 1.05077183e+00 -3.78563046e-01
5.67074895e-01 8.03369999e-01 5.04834294e-01 5.59730411e-01
-1.17023563e+00 -1.28914881e+00 4.85903807e-02 -2.35192031e-01
-1.53115976e+00 -1.95694342e-01 7.35159934e-01 -4.06790674e-01
3.52508873e-01 6.83819830e-01 3.92906874e-01 6.72507405e-01
1.45344108e-01 6.89683795e-01 1.48551095e+00 -2.95057683e-03
3.61146241e-01 7.91120887e-01 5.04953027e-01 4.02481407e-01
8.25595975e-01 2.28827998e-01 -3.63891780e-01 -9.29555714e-01
5.99088013e-01 4.34047073e-01 1.08229602e-02 -4.18740183e-01
-1.13435614e+00 8.02478373e-01 1.48526713e-01 4.23492454e-02
-1.01925150e-01 3.29085052e-01 5.87916970e-01 8.04720581e-01
9.31802839e-02 1.23744711e-01 -7.93080449e-01 2.47063205e-01
-4.31198835e-01 4.27075773e-01 1.01483774e+00 1.27919126e+00
8.81982446e-01 -1.75698817e-01 4.18891877e-01 9.40766409e-02
3.73806000e-01 5.96521139e-01 2.21763417e-01 -1.13469160e+00
7.75245786e-01 4.01434422e-01 3.44265252e-01 -7.98089266e-01
1.34049118e-01 -1.44657895e-01 -1.11712396e+00 3.60181004e-01
7.53193617e-01 -3.82237822e-01 -1.15051329e-01 1.79534864e+00
5.92505038e-01 -8.84856470e-03 4.38711435e-01 7.80152678e-01
5.77171333e-02 5.71793139e-01 -2.87615091e-01 -7.00425744e-01
1.65343738e+00 -6.10605597e-01 -5.33547819e-01 2.14595109e-01
7.67739236e-01 -4.57346886e-01 8.78199935e-01 3.94051820e-01
-8.25449347e-01 -1.30410016e-01 -1.15229964e+00 2.74391979e-01
-2.08727628e-01 -3.35138679e-01 9.71240580e-01 1.07619178e+00
-2.95826733e-01 2.70923883e-01 -3.98331970e-01 2.46862397e-01
5.09292126e-01 8.96586120e-01 -8.37191164e-01 1.95894524e-01
-8.46925974e-01 -7.25873336e-02 -1.02883903e-02 -1.53846502e-01
-6.59324348e-01 -8.39943171e-01 -4.31776077e-01 -7.81428516e-02
4.82978553e-01 -5.65286398e-01 1.02892947e+00 -7.68484473e-01
-1.46422172e+00 6.86272860e-01 2.66423762e-01 -4.92187917e-01
7.39818156e-01 8.13559815e-02 -3.34095687e-01 -1.86895832e-01
-1.53330162e-01 -1.18700787e-01 3.86092097e-01 -1.41915905e+00
-8.21719468e-01 -7.55351543e-01 1.25231549e-01 -2.09765464e-01
-3.59646887e-01 1.58653930e-01 -1.08555280e-01 -2.97353536e-01
-5.91373928e-02 -1.41683066e+00 -4.67042953e-01 9.21024680e-02
-3.66505831e-01 3.70976552e-02 9.47122514e-01 -2.99384892e-01
7.59705245e-01 -2.45571065e+00 -3.21518570e-01 6.39207840e-01
4.22443867e-01 9.86818150e-02 8.25809091e-02 3.49476784e-01
6.82511106e-02 3.89314473e-01 -8.48652273e-02 -2.22685516e-01
4.14912552e-01 -9.21503752e-02 -4.43104267e-01 7.79354334e-01
-4.78676617e-01 5.18763602e-01 -4.77015913e-01 -1.98203728e-01
-1.59184262e-01 -1.25662595e-01 -4.59169567e-01 1.32838175e-01
-1.83100216e-02 3.26815009e-01 -6.38054490e-01 4.79173720e-01
1.27624917e+00 -2.75650084e-01 6.70478821e-01 -2.26528540e-01
1.04045816e-01 -3.33768487e-01 -1.77948987e+00 1.36058795e+00
-2.08900511e-01 -3.22084606e-01 5.95674872e-01 -5.45344412e-01
6.62983000e-01 4.72981662e-01 5.79728305e-01 -5.31048834e-01
4.01716411e-01 5.79333007e-01 -9.10887420e-02 4.31041010e-02
3.98672223e-02 -4.78105009e-01 -7.75017738e-01 1.03716075e+00
-3.92439365e-01 1.96934447e-01 -3.36120635e-01 1.73564956e-01
9.96999085e-01 -4.43026245e-01 2.15987965e-01 -1.23264126e-01
6.38324916e-01 -4.11675036e-01 1.02683151e+00 5.25912285e-01
-3.91348094e-01 -1.21449344e-01 5.50834358e-01 -5.91922522e-01
-8.42851460e-01 -9.07340765e-01 3.22374374e-01 6.50327861e-01
4.07552421e-01 -4.52386945e-01 -8.37041318e-01 -1.06463838e+00
7.03606427e-01 2.80013859e-01 -2.78177202e-01 1.39045743e-02
-2.47263074e-01 -6.20049655e-01 5.86137772e-01 2.73512781e-01
7.56655157e-01 -2.80434847e-01 3.72186862e-02 -1.67607769e-01
1.56231940e-01 -1.14238727e+00 -6.18529499e-01 -1.28894625e-02
-5.48196197e-01 -1.13014674e+00 5.77589124e-03 -4.58701521e-01
7.29567528e-01 3.20796520e-01 2.26921350e-01 -3.39094177e-02
-6.29975274e-02 1.26603514e-01 1.37631595e-01 -5.01623273e-01
-4.82535154e-01 -1.52174950e-01 5.57650387e-01 3.68729800e-01
3.81869137e-01 -6.22611463e-01 -3.89533460e-01 4.05245811e-01
-9.46190417e-01 -2.64564335e-01 5.58003545e-01 4.68061090e-01
8.28671455e-01 1.84190139e-01 5.87958276e-01 -1.24791479e+00
5.60616076e-01 -5.02118886e-01 -1.23998582e+00 3.78225088e-01
-9.55403268e-01 1.96397811e-01 1.12478554e+00 -5.64141870e-01
-6.99773312e-01 1.38131872e-01 4.14282739e-01 -5.26611567e-01
1.72007322e-01 6.93748444e-02 -5.80116093e-01 -3.16991687e-01
2.97091901e-01 3.31408344e-04 2.60132432e-01 -4.61756915e-01
5.41842699e-01 1.15241897e+00 1.97034195e-01 -1.02556038e+00
1.26081336e+00 7.09802687e-01 3.64758939e-01 7.87901878e-02
-3.38619083e-01 -2.07485422e-01 -1.48201197e-01 2.92237908e-01
5.36924124e-01 -9.57392275e-01 -1.86263406e+00 5.74150860e-01
-6.91428185e-01 1.41953111e-01 -2.63692558e-01 5.41653514e-01
-5.38064599e-01 6.55746818e-01 -8.41301918e-01 -1.11822760e+00
-6.97270989e-01 -1.30348575e+00 5.57716250e-01 -1.13513298e-01
1.22267403e-01 -5.58593392e-01 -1.68764800e-01 9.26401615e-01
2.16670379e-01 3.28352332e-01 7.51120389e-01 -1.03684163e+00
-9.36614752e-01 -6.52220488e-01 -1.53541133e-01 4.17347193e-01
1.43173963e-01 -3.45762879e-01 -8.57658088e-01 -6.00232065e-01
1.97537571e-01 -2.67758161e-01 1.29467234e-01 -4.83515352e-01
1.37788200e+00 -7.82874048e-01 -1.30889520e-01 5.55084050e-01
1.39012063e+00 7.75274485e-02 4.23713505e-01 1.15153521e-01
5.98537326e-01 4.73565340e-01 7.68123746e-01 6.31089807e-01
3.66961747e-01 5.96986711e-01 3.79643738e-01 -3.10241263e-02
6.54432535e-01 -1.49016380e-01 3.70104969e-01 9.63505268e-01
1.08839102e-01 2.74484247e-01 -3.61324281e-01 1.39388159e-01
-1.98716962e+00 -1.00513399e+00 -1.94284096e-01 2.78200293e+00
1.12872374e+00 -3.53602991e-02 2.59112626e-01 1.64324231e-02
5.12838423e-01 -1.50800675e-01 -5.94757915e-01 -5.89418948e-01
2.54193157e-01 8.00457597e-02 1.20820284e+00 2.74382859e-01
-8.70989263e-01 5.26913881e-01 4.98240471e+00 8.78352821e-01
-1.06199241e+00 3.00676674e-01 6.72389805e-01 -2.37212986e-01
-4.36441571e-01 4.44872707e-01 -5.84223807e-01 6.90485775e-01
1.08292115e+00 -6.74840391e-01 7.24551618e-01 1.10301852e+00
-2.81260639e-01 3.25278997e-01 -1.42080081e+00 1.22771645e+00
-2.90054977e-01 -1.33644605e+00 -1.18575431e-01 7.20359147e-01
6.93735480e-01 -1.07147738e-01 2.18396708e-01 1.90035388e-01
6.44393265e-01 -7.26790905e-01 6.17958963e-01 3.09406221e-01
8.40453446e-01 -1.11204863e+00 8.00666690e-01 5.35866082e-01
-1.11884034e+00 -4.42452073e-01 -3.01764816e-01 1.52597614e-02
-1.70451343e-01 6.84783459e-01 -2.35744566e-01 1.08363199e+00
6.28517032e-01 -1.76157787e-01 -7.29928073e-03 4.06150579e-01
3.45968723e-01 2.99511164e-01 -5.99237502e-01 1.85508355e-01
-8.15847889e-02 -6.27361000e-01 1.38849497e-01 6.63799644e-01
4.34680730e-01 6.99326321e-02 3.99289906e-01 5.63242853e-01
-6.83320999e-01 3.00605237e-01 -5.28791070e-01 -1.64340466e-01
7.98813105e-01 1.24767029e+00 -3.81157808e-02 -3.04077834e-01
-4.53509271e-01 8.98260474e-01 2.27960035e-01 6.73020557e-02
-8.57363701e-01 -5.08999944e-01 1.07765329e+00 7.53640607e-02
1.28302425e-01 -1.10271320e-01 -3.76534790e-01 -1.18151605e+00
3.67602050e-01 -1.32126653e+00 5.56453645e-01 1.96734332e-02
-1.71734953e+00 1.08210303e-01 -3.34086746e-01 -1.51633668e+00
8.72827619e-02 -4.51871186e-01 -3.53778064e-01 6.86121702e-01
-1.10976481e+00 -1.43719304e+00 1.52422622e-01 8.63157332e-01
-5.13998270e-01 -4.76635367e-01 1.04504156e+00 3.37307602e-01
-4.82715189e-01 1.10598576e+00 5.48096359e-01 1.43403664e-01
8.99502158e-01 -8.28268468e-01 9.38607380e-02 5.47878325e-01
5.81037328e-02 1.08758175e+00 4.16740268e-01 -5.19399822e-01
-2.20139265e+00 -9.60529089e-01 5.45327365e-01 -2.40673915e-01
7.33953834e-01 -7.50808001e-01 -5.98101199e-01 9.72353637e-01
-4.79633994e-02 6.07655287e-01 1.38958097e+00 2.33879983e-01
-1.07799304e+00 -9.64061022e-01 -1.80660582e+00 3.83198619e-01
7.09922135e-01 -8.44029009e-01 9.11832005e-02 4.69903022e-01
8.07947099e-01 -2.23825663e-01 -1.21911216e+00 1.25047952e-01
7.33953774e-01 -5.86914599e-01 6.42959833e-01 -7.14952707e-01
-3.12383085e-01 -6.76062167e-01 -5.91952324e-01 -5.09760082e-01
-8.54695588e-02 -1.09840643e+00 -1.43963188e-01 1.46367836e+00
4.54755217e-01 -1.30432665e+00 7.32242286e-01 1.24234736e+00
6.45150959e-01 -4.54653502e-01 -1.18598902e+00 -1.01624250e+00
3.25856388e-01 -4.41803068e-01 1.09024870e+00 1.32748568e+00
2.26265103e-01 -2.28239089e-01 -6.70457423e-01 4.38608378e-01
1.09253502e+00 4.83134687e-01 1.32180107e+00 -1.21004915e+00
-6.76590443e-01 1.76118612e-02 -3.95045787e-01 -4.74127769e-01
2.65849799e-01 -9.94496226e-01 -6.36682451e-01 -7.69549787e-01
7.18916118e-01 -8.58833730e-01 -6.00834668e-01 4.67857808e-01
2.20141679e-01 1.32503003e-01 5.43046355e-01 3.65396321e-01
-6.14714682e-01 5.72158955e-02 8.20237100e-01 -1.82168305e-01
3.05559840e-02 1.50457352e-01 -1.13902104e+00 4.05488819e-01
7.41895378e-01 -6.71694160e-01 -4.30068970e-01 -1.28745250e-02
3.85124236e-01 1.85946703e-01 2.14934736e-01 -6.25232577e-01
2.37821206e-01 -6.10667706e-01 -6.41066432e-02 -9.14334953e-02
1.73178315e-01 -1.37532187e+00 8.22493196e-01 6.24330223e-01
-6.89795688e-02 -1.86696529e-01 -1.26440927e-01 7.05252230e-01
1.82602912e-01 6.12746030e-02 5.46786726e-01 -1.23788612e-02
2.51041502e-01 5.92888415e-01 1.31405801e-01 -2.86406934e-01
1.64259720e+00 8.50666091e-02 -5.24673700e-01 -3.14361125e-01
-4.54261661e-01 3.62683862e-01 8.37172747e-01 1.22247063e-01
1.66025925e-02 -1.56588483e+00 -4.80001867e-01 3.88627380e-01
5.77035509e-02 -2.04082832e-01 1.34426653e-01 8.32009315e-01
-2.34479800e-01 2.10642353e-01 1.04761412e-02 -1.31659985e-01
-1.57963562e+00 9.88552988e-01 7.45704547e-02 -3.75563741e-01
-1.52649190e-02 4.47153598e-01 1.53246909e-01 -6.92727447e-01
8.79960358e-02 -1.56305447e-01 8.80285501e-01 -3.88178110e-01
5.08711398e-01 4.38390851e-01 -1.95557266e-01 -2.75570512e-01
-2.86945581e-01 4.37755555e-01 -4.49943691e-01 -1.81957930e-01
1.16171205e+00 -3.92819289e-03 -4.09258604e-01 2.91070431e-01
1.37401056e+00 6.78731740e-01 -7.10384488e-01 -2.80808151e-01
-2.84201831e-01 -7.50088990e-01 -2.69686311e-01 -6.01185143e-01
-1.23109746e+00 6.24087453e-01 7.66326010e-01 7.83381164e-02
8.49751651e-01 -2.26743728e-01 1.11210239e+00 5.13997138e-01
1.13809097e+00 -1.23494172e+00 -5.01115263e-01 -1.59679741e-01
3.18581849e-01 -1.12295306e+00 3.96384209e-01 -6.92681313e-01
-7.68542588e-01 7.71810114e-01 4.51886237e-01 -1.03838742e-02
8.00372124e-01 5.91413617e-01 1.01061217e-01 2.21425414e-01
-7.01738596e-01 5.05001307e-01 -3.04196388e-01 4.19122636e-01
-1.54580757e-01 4.51842844e-01 -3.30444157e-01 1.21664023e+00
-3.05156503e-02 1.45391881e-01 4.63758498e-01 9.02882099e-01
1.94267109e-01 -1.81527138e+00 -2.89156199e-01 1.92917496e-01
-6.51911795e-01 3.39730680e-01 -3.94521058e-01 7.36166239e-01
1.33148760e-01 8.27564657e-01 -1.79232076e-01 -5.99047601e-01
9.56526473e-02 8.95980597e-02 3.49274009e-01 -3.27743255e-02
-1.00646985e+00 -6.35572746e-02 4.48376797e-02 -6.79345548e-01
-1.00695141e-01 -6.80998147e-01 -1.52497029e+00 -9.88998115e-01
-5.64772427e-01 3.34167272e-01 7.58431554e-01 5.73041201e-01
7.15896904e-01 -3.32785428e-01 1.30479658e+00 -1.35317057e-01
-1.22900009e+00 -3.78621280e-01 -1.05350220e+00 5.78340650e-01
-1.91492047e-02 3.58569771e-02 -5.17255545e-01 -2.03354925e-01] | [5.834895610809326, 6.667096138000488] |
63645777-50cc-4c10-bd54-d8e3e4e3aad0 | sft-kd-recon-learning-a-student-friendly | 2304.05057 | null | https://arxiv.org/abs/2304.05057v1 | https://arxiv.org/pdf/2304.05057v1.pdf | SFT-KD-Recon: Learning a Student-friendly Teacher for Knowledge Distillation in Magnetic Resonance Image Reconstruction | Deep cascaded architectures for magnetic resonance imaging (MRI) acceleration have shown remarkable success in providing high-quality reconstruction. However, as the number of cascades increases, the improvements in reconstruction tend to become marginal, indicating possible excess model capacity. Knowledge distillation (KD) is an emerging technique to compress these models, in which a trained deep teacher network is used to distill knowledge to a smaller student network such that the student learns to mimic the behavior of the teacher. Most KD methods focus on effectively training the student with a pre-trained teacher unaware of the student model. We propose SFT-KD-Recon, a student-friendly teacher training approach along with the student as a prior step to KD to make the teacher aware of the structure and capacity of the student and enable aligning the representations of the teacher with the student. In SFT, the teacher is jointly trained with the unfolded branch configurations of the student blocks using three loss terms - teacher-reconstruction loss, student-reconstruction loss, and teacher-student imitation loss, followed by KD of the student. We perform extensive experiments for MRI acceleration in 4x and 5x under-sampling on the brain and cardiac datasets on five KD methods using the proposed approach as a prior step. We consider the DC-CNN architecture and setup teacher as D5C5 (141765 parameters), and student as D3C5 (49285 parameters), denoting a compression of 2.87:1. Results show that (i) our approach consistently improves the KD methods with improved reconstruction performance and image quality, and (ii) the student distilled using our approach is competitive with the teacher, with the performance gap reduced from 0.53 dB to 0.03 dB. | ['Mohanasankar Sivaprakasam', 'Keerthi Ram', 'Rahul G S', 'Mohammad Al Fahim', 'Sriprabha Ramanarayanan', 'Matcha Naga Gayathri'] | 2023-04-11 | null | null | null | null | ['image-reconstruction'] | ['computer-vision'] | [ 1.07162483e-01 3.47505718e-01 -1.01399757e-01 -3.10354978e-01
-7.07187831e-01 -1.81412175e-01 2.92582035e-01 -2.79282425e-02
-5.91584802e-01 4.89639759e-01 2.84530252e-01 -4.72093284e-01
-1.21354144e-02 -4.74410236e-01 -1.04387987e+00 -9.32706535e-01
-1.98744133e-01 4.69218314e-01 3.69081855e-01 5.87935634e-02
-1.99515119e-01 5.84990740e-01 -8.16215813e-01 3.26835960e-01
8.10615242e-01 1.00045860e+00 4.27926809e-01 7.09225893e-01
3.35212052e-01 1.16608322e+00 -4.53878075e-01 -3.89329493e-01
2.43254483e-01 -3.89980853e-01 -1.02703583e+00 6.54019862e-02
2.40516439e-01 -7.15766072e-01 -8.21821988e-01 9.70286667e-01
8.23089421e-01 -3.28095979e-04 4.98479307e-01 -7.92418540e-01
-2.51360744e-01 8.55661333e-01 -9.38899100e-01 5.88795960e-01
-4.88073647e-01 2.51219541e-01 4.48916376e-01 -8.65603626e-01
4.03282642e-01 1.11669266e+00 5.94121337e-01 6.18183434e-01
-1.24279022e+00 -9.65821981e-01 4.03547697e-02 2.20512047e-01
-1.26765239e+00 -3.17989469e-01 6.05515718e-01 -2.37004966e-01
7.48954177e-01 -1.76762566e-01 7.15539098e-01 9.30594683e-01
2.52653599e-01 7.98635364e-01 9.53722954e-01 -1.01210438e-01
4.34714183e-02 2.19246253e-01 2.41148397e-01 7.78416991e-01
-5.60142063e-02 2.26217091e-01 -4.33122545e-01 -1.13876455e-01
1.18686700e+00 4.35081683e-02 -4.02179927e-01 -2.66024441e-01
-1.06306875e+00 7.34100163e-01 9.18976784e-01 9.26578194e-02
-7.26408899e-01 3.46712381e-01 4.50352252e-01 4.10371542e-01
3.81468058e-01 2.27646455e-01 -5.80436945e-01 1.84461609e-01
-1.03455794e+00 1.87520638e-01 5.03038049e-01 8.23648870e-01
3.93295646e-01 4.20523494e-01 -2.90115297e-01 7.70982325e-01
8.11383799e-02 3.54345948e-01 7.00428247e-01 -8.35770249e-01
3.97050440e-01 2.71629781e-01 -4.07532096e-01 -6.90789461e-01
-2.05347463e-01 -1.11680865e+00 -1.11912906e+00 -5.33146318e-03
1.97969258e-01 -2.60932058e-01 -1.07547438e+00 1.81003630e+00
4.37280536e-01 7.08585322e-01 1.46794155e-01 9.19191658e-01
9.05001938e-01 6.16586804e-01 8.28515217e-02 -1.65728509e-01
1.25276852e+00 -1.21386313e+00 -2.91840285e-01 -3.20124105e-02
6.89341426e-01 -6.20747209e-01 9.15226460e-01 5.58198869e-01
-1.44109142e+00 -5.90430498e-01 -1.02970147e+00 1.91224232e-01
3.99918139e-01 1.29426852e-01 3.70384663e-01 2.26236001e-01
-1.13013780e+00 8.48426104e-01 -1.18113947e+00 3.69375437e-01
8.87008905e-01 5.62495828e-01 -1.16061317e-02 -3.39549333e-01
-1.02637970e+00 7.58148968e-01 3.66651684e-01 -4.62202765e-02
-1.71101773e+00 -1.36871421e+00 -4.38424408e-01 2.99624681e-01
8.66653547e-02 -7.77038753e-01 1.22955048e+00 -1.08151889e+00
-1.62084961e+00 5.88213742e-01 3.81280869e-01 -7.97050238e-01
7.66873777e-01 -3.53226095e-01 -1.50010556e-01 6.44777894e-01
-2.16995984e-01 1.02329290e+00 9.93841231e-01 -1.11324120e+00
-3.06709260e-01 -2.19300449e-01 6.96145296e-02 3.28089386e-01
-2.19502583e-01 -1.62753403e-01 -4.67092186e-01 -8.34605455e-01
1.04254492e-01 -9.91535008e-01 -3.12209785e-01 5.34046628e-02
-3.47228467e-01 1.89404562e-01 7.91083038e-01 -8.25900555e-01
9.97539818e-01 -2.27879953e+00 1.28168076e-01 2.25572363e-01
5.27395129e-01 4.55518395e-01 -2.64676601e-01 -1.38534471e-01
-5.02197981e-01 -3.17660630e-01 -1.34858817e-01 -3.62103850e-01
-6.88125789e-01 4.38204914e-01 -3.69634181e-01 5.33442795e-01
2.05615938e-01 8.28094125e-01 -9.11148071e-01 -3.85556191e-01
-7.93496892e-02 8.26093376e-01 -9.64591503e-01 3.43404114e-01
1.61372393e-01 8.19165409e-01 -4.29585099e-01 2.75692791e-01
6.75225675e-01 -5.65337539e-01 3.05813849e-01 -3.69991899e-01
3.07397783e-01 3.14170808e-01 -9.52404082e-01 1.57727194e+00
-3.99466068e-01 3.07706177e-01 1.65177748e-01 -1.35308766e+00
7.10150003e-01 4.75945055e-01 4.08279300e-01 -5.00065744e-01
1.66590393e-01 1.17398158e-01 4.13203478e-01 -3.11647177e-01
-6.27322719e-02 -2.45354161e-01 4.31733757e-01 3.84633452e-01
3.53016287e-01 -3.38356122e-02 -3.74468029e-01 5.37471652e-01
1.05478585e+00 -2.79696882e-01 -1.65982619e-01 -3.80499005e-01
2.85750896e-01 -2.16713056e-01 4.91891116e-01 6.34940028e-01
-9.54220816e-02 4.06716824e-01 4.62821096e-01 -3.86372596e-01
-1.16536117e+00 -1.08767176e+00 -1.02098495e-01 7.12103426e-01
-1.79786101e-01 -3.46712694e-02 -8.41547668e-01 -7.10063934e-01
-3.26460809e-01 4.12873268e-01 -4.88438666e-01 -4.64997381e-01
-9.32188630e-01 -7.72766113e-01 5.81051826e-01 6.57380819e-01
8.63705814e-01 -7.67553806e-01 -6.07065499e-01 2.95184404e-01
-7.49694556e-02 -1.12905037e+00 -5.88673949e-01 3.03303093e-01
-1.48598635e+00 -9.53330398e-01 -8.81223857e-01 -9.57677186e-01
1.06457853e+00 2.77480513e-01 9.54188228e-01 4.21273440e-01
-1.03692114e-02 1.89895123e-01 -6.93465173e-02 -5.95318489e-02
-4.39158022e-01 4.79227006e-02 1.55735642e-01 -1.27340630e-01
-3.39379847e-01 -8.73257160e-01 -9.51922417e-01 1.43908396e-01
-9.69347596e-01 2.49492973e-01 8.70003343e-01 1.20814395e+00
6.90931022e-01 5.61627522e-02 5.03835618e-01 -8.96553755e-01
2.08836690e-01 -5.51358163e-01 -3.34626853e-01 -1.30296275e-01
-6.63601935e-01 8.92624781e-02 7.22853303e-01 -9.80187118e-01
-9.15771127e-01 9.31500345e-02 -1.73043430e-01 -1.13033533e+00
3.03480923e-01 5.62826574e-01 1.24655366e-01 -7.08387643e-02
5.22206128e-01 3.90620083e-01 1.26101390e-01 -5.30511975e-01
2.01147586e-01 5.97701371e-02 8.65188003e-01 -7.38411367e-01
6.95271850e-01 3.49386871e-01 -1.16433233e-01 -6.89228356e-01
-6.99688137e-01 5.78091759e-03 -4.11001772e-01 -1.18191428e-01
5.77187240e-01 -1.23276317e+00 -4.67824847e-01 6.34780467e-01
-9.31406617e-01 -6.00330174e-01 -4.12196815e-01 7.68622100e-01
-2.44844213e-01 8.59881341e-02 -1.00648034e+00 -8.64989385e-02
-6.26424670e-01 -1.57458794e+00 4.74480689e-01 2.31073231e-01
1.62098318e-01 -9.24462020e-01 -3.97975385e-01 4.82462108e-01
5.93876421e-01 2.91079711e-02 1.30025411e+00 -7.66098201e-01
-4.81795996e-01 1.07540809e-01 -1.64147511e-01 6.26296222e-01
-2.06788898e-01 -5.24886847e-01 -7.97508836e-01 -7.06833303e-01
3.48811537e-01 -5.10952294e-01 7.92759478e-01 5.42993248e-01
1.46155870e+00 -5.41357517e-01 -1.50984243e-01 7.56462038e-01
1.24491334e+00 4.95868176e-02 6.70454383e-01 7.44592920e-02
7.71632612e-01 2.62456328e-01 1.31320637e-02 2.05785573e-01
3.64339024e-01 3.05281907e-01 4.78597641e-01 -3.08499366e-01
-4.45561945e-01 -2.65705317e-01 3.73007327e-01 1.38904846e+00
8.17009732e-02 3.03251833e-01 -8.67637575e-01 5.00931680e-01
-1.31265044e+00 -5.22324741e-01 1.27692297e-01 1.98560297e+00
1.19492662e+00 2.90353090e-01 1.16733626e-01 1.52372286e-01
4.59037662e-01 -1.01682290e-01 -8.47605526e-01 -1.00263953e-01
1.95427924e-01 4.02052879e-01 4.20908719e-01 4.78826374e-01
-8.05575550e-01 7.59110451e-01 5.25681496e+00 9.14929032e-01
-1.56394136e+00 4.24184144e-01 1.08757985e+00 -2.61268407e-01
1.06361344e-01 -2.13319793e-01 -6.46127105e-01 3.69230777e-01
1.07675815e+00 1.39204100e-01 4.39757198e-01 7.28458285e-01
1.02062270e-01 3.26546505e-02 -1.05288255e+00 7.18788326e-01
-1.90504521e-01 -1.28224242e+00 1.72358811e-01 4.28481773e-02
8.98796797e-01 1.68791771e-01 3.38501275e-01 4.18644100e-01
2.89976865e-01 -9.71165955e-01 5.99826217e-01 2.57365882e-01
7.67958701e-01 -9.74395752e-01 6.70330703e-01 5.36816478e-01
-7.75508404e-01 8.08516145e-02 -2.80685872e-01 2.02752754e-01
-2.32529432e-01 5.41186631e-01 -1.10803568e+00 3.25807631e-01
9.25761342e-01 4.66193318e-01 -2.19297796e-01 8.19831371e-01
-3.14193428e-01 1.10506618e+00 -2.51154959e-01 4.50914741e-01
4.28771406e-01 1.06258951e-01 4.47658181e-01 1.03998709e+00
2.83056721e-02 5.20524800e-01 2.92798817e-01 6.52845085e-01
-5.00744045e-01 -2.14790195e-01 -8.45570788e-02 2.34354466e-01
3.74834806e-01 1.03655076e+00 -6.68226063e-01 -6.37102306e-01
-2.16566429e-01 8.46236587e-01 4.52465653e-01 2.74239600e-01
-9.55721021e-01 -6.54009124e-03 4.41637039e-01 1.94382668e-01
5.69047809e-01 -6.30410686e-02 -4.47923206e-02 -8.78935039e-01
-1.30952895e-01 -1.02092087e+00 2.86036283e-01 -6.48787022e-01
-9.45296466e-01 7.51907706e-01 1.31231576e-01 -1.03103042e+00
-5.51150404e-02 -2.10336387e-01 -6.75041318e-01 8.94010365e-01
-1.59383690e+00 -9.60105479e-01 -7.48920590e-02 7.97005355e-01
4.38001096e-01 -3.61470431e-02 4.78038341e-01 6.58055425e-01
-5.92220843e-01 8.63456666e-01 7.00918734e-02 2.93142706e-01
3.85450006e-01 -8.75058830e-01 3.25438101e-03 6.20792568e-01
-1.71996862e-01 6.05089843e-01 5.54241836e-01 -5.40852726e-01
-1.24089134e+00 -1.24012375e+00 4.59901154e-01 4.20840271e-02
6.21483862e-01 -8.09781551e-02 -1.26240981e+00 6.10028148e-01
4.31805477e-02 3.48527521e-01 3.26262593e-01 -4.63669419e-01
-2.36493185e-01 -2.35194221e-01 -1.22166121e+00 3.43344897e-01
6.09545887e-01 -3.90916884e-01 -4.05705094e-01 3.32808882e-01
9.44987535e-01 -8.92497718e-01 -1.10844803e+00 3.15486223e-01
4.21033293e-01 -5.35651207e-01 1.09685624e+00 -6.11505032e-01
6.94740236e-01 -6.54452294e-02 1.50611833e-01 -1.48978007e+00
-1.86941877e-01 -2.94686466e-01 -3.16038638e-01 8.34085166e-01
2.84468949e-01 -4.13309664e-01 9.17659044e-01 3.64875823e-01
-4.62104142e-01 -1.40613556e+00 -1.02832758e+00 -6.28063381e-01
5.18530130e-01 -1.49921730e-01 4.27853495e-01 1.05463600e+00
-4.25674707e-01 4.46752578e-01 -4.37264323e-01 3.17160934e-01
5.22446990e-01 -2.46398136e-01 4.26147640e-01 -7.16118097e-01
-6.58337235e-01 -2.07699537e-01 -5.51932212e-03 -1.44774759e+00
-1.22774594e-01 -1.06335783e+00 -2.70978928e-01 -9.24174964e-01
3.60463679e-01 -6.05204403e-01 -4.38877970e-01 5.25261819e-01
-1.97551578e-01 6.44856598e-03 1.33482620e-01 3.28573644e-01
-1.55731589e-01 6.38829589e-01 1.77922571e+00 -2.47244090e-01
-1.83681339e-01 9.12985802e-02 -5.87645888e-01 7.76942134e-01
6.33122563e-01 -7.17100501e-01 -6.58574820e-01 -7.78950453e-01
-3.03981185e-01 3.80108654e-01 4.83054012e-01 -9.89205480e-01
4.25034106e-01 4.49723810e-01 6.90210283e-01 -4.71053630e-01
2.42675871e-01 -8.34793687e-01 -3.84431565e-03 9.65368330e-01
-4.03859764e-01 1.21775404e-01 3.65567118e-01 3.92761052e-01
-1.03264511e-01 -1.14662334e-01 1.31690156e+00 -8.87732878e-02
-1.85336277e-01 5.45069396e-01 -1.17036588e-01 1.23252556e-01
5.91715693e-01 7.56676942e-02 -1.08147651e-01 -3.10890585e-01
-8.95178854e-01 4.27352816e-01 -1.40302151e-01 -3.65641229e-02
9.25889015e-01 -1.21342874e+00 -8.30981374e-01 1.99926317e-01
-5.72963715e-01 3.62622231e-01 5.34299493e-01 1.17602396e+00
-4.75218058e-01 1.28847986e-01 -1.34232610e-01 -7.42033422e-01
-1.12942934e+00 4.31211472e-01 5.62848330e-01 -6.20076299e-01
-1.00654328e+00 1.15180576e+00 4.92424756e-01 -4.26906556e-01
5.30211747e-01 -4.71265703e-01 -3.88445184e-02 -3.32543761e-01
5.61834872e-01 3.87599409e-01 1.60353675e-01 -3.62100154e-01
-2.44103923e-01 1.62926704e-01 -6.34512842e-01 6.21718056e-02
1.60388792e+00 3.09608012e-01 2.15492442e-01 -1.51494086e-01
1.34276342e+00 -5.05774260e-01 -1.41689968e+00 -7.46656656e-01
-3.54341209e-01 -6.86398521e-02 5.16001523e-01 -9.25538659e-01
-1.78972626e+00 1.03235507e+00 8.29452276e-01 -5.56510329e-01
1.35462117e+00 6.20183982e-02 1.01592350e+00 1.91961259e-01
2.09886909e-01 -6.28801763e-01 5.79352558e-01 4.85223293e-01
8.60463440e-01 -9.66442347e-01 1.41827658e-01 -1.49647504e-01
-7.52203524e-01 9.03243423e-01 7.38354564e-01 -3.50493848e-01
8.09877157e-01 2.51451254e-01 -1.34002000e-01 -3.03426415e-01
-8.90554488e-01 5.44508576e-01 2.43449092e-01 2.39191413e-01
1.85654268e-01 -7.42618553e-03 1.92357048e-01 7.36319780e-01
-1.78143561e-01 -1.23730963e-02 3.84667873e-01 8.14512491e-01
-2.46621907e-01 -7.16144145e-01 -1.98170036e-01 6.35382056e-01
-6.69722617e-01 -3.43296915e-01 2.57679939e-01 7.45440841e-01
3.10483407e-02 4.64598954e-01 -5.40583357e-02 -3.96069765e-01
4.07054007e-01 -2.97884762e-01 5.10176301e-01 -5.13794661e-01
-1.01238286e+00 4.12547439e-01 -1.96712300e-01 -5.02801180e-01
-1.59378380e-01 -5.19295931e-01 -1.42503726e+00 -4.41456199e-01
-4.39561039e-01 2.15567276e-01 4.30431068e-01 7.75623441e-01
3.30126286e-01 8.50345612e-01 8.03807557e-01 -7.26604104e-01
-9.14360166e-01 -9.09456968e-01 -4.81582671e-01 1.62808776e-01
4.64794457e-01 -5.14714241e-01 -2.30103850e-01 -4.78641503e-03] | [13.75632381439209, -2.348465919494629] |
485ea4f8-7e77-45e9-a864-a7f73491d3a8 | an-adaptive-simulated-annealing-based-machine | 2212.11892 | null | https://arxiv.org/abs/2212.11892v1 | https://arxiv.org/pdf/2212.11892v1.pdf | An Adaptive Simulated Annealing-Based Machine Learning Approach for Developing an E-Triage Tool for Hospital Emergency Operations | Patient triage at emergency departments (EDs) is necessary to prioritize care for patients with critical and time-sensitive conditions. Different tools are used for patient triage and one of the most common ones is the emergency severity index (ESI), which has a scale of five levels, where level 1 is the most urgent and level 5 is the least urgent. This paper proposes a framework for utilizing machine learning to develop an e-triage tool that can be used at EDs. A large retrospective dataset of ED patient visits is obtained from the electronic health record of a healthcare provider in the Midwest of the US for three years. However, the main challenge of using machine learning algorithms is that most of them have many parameters and without optimizing these parameters, developing a high-performance model is not possible. This paper proposes an approach to optimize the hyperparameters of machine learning. The metaheuristic optimization algorithms simulated annealing (SA) and adaptive simulated annealing (ASA) are proposed to optimize the parameters of extreme gradient boosting (XGB) and categorical boosting (CaB). The newly proposed algorithms are SA-XGB, ASA-XGB, SA-CaB, ASA-CaB. Grid search (GS), which is a traditional approach used for machine learning fine-tunning is also used to fine-tune the parameters of XGB and CaB, which are named GS-XGB and GS-CaB. The six algorithms are trained and tested using eight data groups obtained from the feature selection phase. The results show ASA-CaB outperformed all the proposed algorithms with accuracy, precision, recall, and f1 of 83.3%, 83.2%, 83.3%, 83.2%, respectively. | ['Dursun Delen', 'Mohammad Firouz', 'Mohammed Al-Maamari', 'Abdulaziz Ahmed'] | 2022-12-22 | null | null | null | null | ['metaheuristic-optimization'] | ['methodology'] | [ 7.01962709e-02 -4.53464359e-01 3.78102474e-02 -3.49361926e-01
-5.75186610e-01 -2.58995861e-01 -6.65525498e-04 7.92376220e-01
-5.09025693e-01 9.25586879e-01 2.83725232e-01 -5.80871165e-01
-9.40973461e-01 -5.60523689e-01 -8.53972360e-02 -8.41690481e-01
-1.19597219e-01 7.88686514e-01 -1.37449829e-02 -2.93236166e-01
5.57511091e-01 4.59004611e-01 -1.60080075e+00 5.90699017e-01
1.16651428e+00 1.10167730e+00 6.43051788e-02 7.86646068e-01
-5.50707541e-02 2.45500207e-01 -4.04889226e-01 -3.53219509e-02
5.02463102e-01 -6.12100124e-01 -6.92829967e-01 -3.28645796e-01
-3.39301407e-01 1.76515773e-01 5.36109626e-01 3.50552112e-01
8.64037812e-01 2.80103862e-01 6.98865354e-01 -1.26656830e+00
2.58151978e-01 1.66255087e-01 -5.01183271e-01 6.04054332e-01
3.82648528e-01 4.35411707e-02 5.00881851e-01 -3.14458281e-01
1.21678591e-01 1.20451701e+00 6.76279068e-01 3.03275287e-01
-7.72256494e-01 -5.36700547e-01 -5.20411134e-02 5.30269206e-01
-1.10531092e+00 8.97698700e-02 5.38427889e-01 -4.97307479e-01
1.03516400e+00 7.62354970e-01 8.95944059e-01 4.02122021e-01
5.03146768e-01 2.11334288e-01 1.59817255e+00 -7.29943693e-01
4.60210979e-01 1.27192408e-01 2.85162896e-01 5.18670380e-01
3.83467495e-01 2.09907770e-01 -3.56816888e-01 -7.19296575e-01
8.59755948e-02 3.36799592e-01 -1.50323391e-01 -1.09312065e-01
-9.47323143e-01 9.38231826e-01 4.11017627e-01 2.89849639e-01
-9.73536313e-01 -6.50229871e-01 7.63838768e-01 1.24915250e-01
-2.37799566e-02 5.09104013e-01 -7.67682016e-01 -3.77278894e-01
-5.81886470e-01 2.06156492e-01 6.73782110e-01 3.25264394e-01
2.25779891e-01 -3.29618871e-01 -2.34438285e-01 6.91781342e-01
2.69385159e-01 3.13331336e-01 8.58132660e-01 -2.95227170e-01
5.61420083e-01 9.23054755e-01 3.89967471e-01 -9.77701068e-01
-8.12238812e-01 -3.10861439e-01 -7.61152327e-01 -2.41818186e-03
-4.19408269e-02 -4.67177629e-01 -9.06941116e-01 1.07663143e+00
6.93141401e-01 -4.69258577e-02 2.18409494e-01 7.27797687e-01
6.80079222e-01 5.90459466e-01 4.12105590e-01 -6.58137143e-01
1.41946590e+00 -7.94503450e-01 -5.44229925e-01 1.74055006e-02
6.16609752e-01 -8.05907130e-01 8.15856576e-01 7.50659466e-01
-9.13662791e-01 -3.32640678e-01 -9.08261359e-01 7.48060644e-01
-3.65648717e-01 -1.24455228e-01 5.87954104e-01 7.38150179e-01
-7.92140901e-01 4.37419116e-01 -7.19814241e-01 -6.10749722e-01
8.10332298e-02 6.88377798e-01 -1.73311710e-01 -7.94116110e-02
-9.55478907e-01 1.11253929e+00 4.84279305e-01 5.67736439e-02
-2.92119563e-01 -3.81630063e-01 -3.88256699e-01 -5.17378226e-02
-4.83764745e-02 -8.39851499e-01 8.27089787e-01 -7.11383462e-01
-1.28856552e+00 6.81738615e-01 -1.38412163e-01 -2.93350428e-01
3.62458348e-01 -1.85069039e-01 -3.98410320e-01 5.10294251e-02
-1.68355197e-01 1.24337092e-01 3.40555072e-01 -1.19419765e+00
-1.06268966e+00 -7.40226865e-01 -3.09040427e-01 4.76560295e-01
-8.61754864e-02 2.51705199e-01 3.71329695e-01 -2.78707772e-01
3.42393547e-01 -8.42374146e-01 -4.40292358e-01 -9.45658982e-01
-3.54428403e-02 -1.27091110e-01 5.66320896e-01 -8.64324510e-01
1.71341121e+00 -1.81150091e+00 -3.89154032e-02 6.47065878e-01
-3.34384710e-01 5.45696795e-01 4.80298072e-01 7.82289267e-01
-3.01490963e-01 -9.90511850e-02 -2.40859985e-01 2.38414526e-01
-3.64058465e-01 2.94984907e-01 1.52991980e-01 7.35388622e-02
-2.30212599e-01 4.27651018e-01 -7.24000633e-01 -5.52941799e-01
5.07584512e-01 2.98974991e-01 -6.09941840e-01 5.42499840e-01
3.82515341e-01 8.16276431e-01 -6.12165034e-01 6.85604274e-01
6.23592675e-01 -8.23838562e-02 1.38011515e-01 -2.09238529e-01
-1.22075602e-01 1.06417641e-01 -1.38982475e+00 7.78839827e-01
-5.38561225e-01 -4.41229939e-01 -7.90542066e-02 -1.30083299e+00
1.14227247e+00 4.26800847e-01 8.88128638e-01 -5.86626351e-01
4.70204800e-01 3.01847845e-01 8.06570128e-02 -1.11578596e+00
-8.37891400e-02 -2.84062237e-01 9.97832641e-02 4.31783766e-01
-5.40195465e-01 1.58086076e-01 1.70482442e-01 -2.54331857e-01
1.03825724e+00 -4.46027011e-01 7.32170343e-01 -2.44886518e-01
8.65766943e-01 1.51052132e-01 7.78114617e-01 5.10320902e-01
-2.32110485e-01 3.31216544e-01 1.20439075e-01 -8.58441472e-01
-7.48698354e-01 -6.63415968e-01 -2.66746700e-01 9.58564162e-01
-1.17982149e-01 -3.41472775e-02 -5.76812804e-01 -3.79598439e-01
9.27585550e-03 7.19529152e-01 -4.06451553e-01 -2.13572726e-01
-7.12695956e-01 -1.28116643e+00 -1.52165204e-01 2.00508684e-01
5.83144367e-01 -1.10551691e+00 -1.36602592e+00 3.23391676e-01
-1.38930112e-01 -4.43749487e-01 6.42928779e-02 3.13313425e-01
-1.20171607e+00 -1.09187806e+00 -3.25613171e-01 -5.53520322e-01
6.89805329e-01 -1.24353416e-01 1.03049588e+00 3.31424922e-01
-5.98226190e-01 1.27848238e-01 -8.18932772e-01 -5.24122715e-01
-3.30593944e-01 -3.55659537e-02 4.79450822e-02 3.34173217e-02
4.59603578e-01 -4.50572252e-01 -9.82767105e-01 3.08879435e-01
-7.01876700e-01 -2.68467963e-01 6.85933828e-01 1.05110693e+00
5.73865414e-01 -8.36057141e-02 7.61142612e-01 -8.33995283e-01
9.68905210e-01 -6.94594145e-01 -2.94338077e-01 3.83434832e-01
-1.08024263e+00 -1.61993638e-01 4.93130982e-01 -2.28960663e-01
-7.25084662e-01 -2.20704991e-02 -2.66469091e-01 1.63677588e-01
-3.61603856e-01 5.14264405e-01 1.39375567e-01 -1.13544492e-02
6.63354695e-01 -1.26535431e-01 -3.39673400e-01 -4.44813102e-01
-5.61634302e-01 1.20548594e+00 -4.11302336e-02 -4.54432100e-01
7.21865334e-04 6.38074800e-02 2.20903158e-01 -2.24753127e-01
-6.12059414e-01 -8.78365397e-01 -3.00538778e-01 -1.80418164e-01
8.23772073e-01 -3.99637312e-01 -8.87838125e-01 1.01798527e-01
-4.70711589e-01 2.16474339e-01 1.70407459e-01 7.70669937e-01
-3.98342341e-01 6.99520186e-02 -2.72989459e-02 -1.11136091e+00
-1.03191483e+00 -1.23335993e+00 6.78671181e-01 3.77265722e-01
-4.00901318e-01 -7.72895515e-01 2.05730349e-02 6.17033362e-01
4.87327099e-01 8.65260303e-01 1.23215306e+00 -8.61798048e-01
9.05327573e-02 -2.91097730e-01 2.60245651e-01 2.12593302e-01
4.05289590e-01 -7.93792829e-02 -4.76750731e-01 -3.48450989e-01
3.05128902e-01 3.07506658e-02 1.77217081e-01 6.89439118e-01
1.01019120e+00 -4.48875070e-01 -4.68012124e-01 6.14826620e-01
1.59143770e+00 1.23488283e+00 4.40644979e-01 8.93397689e-01
1.64945483e-01 4.86817300e-01 1.28804171e+00 9.52545881e-01
3.82624775e-01 3.95625830e-01 4.99116004e-01 -2.53432412e-02
6.41149640e-01 2.79685318e-01 -2.19262674e-01 8.31640005e-01
-2.12206706e-01 -1.40779525e-01 -1.34417236e+00 3.51462692e-01
-1.82631779e+00 -6.40972674e-01 -2.33283505e-01 2.39669299e+00
7.23239183e-01 2.85908747e-02 3.66870165e-01 6.94336295e-01
6.76491261e-01 -5.47464430e-01 -3.53246272e-01 -1.28285134e+00
3.53060901e-01 4.05108780e-01 4.76430237e-01 3.42223883e-01
-1.00099409e+00 2.85478801e-01 5.34443092e+00 2.02500179e-01
-1.17895329e+00 -6.26147166e-02 8.98593068e-01 1.96213163e-02
2.46084794e-01 -1.97534170e-02 -7.31433690e-01 8.68902326e-01
1.08384895e+00 8.20334628e-02 4.86533999e-01 7.60047615e-01
5.11198282e-01 -4.93581444e-01 -7.48951197e-01 1.15399623e+00
-1.48741663e-01 -8.60403061e-01 -1.03707299e-01 -3.41641724e-01
8.59698236e-01 -2.46051416e-01 -2.24433362e-01 9.78882834e-02
7.02257082e-02 -9.72060084e-01 2.47106962e-02 5.83624840e-01
2.02780202e-01 -1.05695438e+00 1.31545317e+00 3.79500031e-01
-7.56408334e-01 -7.64539421e-01 5.12018576e-02 6.19447306e-02
3.65036041e-01 4.82563049e-01 -1.03779423e+00 7.05672026e-01
1.23828018e+00 -1.40054971e-01 -3.30293238e-01 1.56774974e+00
3.19394886e-01 5.48490822e-01 -5.19186020e-01 -2.57420361e-01
2.95596540e-01 -3.27288806e-01 3.40470880e-01 1.04557037e+00
5.74145913e-01 4.83568132e-01 2.17451587e-01 -8.93852711e-02
7.28449583e-01 6.32996202e-01 -1.74369231e-01 4.16220397e-01
5.13390660e-01 1.08406663e+00 -7.31966078e-01 -3.44770432e-01
1.61575258e-01 4.48994100e-01 -2.01963663e-01 -4.08761986e-02
-6.99000716e-01 -4.29371476e-01 9.05023962e-02 1.98351249e-01
1.45159259e-01 2.30956569e-01 -5.10900855e-01 -3.32485676e-01
-1.02753140e-01 -1.31175649e+00 1.04568172e+00 -5.44092417e-01
-1.16004825e+00 8.12087119e-01 3.56457770e-01 -1.17462838e+00
-3.32842618e-01 -4.33498442e-01 -4.89172786e-01 9.24736083e-01
-1.17309654e+00 -6.71631455e-01 -8.05745006e-01 6.43620729e-01
3.18173140e-01 -2.41757348e-01 9.12801564e-01 3.63242239e-01
-4.67390060e-01 3.97012085e-01 2.28897050e-01 -5.95186949e-01
4.97156173e-01 -8.90970230e-01 -4.24736410e-01 2.64376402e-01
-7.59099543e-01 5.99652290e-01 8.57272387e-01 -6.25580609e-01
-1.08037043e+00 -6.70731902e-01 8.74324977e-01 -7.74008334e-02
-8.58903825e-02 3.09001744e-01 -7.02104688e-01 2.47696653e-01
8.30489397e-02 -3.36041808e-01 9.18365121e-01 -8.60888660e-02
5.17711699e-01 -4.51464713e-01 -1.68481433e+00 2.25088820e-01
5.15888155e-01 4.20074373e-01 -6.00667775e-01 4.26353842e-01
7.85467122e-03 -5.42911530e-01 -1.18736005e+00 8.23350310e-01
4.50648457e-01 -1.26215470e+00 1.00693190e+00 -6.97067440e-01
-7.07274005e-02 -1.10075302e-01 3.35581303e-02 -1.35027850e+00
-2.38511339e-01 -4.36707199e-01 2.45621219e-01 9.31896389e-01
1.86920911e-01 -1.09391773e+00 4.56633061e-01 7.04861760e-01
-1.39227271e-01 -1.33803558e+00 -8.54246438e-01 -3.71871322e-01
-2.91205734e-01 2.10387960e-01 9.99831378e-01 8.95685792e-01
-1.57386184e-01 3.59819867e-02 -6.95874020e-02 -8.40141401e-02
2.82422185e-01 2.29039803e-01 3.52017164e-01 -1.25566673e+00
-2.01906696e-01 -1.41483665e-01 -2.70191967e-01 4.68772650e-02
-7.25523114e-01 -4.75699067e-01 -1.98780715e-01 -1.86226237e+00
9.69662070e-02 -1.00151098e+00 -6.71606898e-01 3.09718758e-01
-3.41395617e-01 -3.79290491e-01 -1.36919886e-01 9.29491818e-02
-2.42556203e-02 2.13736713e-01 1.05350685e+00 4.03772920e-01
-7.61799097e-01 4.38189745e-01 -4.34755355e-01 5.05012035e-01
1.29130530e+00 -6.78876460e-01 -4.12311494e-01 -4.70937490e-02
1.62727058e-01 2.11595640e-01 -1.23236790e-01 -9.49948251e-01
4.23892140e-02 -4.98354167e-01 3.98981273e-01 -7.56094873e-01
1.50469616e-01 -1.01145947e+00 3.85354847e-01 8.70316684e-01
-2.15305328e-01 9.08220232e-01 1.30619735e-01 2.06736967e-01
-3.02209347e-01 -3.97522777e-01 7.78338492e-01 -2.57005662e-01
-3.40310693e-01 -2.03074105e-02 -8.27943906e-02 -1.16168648e-01
1.38766813e+00 -4.59208310e-01 -4.02403660e-02 -7.35072345e-02
-6.52132094e-01 2.01957524e-01 -5.46687841e-02 1.96636304e-01
6.05454087e-01 -1.08705235e+00 -6.06263816e-01 1.55119091e-01
5.16332239e-02 1.78652219e-02 3.57512146e-01 1.30521059e+00
-9.89971459e-01 5.34532607e-01 -5.15491664e-01 -5.72516739e-01
-1.58981764e+00 5.97651184e-01 1.79210484e-01 -5.19267380e-01
-2.63718218e-01 7.18942702e-01 -4.80471849e-01 -2.71852136e-01
2.17749715e-01 -2.21433654e-01 -7.02390790e-01 -8.14400911e-02
3.12970847e-01 7.82988250e-01 5.57619870e-01 -6.31263256e-01
-6.66846931e-01 7.70626903e-01 1.61352500e-01 1.84212416e-01
1.54522562e+00 2.61261947e-02 -3.40597481e-01 9.74429995e-02
9.10211086e-01 -3.62902462e-01 -3.49475831e-01 3.66862625e-01
1.24044031e-01 -5.44471443e-01 5.83771430e-02 -1.16805196e+00
-6.60671234e-01 6.50700986e-01 1.23684943e+00 1.62312657e-01
1.74011016e+00 -4.83173519e-01 7.39711404e-01 2.98572212e-01
2.56100178e-01 -1.16602564e+00 -3.33308697e-01 2.66782373e-01
8.51859987e-01 -1.26462555e+00 -8.01653489e-02 -2.13482399e-02
-9.56202328e-01 9.36781347e-01 4.86891478e-01 -4.25563827e-02
7.62299538e-01 -2.96128783e-02 6.47382215e-02 -1.40873075e-01
-8.00105095e-01 1.54134318e-01 1.84360340e-01 3.69429737e-01
3.00054282e-01 1.53100029e-01 -1.05307436e+00 5.54956496e-01
-2.49934956e-01 1.88957915e-01 -5.88922650e-02 1.40130162e+00
-5.37009537e-01 -1.19002366e+00 -8.46246183e-01 8.44350100e-01
-6.78631842e-01 8.45880806e-02 -9.98982564e-02 5.68138361e-01
4.78613079e-01 1.17332602e+00 -2.66405553e-01 -4.99869227e-01
4.48993802e-01 1.44161075e-01 2.98466623e-01 -2.33065352e-01
-1.21163845e+00 -2.06734717e-01 1.51657104e-01 -2.00238228e-01
-3.63427669e-01 -7.04217792e-01 -1.34396708e+00 -2.06960827e-01
-3.25542897e-01 6.70986891e-01 1.13005185e+00 7.69665956e-01
4.51613665e-01 5.43817222e-01 8.77455950e-01 -3.85665357e-01
-5.83497703e-01 -9.09504116e-01 -2.18995109e-01 4.15345162e-01
2.31914282e-01 -9.74404752e-01 -4.86792982e-01 -2.01470897e-01] | [8.453534126281738, 4.792398452758789] |
964edbc9-66c4-4a11-9227-ff7c763152a3 | comparing-rule-based-and-deep-learning-models | 1703.08705 | null | http://arxiv.org/abs/1703.08705v1 | http://arxiv.org/pdf/1703.08705v1.pdf | Comparing Rule-Based and Deep Learning Models for Patient Phenotyping | Objective: We investigate whether deep learning techniques for natural
language processing (NLP) can be used efficiently for patient phenotyping.
Patient phenotyping is a classification task for determining whether a patient
has a medical condition, and is a crucial part of secondary analysis of
healthcare data. We assess the performance of deep learning algorithms and
compare them with classical NLP approaches.
Materials and Methods: We compare convolutional neural networks (CNNs),
n-gram models, and approaches based on cTAKES that extract pre-defined medical
concepts from clinical notes and use them to predict patient phenotypes. The
performance is tested on 10 different phenotyping tasks using 1,610 discharge
summaries extracted from the MIMIC-III database.
Results: CNNs outperform other phenotyping algorithms in all 10 tasks. The
average F1-score of our model is 76 (PPV of 83, and sensitivity of 71) with our
model having an F1-score up to 37 points higher than alternative approaches. We
additionally assess the interpretability of our model by presenting a method
that extracts the most salient phrases for a particular prediction.
Conclusion: We show that NLP methods based on deep learning improve the
performance of patient phenotyping. Our CNN-based algorithm automatically
learns the phrases associated with each patient phenotype. As such, it reduces
the annotation complexity for clinical domain experts, who are normally
required to develop task-specific annotation rules and identify relevant
phrases. Our method performs well in terms of both performance and
interpretability, which indicates that deep learning is an effective approach
to patient phenotyping based on clinicians' notes. | ['Leo Anthony Celi', 'John Foote Jr.', 'Franck Dernoncourt', 'David W. Grant', 'Yeran Li', 'Edward T. Moseley', 'Joy T. Wu', 'Jonathan Welt', 'Eric T. Carlson', 'Sebastian Gehrmann', 'Patrick D. Tyler'] | 2017-03-25 | null | null | null | null | ['patient-phenotyping'] | ['medical'] | [ 3.10097575e-01 4.36746091e-01 -2.94580072e-01 -4.40905720e-01
-9.37629759e-01 -4.48914737e-01 -1.01125814e-01 1.00174999e+00
-5.30228496e-01 8.28060448e-01 5.22227705e-01 -5.85292220e-01
-4.42227453e-01 -6.86058402e-01 -4.39470708e-01 -5.75334311e-01
-1.24564976e-01 1.15998423e+00 -5.75329602e-01 2.18273237e-01
-1.92068890e-01 5.02288938e-01 -1.01923597e+00 7.91448534e-01
8.24200451e-01 1.00793493e+00 -1.36393920e-01 7.87308395e-01
-8.61669779e-02 8.88683856e-01 -7.18233824e-01 -4.99583483e-01
-6.56086951e-02 -2.55188763e-01 -9.53045130e-01 -3.00700009e-01
7.23417699e-02 -1.87127814e-01 2.86143124e-01 8.82988393e-01
8.66411626e-01 -4.77912962e-01 5.98633051e-01 -7.49278426e-01
-3.58363211e-01 7.32044220e-01 5.61072491e-02 2.63425887e-01
5.42825282e-01 4.58654612e-01 1.24891293e+00 -4.93028194e-01
5.04136384e-01 8.58543634e-01 1.23596334e+00 5.05348325e-01
-1.32124984e+00 -5.72044313e-01 -2.68476486e-01 -3.20256054e-02
-1.22583604e+00 -1.39645010e-01 1.68355405e-01 -6.90582514e-01
1.22213352e+00 3.58734131e-01 7.46624410e-01 1.11667645e+00
3.81484687e-01 6.95949018e-01 5.43960154e-01 -2.58461982e-01
2.11968437e-01 -2.72086076e-02 3.72094840e-01 8.42001975e-01
5.28971910e-01 -8.92787129e-02 -2.55626351e-01 -8.06287527e-01
1.75410181e-01 9.08751264e-02 -3.23044151e-01 2.30517149e-01
-1.35527694e+00 8.26920033e-01 2.40621567e-02 3.49283785e-01
-7.60996461e-01 -2.07247540e-01 7.49406099e-01 9.99332741e-02
3.49916995e-01 1.08004642e+00 -1.03927207e+00 -2.28872135e-01
-9.80161250e-01 4.34713572e-01 1.07592523e+00 8.18286300e-01
7.21071362e-02 -3.31209838e-01 -7.58211732e-01 8.26608956e-01
-9.17165503e-02 3.37081432e-01 6.71998978e-01 -6.33718133e-01
3.71753246e-01 1.01934135e+00 -9.13279876e-03 -8.77112627e-01
-1.10863400e+00 -4.97266680e-01 -9.92154241e-01 -4.28679287e-01
3.09432715e-01 -6.30685329e-01 -8.59573066e-01 1.58943796e+00
-9.83581543e-02 -1.13464937e-01 3.00675243e-01 3.92149299e-01
1.09111416e+00 4.04244155e-01 6.69412971e-01 -1.30680516e-01
1.83065617e+00 -4.04323548e-01 -6.33515179e-01 1.25805512e-01
1.21920264e+00 -4.17143404e-01 8.23205650e-01 5.79299510e-01
-9.32016969e-01 -2.55774766e-01 -6.62637472e-01 -3.91159090e-04
-3.27616721e-01 4.58225906e-01 5.08709848e-01 6.90621614e-01
-9.99843895e-01 8.02062571e-01 -6.44241989e-01 -4.54220921e-01
9.47845578e-01 7.60510087e-01 -3.95782858e-01 1.09046839e-01
-1.13339674e+00 6.71447098e-01 8.20080221e-01 -1.64445743e-01
-5.40570855e-01 -1.27023041e+00 -8.14248443e-01 6.51004910e-01
3.66308587e-03 -1.32927346e+00 1.29021156e+00 -6.72356546e-01
-9.46600735e-01 1.04565287e+00 -2.02404648e-01 -8.77774656e-01
2.61152387e-01 -1.74873695e-01 -4.42642510e-01 3.35510254e-01
1.86890796e-01 7.07175374e-01 2.70316392e-01 -6.13341749e-01
-8.68508518e-01 -2.27562442e-01 -1.69803843e-01 -1.17977090e-01
-3.00582737e-01 1.63200185e-01 -7.81504437e-02 -5.32829583e-01
-5.12447596e-01 -8.34906697e-01 -4.26198423e-01 -2.16188908e-01
-7.80692160e-01 -4.04038191e-01 3.04412931e-01 -8.68216336e-01
1.32165551e+00 -2.10474563e+00 -2.63537198e-01 6.16925098e-02
8.59399319e-01 6.40031457e-01 4.95654764e-03 4.25649375e-01
-4.81021523e-01 6.05397880e-01 -4.93121892e-01 -2.44915396e-01
-1.45227522e-01 1.37398034e-01 -1.18646622e-01 2.34767459e-02
7.03424454e-01 1.23555982e+00 -9.65028048e-01 -4.45995927e-01
3.67162786e-02 4.12246704e-01 -8.08046401e-01 3.08942914e-01
-3.46552312e-01 5.39775074e-01 -3.75448763e-01 7.70407021e-01
2.20568329e-01 -7.41717517e-01 4.29284930e-01 5.22445142e-02
4.67709690e-01 4.31913257e-01 -3.73282462e-01 1.19741797e+00
-3.24659795e-01 4.33608711e-01 -3.34462851e-01 -9.90304768e-01
6.94371879e-01 7.97807395e-01 7.87699878e-01 -1.96390018e-01
1.60578400e-01 1.80531830e-01 3.53648633e-01 -1.05438256e+00
-1.12055331e-01 -3.56089234e-01 -1.20919742e-01 2.60699511e-01
-4.49746661e-02 2.30162308e-01 1.59484908e-01 -8.69279876e-02
1.63447988e+00 -4.62245047e-01 9.56727624e-01 -4.01545197e-01
5.41269898e-01 2.08026469e-01 8.74637008e-01 8.14259887e-01
-1.04956850e-01 5.09924293e-01 9.58196938e-01 -1.01892352e+00
-9.31528270e-01 -6.19028926e-01 -3.47743720e-01 5.49345851e-01
-8.00187349e-01 -5.30084252e-01 -5.83084583e-01 -8.21016252e-01
2.36020476e-01 5.92064440e-01 -8.56488585e-01 -4.66257557e-02
-4.82972831e-01 -1.19832385e+00 9.37630892e-01 7.61136115e-01
1.92338619e-02 -1.57434201e+00 -7.85301626e-01 5.32156229e-01
-3.60781461e-01 -1.21620572e+00 -8.03542063e-02 4.32559252e-01
-9.55464840e-01 -1.36126876e+00 -6.00606084e-01 -6.85604095e-01
6.12184882e-01 -8.45403969e-01 1.41598797e+00 2.02642679e-01
-4.99109745e-01 3.43462154e-02 -4.36577976e-01 -8.02808046e-01
-6.25467777e-01 4.31709826e-01 5.77041414e-03 -1.77350596e-01
1.16140568e+00 -4.89752591e-01 -6.21164382e-01 -3.01328272e-01
-7.96509445e-01 -3.44208889e-02 8.80598783e-01 1.01197994e+00
5.99601567e-01 -3.55636217e-02 6.92048311e-01 -1.39127958e+00
9.12076354e-01 -5.95316768e-01 -2.27894783e-01 3.27406637e-02
-7.16688931e-01 1.62002444e-01 1.01423573e+00 -5.11451103e-02
-4.82370377e-01 3.66045594e-01 -5.28410733e-01 -4.19039913e-02
-7.62973487e-01 6.44612908e-01 -4.36386243e-02 4.39789981e-01
6.79473758e-01 6.57446906e-02 -1.57390952e-01 -5.73245585e-01
-2.35827059e-01 7.34155655e-01 2.71392643e-01 -3.94326687e-01
1.99471802e-01 3.63798529e-01 1.15474716e-01 -5.71755052e-01
-1.11275005e+00 -6.34441495e-01 -6.10542297e-01 4.34667259e-01
1.27571988e+00 -6.76337302e-01 -1.25874674e+00 -5.50880991e-02
-1.22734749e+00 -1.83421731e-01 -3.75610560e-01 4.19974297e-01
-4.38975155e-01 3.29514563e-01 -7.80855119e-01 -3.64829600e-01
-9.45971668e-01 -1.08764350e+00 1.36176789e+00 -1.45030543e-01
-1.02213669e+00 -1.17344129e+00 1.53106332e-01 2.86449611e-01
1.39569342e-01 5.43992519e-01 1.75629890e+00 -1.41535890e+00
8.11691061e-02 -2.32831061e-01 -3.69965822e-01 1.40378565e-01
1.41113818e-01 -2.55834192e-01 -1.05312085e+00 -4.12189960e-02
-1.33804172e-01 -8.98096338e-02 8.33271205e-01 6.74369991e-01
1.70398295e+00 -4.83190149e-01 -4.91029233e-01 7.88042128e-01
1.33895969e+00 5.25881946e-01 4.39843625e-01 8.13659579e-02
6.39970303e-01 6.77379668e-01 1.96004897e-01 5.28717637e-01
2.05525637e-01 2.18591005e-01 1.03678055e-01 -2.62691528e-01
2.82481581e-01 -3.85148972e-02 -1.50296867e-01 4.67007965e-01
1.73146022e-03 -2.92479187e-01 -1.52928019e+00 7.21777976e-01
-1.75868368e+00 -5.50020099e-01 -1.31769001e-01 1.73405039e+00
1.00943792e+00 1.58404112e-02 6.96518868e-02 1.46664739e-01
4.39911097e-01 -3.88644397e-01 -4.40341622e-01 -6.37060583e-01
7.35246465e-02 6.70376718e-01 3.93451363e-01 1.03506245e-01
-1.11511469e+00 6.56278193e-01 6.81509686e+00 3.93383384e-01
-8.70508075e-01 -1.64359257e-01 1.03945625e+00 -3.87186930e-02
7.72200897e-02 -6.49802506e-01 -8.94477785e-01 5.53624451e-01
1.38543129e+00 2.24625878e-02 -1.35719568e-01 7.68965483e-01
4.87681836e-01 2.25846842e-01 -1.68850458e+00 1.03788626e+00
3.11687938e-03 -1.57950819e+00 3.48270506e-01 1.49696842e-01
5.81784844e-01 -1.61897317e-02 -7.13477805e-02 2.05453649e-01
3.79173040e-01 -1.47931194e+00 -4.17122580e-02 4.83041167e-01
9.47834373e-01 -6.12693906e-01 1.41513371e+00 1.21574096e-01
-6.88938141e-01 -3.11616421e-01 -2.03388602e-01 6.44898042e-02
1.03068434e-01 9.24160779e-01 -1.65865326e+00 5.01041591e-01
7.67363906e-01 7.49613583e-01 -4.45446342e-01 1.06284440e+00
-1.27807155e-01 8.65756691e-01 -9.58389714e-02 -8.39220062e-02
2.78366238e-01 3.00887585e-01 4.19338971e-01 1.65883601e+00
1.97529793e-01 3.65694940e-01 1.20510049e-01 8.39104712e-01
-3.21199179e-01 3.66257697e-01 -5.46679676e-01 -4.08256859e-01
1.21691652e-01 1.09147382e+00 -5.82051694e-01 -7.93063343e-01
-2.71422062e-02 5.35488904e-01 2.71210093e-02 1.34959653e-01
-5.55702448e-01 -3.65206391e-01 7.01439559e-01 1.37964934e-01
2.71530926e-01 5.77660501e-01 -8.12269747e-01 -8.04412544e-01
-7.29514211e-02 -1.17272139e+00 8.15450668e-01 -5.26235759e-01
-1.46877432e+00 7.73869753e-01 -4.34079945e-01 -1.22172678e+00
-3.68238062e-01 -9.23252344e-01 -2.65399843e-01 8.56168807e-01
-1.55372047e+00 -7.60365605e-01 -2.72526652e-01 3.33281994e-01
2.90052176e-01 -2.68949568e-01 1.44156086e+00 3.45700413e-01
-5.78574598e-01 5.21306336e-01 -2.00127736e-01 6.02844834e-01
6.52706325e-01 -1.47278464e+00 3.99979085e-01 2.62632012e-01
-1.95861578e-01 7.49412656e-01 3.95649135e-01 -7.36063600e-01
-6.99512720e-01 -1.46191847e+00 1.49355745e+00 -6.93522930e-01
3.52675378e-01 1.08245676e-02 -8.85693133e-01 6.70133233e-01
-6.01061620e-02 -3.07051599e-01 1.57473016e+00 1.90537095e-01
-1.23171523e-01 1.13860771e-01 -1.27250922e+00 3.49019825e-01
6.90861881e-01 -3.74845386e-01 -7.88621247e-01 6.85606241e-01
8.24344516e-01 -2.09504336e-01 -1.23533726e+00 6.82927728e-01
5.77759743e-01 -6.64812207e-01 7.89158642e-01 -1.28962958e+00
8.20879400e-01 1.06944166e-01 2.76126266e-01 -1.19930267e+00
-4.69549716e-01 -4.52247173e-01 7.97073320e-02 4.78121430e-01
9.85319316e-01 -7.24570513e-01 8.44179392e-01 6.55251026e-01
-1.86503641e-02 -1.15876198e+00 -4.85673845e-01 -2.71915495e-01
6.04666583e-03 -3.78993422e-01 8.87624383e-01 1.13548589e+00
2.33426481e-01 4.12671089e-01 -5.13453335e-02 2.82886088e-01
9.91513804e-02 -2.80916430e-02 3.01694065e-01 -1.66918576e+00
-4.38118309e-01 -6.24409854e-01 -4.71957117e-01 -4.12534058e-01
1.27148241e-01 -1.00555456e+00 -1.29360154e-01 -1.65059507e+00
4.03654784e-01 -3.43415946e-01 -4.75960314e-01 1.03762317e+00
-4.11753714e-01 -9.06705100e-04 -2.29829341e-01 1.00372195e-01
-2.55762875e-01 -1.86092332e-01 7.24696338e-01 -2.77498662e-01
-3.41323227e-01 5.08460365e-02 -1.02619588e+00 8.16364288e-01
1.03426516e+00 -7.45809972e-01 -4.66126166e-02 -3.33465725e-01
4.42178220e-01 1.59990266e-01 2.38272160e-01 -8.13880742e-01
-4.25235778e-02 2.08898872e-01 5.14326930e-01 -4.60649431e-01
-6.18818635e-03 -7.59917796e-01 -2.04504561e-02 9.85770047e-01
-6.09124780e-01 2.79610485e-01 4.89189744e-01 4.05938596e-01
-2.40555853e-01 -1.41948789e-01 4.85037625e-01 -2.65121967e-01
-2.34610945e-01 3.00556123e-01 -5.78287899e-01 1.30379573e-01
8.26270640e-01 6.05418682e-02 -1.35523647e-01 -2.13669255e-01
-1.09447610e+00 2.38692120e-01 -1.17077613e-02 3.32986005e-02
4.99774665e-01 -8.62223327e-01 -9.10556197e-01 1.07345290e-01
4.35908914e-01 8.93324092e-02 1.17341243e-01 1.03279626e+00
-1.12170434e+00 1.02806175e+00 -1.13320939e-01 -6.56621575e-01
-1.38667226e+00 6.25176907e-01 3.76289845e-01 -8.59399498e-01
-9.37736630e-01 7.96196282e-01 3.17090809e-01 -3.94718438e-01
2.52348691e-01 -1.08763385e+00 -5.59826612e-01 1.99683886e-02
8.37678730e-01 -1.18213706e-01 3.79733890e-01 -2.15862051e-01
-4.69556183e-01 2.66129553e-01 -1.81242809e-01 4.50315237e-01
1.53555012e+00 6.26421332e-01 -2.63420939e-01 1.26864702e-01
1.07210827e+00 -2.03679308e-01 -3.44770193e-01 2.55186926e-03
3.44075501e-01 9.71110612e-02 -2.97249287e-01 -1.17774737e+00
-8.00130785e-01 8.64761829e-01 3.45713168e-01 2.42491513e-01
1.21544433e+00 -9.30349529e-02 9.65256810e-01 6.31395578e-01
-1.74445599e-01 -6.35791898e-01 -4.44822282e-01 3.74153733e-01
5.48651695e-01 -1.16885662e+00 -2.11392149e-01 -3.83149326e-01
-5.89013040e-01 1.13349891e+00 1.28458709e-01 6.85574710e-02
6.46984220e-01 3.48817050e-01 8.14036578e-02 -6.63232088e-01
-8.36695790e-01 -2.65835263e-02 2.31850237e-01 5.57231724e-01
5.85983336e-01 4.68343735e-01 -3.66191357e-01 9.79591131e-01
-4.09582973e-01 2.60901541e-01 2.53497869e-01 5.11637926e-01
-7.96871930e-02 -1.06270528e+00 -1.03980288e-01 1.04537964e+00
-1.03565121e+00 -6.65582418e-01 -5.91423631e-01 5.51697493e-01
4.73624557e-01 6.59852147e-01 -1.12386756e-01 -2.99954712e-01
3.01617891e-01 5.74368417e-01 -9.87730548e-02 -1.18792510e+00
-1.07952857e+00 -3.07950050e-01 4.16896760e-01 -5.00269771e-01
-3.63980591e-01 -5.11390567e-01 -1.14144504e+00 3.21456939e-02
2.14440972e-01 1.49996236e-01 9.87441018e-02 1.01562262e+00
8.21268678e-01 7.99122691e-01 -5.95690534e-02 -3.05782631e-02
-3.24184865e-01 -8.87399077e-01 -2.72228211e-01 4.05271292e-01
4.32395220e-01 -1.86938047e-01 -2.33541895e-02 2.68530667e-01] | [8.062433242797852, 7.078860759735107] |
8c998420-1431-4fd4-a8ec-b4ccf6b38a48 | image-shape-manipulation-from-a-single | 2109.06151 | null | https://arxiv.org/abs/2109.06151v3 | https://arxiv.org/pdf/2109.06151v3.pdf | Image Shape Manipulation from a Single Augmented Training Sample | In this paper, we present DeepSIM, a generative model for conditional image manipulation based on a single image. We find that extensive augmentation is key for enabling single image training, and incorporate the use of thin-plate-spline (TPS) as an effective augmentation. Our network learns to map between a primitive representation of the image to the image itself. The choice of a primitive representation has an impact on the ease and expressiveness of the manipulations and can be automatic (e.g. edges), manual (e.g. segmentation) or hybrid such as edges on top of segmentations. At manipulation time, our generator allows for making complex image changes by modifying the primitive input representation and mapping it through the network. Our method is shown to achieve remarkable performance on image manipulation tasks. | ['Yedid Hoshen', 'Nir Zabari', 'Eliahu Horwitz', 'Yael Vinker'] | 2021-09-13 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Vinker_Image_Shape_Manipulation_From_a_Single_Augmented_Training_Sample_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Vinker_Image_Shape_Manipulation_From_a_Single_Augmented_Training_Sample_ICCV_2021_paper.pdf | iccv-2021-1 | ['sketch-to-image-translation'] | ['computer-vision'] | [ 6.95301712e-01 4.39012825e-01 -1.89183224e-02 -2.07753018e-01
-3.82277429e-01 -8.75714362e-01 9.91031945e-01 -3.85965407e-01
-3.88977349e-01 4.54809934e-01 -5.94461784e-02 -4.93888229e-01
2.86178827e-01 -9.38352644e-01 -1.16982365e+00 -5.47447264e-01
-1.87182648e-03 4.54360098e-01 2.51276344e-01 -3.58290911e-01
2.43889257e-01 7.74153531e-01 -1.36383915e+00 1.28098875e-01
5.93354702e-01 7.42696524e-01 4.06420261e-01 9.50238287e-01
-1.23696774e-01 2.24152192e-01 -7.79075027e-01 -3.31765860e-01
6.75358534e-01 -2.88641542e-01 -7.11908102e-01 5.44676900e-01
3.02303880e-01 -4.60140586e-01 -2.21192703e-01 9.29507256e-01
1.86203584e-01 9.33459625e-02 7.63702691e-01 -1.44913435e+00
-6.83458507e-01 5.82441330e-01 -4.83304054e-01 -2.74196327e-01
1.61424354e-01 5.82263410e-01 6.00400805e-01 -8.81109536e-01
8.95810127e-01 1.07511449e+00 4.85537082e-01 5.58185935e-01
-1.57561898e+00 -4.09023941e-01 2.64298394e-02 -5.93304157e-01
-1.14899886e+00 -5.51077724e-01 6.76672041e-01 -5.46334743e-01
7.90363014e-01 2.04792067e-01 8.05984855e-01 9.34127927e-01
2.07681626e-01 4.47722584e-01 1.09903610e+00 -6.74496174e-01
6.31251559e-02 -1.17498031e-02 -4.61976290e-01 8.16063285e-01
-1.82661284e-02 4.29572724e-02 -1.24682106e-01 -2.97199637e-02
1.63528156e+00 -2.75039285e-01 -2.55187392e-01 -4.51483011e-01
-1.14130533e+00 5.22146642e-01 5.81268191e-01 2.03523389e-03
-2.86394775e-01 7.42805302e-01 1.42687097e-01 1.80303186e-01
9.33742337e-03 7.38116682e-01 -2.75403887e-01 -1.50884837e-01
-9.45623994e-01 2.71334231e-01 5.75102508e-01 1.25423598e+00
8.78489673e-01 3.15336496e-01 -2.43647069e-01 4.94933665e-01
6.20356202e-02 1.67283610e-01 2.19415769e-01 -1.20673323e+00
3.15999418e-01 3.03885818e-01 1.66235849e-01 -5.88725209e-01
6.85289130e-02 -9.42332000e-02 -6.88580394e-01 8.04069102e-01
4.62562978e-01 -3.27390730e-01 -1.62480497e+00 1.72455275e+00
1.60212219e-01 7.77956545e-02 -3.08002084e-01 4.62677568e-01
5.82704246e-01 5.71955800e-01 2.33754709e-01 2.29600936e-01
1.12445879e+00 -9.26863015e-01 -4.17286336e-01 -2.55890667e-01
1.80505842e-01 -6.59698129e-01 1.22469079e+00 3.35994899e-01
-1.51807404e+00 -5.62256813e-01 -9.01457012e-01 -2.15206459e-01
-5.44404685e-01 1.86624274e-01 7.40351379e-01 5.12983859e-01
-1.48518479e+00 7.45082498e-01 -9.12846029e-01 -6.76454371e-03
4.53787476e-01 6.81724429e-01 -5.45133233e-01 2.37307563e-01
-8.17615986e-01 8.34476888e-01 6.14365518e-01 -1.03854254e-01
-9.81534779e-01 -6.76518977e-01 -1.06741059e+00 1.04607932e-01
2.65621215e-01 -1.08842289e+00 1.23607934e+00 -1.21021855e+00
-1.81052518e+00 9.59522247e-01 2.37488419e-01 -4.38816637e-01
7.42574513e-01 5.79747148e-02 3.31514120e-01 2.33485028e-01
-9.61205438e-02 1.43148613e+00 1.40857220e+00 -1.58821356e+00
-2.85632730e-01 1.52674362e-01 6.42597258e-01 2.34450147e-01
1.12135306e-01 3.39605077e-03 -6.48792386e-01 -8.25495183e-01
-8.53556991e-02 -1.19120407e+00 -4.20785010e-01 3.55829120e-01
-6.66749418e-01 1.62779003e-01 9.56745863e-01 -7.41182685e-01
8.20520163e-01 -2.13429713e+00 2.78337777e-01 4.57623839e-01
8.00217539e-02 2.13282764e-01 -1.77199915e-01 3.04928839e-01
-3.04886967e-01 5.86959004e-01 -4.45624739e-01 -4.61900860e-01
-1.12880155e-01 2.98554361e-01 -1.80300221e-01 1.53129369e-01
4.84324723e-01 1.24616814e+00 -6.06279790e-01 -4.01422828e-01
2.42285535e-01 6.41864121e-01 -7.57285416e-01 2.21325561e-01
-5.76944411e-01 7.53159225e-01 -1.84034914e-01 4.30513501e-01
4.85100359e-01 -1.87986642e-01 -1.82546005e-01 -2.82907218e-01
-2.22189993e-01 2.88669705e-01 -9.25771654e-01 1.77175963e+00
-5.72298825e-01 3.72675687e-01 2.85797715e-01 -5.36212564e-01
6.33696079e-01 3.05439115e-01 2.14406505e-01 4.77713346e-02
1.49140343e-01 -1.08699523e-01 1.93092689e-01 -1.58936575e-01
5.97937167e-01 -1.27804494e-02 -5.53644523e-02 6.16806686e-01
3.41731608e-02 -8.39375436e-01 3.64040643e-01 2.41242439e-01
7.99354851e-01 7.47270107e-01 3.65586042e-01 -1.78210601e-01
3.30513343e-02 3.85859385e-02 2.17806309e-01 7.83092260e-01
3.61005336e-01 8.97368252e-01 5.26630938e-01 -2.93492377e-02
-1.32227218e+00 -1.04743958e+00 3.06036249e-02 9.51931477e-01
4.51951399e-02 -2.57280588e-01 -8.66999686e-01 -5.57503819e-01
-8.64882469e-02 7.49506354e-01 -5.45708299e-01 -9.81035754e-02
-7.40069389e-01 -3.26924890e-01 4.90016758e-01 7.77380526e-01
6.30615592e-01 -1.21416044e+00 -6.75493538e-01 3.81423719e-02
1.80574268e-01 -1.04710269e+00 -6.49151087e-01 3.57722729e-01
-8.44293177e-01 -6.48009896e-01 -5.96245646e-01 -8.63358021e-01
1.23459566e+00 -2.67396943e-04 1.13991988e+00 2.33806819e-01
-3.94901305e-01 5.64283133e-01 -9.33601707e-02 -2.15721473e-01
-7.53820956e-01 -1.38067445e-02 -3.67611408e-01 -2.95305222e-01
-6.55568421e-01 -8.29907894e-01 -6.19884729e-01 5.03409058e-02
-1.31336987e+00 5.15287876e-01 6.25033319e-01 7.81979442e-01
6.74334586e-01 1.85185775e-01 1.10707188e-03 -9.37180638e-01
6.47826970e-01 -8.34419802e-02 -6.26004279e-01 4.97087017e-02
-1.71177819e-01 1.59132212e-01 4.85699922e-01 -6.17430270e-01
-1.15432227e+00 4.37585354e-01 -8.25134218e-02 -4.66674805e-01
-3.22000682e-01 4.45599347e-01 -8.67907107e-02 -4.43230450e-01
6.31751180e-01 8.48379359e-02 1.84816681e-02 -3.18606943e-01
8.22307944e-01 1.78030327e-01 8.05543125e-01 -8.32197666e-01
1.14375710e+00 3.39600682e-01 1.15251362e-01 -5.72375238e-01
-1.52965337e-01 2.00976774e-01 -1.02821803e+00 -4.47842516e-02
9.94659066e-01 -6.31966889e-01 -3.82373512e-01 5.56805789e-01
-1.13792121e+00 -9.99866009e-01 -4.11767960e-01 -1.45660281e-01
-6.34984910e-01 2.47903522e-02 -7.75592923e-01 -3.65677893e-01
-1.59526795e-01 -1.43920279e+00 1.20513451e+00 3.42693925e-01
-3.05164516e-01 -9.21786547e-01 -4.40186560e-01 2.70331334e-02
5.84744275e-01 6.33186996e-01 9.93285477e-01 -1.23945899e-01
-1.08191419e+00 -2.57396132e-01 -4.48169000e-02 2.94119626e-01
2.93850124e-01 5.39542496e-01 -7.71747887e-01 -1.81221634e-01
-3.21621358e-01 -2.95758218e-01 6.96992338e-01 2.50193954e-01
1.40529573e+00 -4.44178641e-01 -3.89066964e-01 8.60014915e-01
1.36018324e+00 2.43064642e-01 1.09107077e+00 4.03980702e-01
7.43849695e-01 5.67883961e-02 1.78722978e-01 2.15283215e-01
1.71341658e-01 4.89153415e-01 4.96815771e-01 -2.91243106e-01
-3.44974786e-01 -3.30088854e-01 2.56236911e-01 1.30375639e-01
-3.00858587e-01 -1.34016365e-01 -8.36407542e-01 1.99441895e-01
-1.29925764e+00 -7.44057834e-01 3.14799994e-01 2.15966773e+00
1.04428124e+00 4.10461128e-01 -3.27169187e-02 -6.23331740e-02
7.29457617e-01 -2.38515213e-02 -3.77943993e-01 -5.78258514e-01
1.64423123e-01 5.34332633e-01 6.86535716e-01 5.66768408e-01
-1.17858410e+00 1.13565910e+00 6.87018108e+00 5.52279413e-01
-1.26025367e+00 -2.15371236e-01 6.13853991e-01 2.16595531e-01
-3.87305766e-01 1.27345785e-01 -6.25528574e-01 3.13492358e-01
2.07684934e-01 -1.65260836e-01 7.51384974e-01 6.71277404e-01
2.58706231e-02 -2.27075472e-01 -1.16967797e+00 6.20420516e-01
-1.49183691e-01 -1.31771863e+00 4.00315613e-01 4.26308624e-02
8.72546673e-01 -2.96197891e-01 2.71907926e-01 8.71199220e-02
6.04323387e-01 -1.09696054e+00 8.45113516e-01 3.63247484e-01
1.11409318e+00 -5.63597798e-01 -2.61695553e-02 3.68126601e-01
-9.75893497e-01 3.33716452e-01 1.17771387e-01 1.88525289e-01
2.63955742e-01 9.38075110e-02 -1.03347063e+00 6.75832927e-02
3.48443985e-01 7.72464499e-02 -4.72809523e-01 8.98302972e-01
-6.42634630e-01 4.05368358e-01 -5.09354770e-01 4.94854629e-01
1.46228209e-01 -3.87000501e-01 6.00059569e-01 1.23082924e+00
2.14586273e-01 1.48819804e-01 3.51345092e-01 1.29132903e+00
-3.27730566e-01 -3.05059075e-01 -8.31860602e-01 -3.14373821e-01
3.72465879e-01 1.32571030e+00 -1.01530874e+00 -5.03352523e-01
-8.64378884e-02 1.15389812e+00 1.39615983e-01 5.09361207e-01
-8.58944297e-01 -4.78972226e-01 4.44152653e-01 3.57129663e-01
5.66541493e-01 -7.41706312e-01 -4.36254740e-01 -8.18698347e-01
-1.82029128e-01 -8.41161191e-01 -2.16821641e-01 -1.10670555e+00
-6.85140550e-01 3.86987418e-01 3.20634842e-01 -9.92781162e-01
-4.24900562e-01 -5.78724682e-01 -9.27815676e-01 9.89220560e-01
-1.05172372e+00 -1.40981543e+00 -2.57121563e-01 4.62541223e-01
3.66546720e-01 2.08871558e-01 6.93130553e-01 -1.34768933e-02
-2.63006270e-01 5.20171762e-01 -6.42601907e-01 2.96884239e-01
4.08743382e-01 -1.40912843e+00 8.09221566e-01 8.34430814e-01
1.04675137e-01 9.50550675e-01 5.84369242e-01 -7.13523626e-01
-1.29807448e+00 -8.93193662e-01 2.44631901e-01 -3.46715242e-01
3.93832177e-01 -4.24378633e-01 -7.69177198e-01 1.12176764e+00
4.54249918e-01 -1.58410564e-01 2.33720005e-01 -4.48423415e-01
-2.89942175e-01 3.66206944e-01 -1.33475542e+00 1.17730200e+00
1.09733510e+00 -5.22222757e-01 -3.40262711e-01 1.93753853e-01
7.62473702e-01 -9.45880413e-01 -8.38570535e-01 4.03100640e-01
3.90084118e-01 -7.03183055e-01 1.09995317e+00 -5.16590238e-01
5.74138999e-01 -4.35620487e-01 3.13517839e-01 -1.54386067e+00
-4.24628496e-01 -8.71129632e-01 9.16373879e-02 1.06956089e+00
4.86660361e-01 -5.02526820e-01 7.75102615e-01 1.06161964e+00
-2.67422438e-01 -6.38148725e-01 -3.15855294e-01 -5.95830381e-01
7.72646293e-02 -1.86808839e-01 5.91644406e-01 6.55569851e-01
-1.70132861e-01 2.80311294e-02 -1.26753867e-01 7.85181150e-02
3.00792485e-01 7.04475269e-02 1.12363827e+00 -7.91024387e-01
-5.77093720e-01 -6.59883201e-01 -3.40254486e-01 -1.25906217e+00
-8.99308082e-03 -9.20480967e-01 2.09095195e-01 -1.45564997e+00
-1.44244373e-01 -6.66649878e-01 3.06067377e-01 7.92961299e-01
-1.89543456e-01 3.41648668e-01 4.35731620e-01 3.13665755e-02
1.29762605e-01 2.71861732e-01 1.69808328e+00 -1.57993913e-01
-4.44527626e-01 -1.11972377e-01 -6.70180500e-01 7.55845308e-01
1.07183349e+00 -1.61928609e-01 -4.04669136e-01 -5.17624736e-01
8.53646770e-02 1.31163718e-02 4.27721709e-01 -8.26023042e-01
8.27615038e-02 -1.96630567e-01 4.71842915e-01 -6.02127649e-02
5.35931110e-01 -5.93727171e-01 3.76360148e-01 3.60207975e-01
-4.51418370e-01 2.01355606e-01 3.77853215e-01 2.60593355e-01
5.21060973e-02 -4.26344782e-01 8.51904929e-01 -5.14499187e-01
-5.44376791e-01 3.61890107e-01 -2.47482657e-01 -1.24654025e-01
9.04862046e-01 -2.70382196e-01 -9.47143957e-02 -5.61169207e-01
-1.03307343e+00 -1.07768700e-01 9.41645503e-01 2.22161710e-01
4.57337469e-01 -1.20493221e+00 -2.63738662e-01 2.52989680e-01
-3.33944768e-01 3.79305184e-01 -4.82301384e-01 4.16107714e-01
-8.75890195e-01 -1.49814576e-01 -4.25751418e-01 -4.76683527e-01
-1.24334216e+00 4.30972606e-01 3.08834195e-01 -8.68911818e-02
-6.30646110e-01 7.98301458e-01 3.19031924e-01 -1.25529945e-01
1.39924437e-01 -4.59147811e-01 2.28498116e-01 -4.26168561e-01
9.10108835e-02 -1.58797689e-02 -2.57693082e-01 -4.91718799e-01
1.31116673e-01 4.63626444e-01 2.41071209e-02 -4.49493706e-01
1.17263329e+00 8.82265493e-02 -3.15149933e-01 1.45575210e-01
9.68000889e-01 -9.07081645e-03 -1.66136312e+00 1.48616403e-01
-4.94530350e-01 -4.51143265e-01 -4.69538793e-02 -6.93633199e-01
-1.01310265e+00 7.42019057e-01 4.64918390e-02 2.05915943e-01
1.00825989e+00 -1.40864342e-01 6.84431136e-01 4.21756297e-01
5.26658654e-01 -9.14694905e-01 2.52646387e-01 4.08738911e-01
1.28937948e+00 -8.99814606e-01 1.78952310e-02 -7.61370897e-01
-5.80352068e-01 1.18880498e+00 6.44078791e-01 -3.64573181e-01
6.46571875e-01 8.43914568e-01 -1.18388250e-01 -1.86013401e-01
-4.04540539e-01 -1.23650007e-01 2.87794918e-01 6.15450501e-01
3.16626161e-01 6.73741698e-02 2.49509960e-02 -2.53748029e-01
-4.84594584e-01 6.75553232e-02 5.83087683e-01 1.10125422e+00
-2.63622791e-01 -1.23683524e+00 -3.42419207e-01 5.37398994e-01
-2.92049438e-01 -2.09917605e-01 -3.20118755e-01 9.76870477e-01
3.89971286e-01 4.33075726e-01 1.94836348e-01 -5.31346798e-02
1.80513307e-01 1.47518516e-01 1.02107763e+00 -9.20054674e-01
-7.31035292e-01 4.79410253e-02 9.65926622e-04 -3.46225858e-01
-2.70239949e-01 -5.61966777e-01 -1.27992105e+00 -4.91227061e-02
-1.75644591e-01 -3.87404621e-01 7.72714913e-01 7.49613523e-01
3.17208558e-01 4.04971272e-01 4.08466041e-01 -1.46099067e+00
-4.16603625e-01 -7.26584256e-01 -1.77669093e-01 4.48304802e-01
2.00583220e-01 -5.25349379e-01 -2.16162965e-01 7.21818984e-01] | [11.48619270324707, -0.41026636958122253] |
85fa9172-c98b-4057-a7b7-d88f3ce1c341 | change-detection-needs-change-information | 2304.12639 | null | https://arxiv.org/abs/2304.12639v1 | https://arxiv.org/pdf/2304.12639v1.pdf | Change detection needs change information: improving deep 3D point cloud change detection | Change detection is an important task to rapidly identify modified areas, in particular when multi-temporal data are concerned. In landscapes with complex geometry such as urban environment, vertical information turn out to be a very useful knowledge not only to highlight changes but also to classify them into different categories. In this paper, we focus on change segmentation directly using raw 3D point clouds (PCs), to avoid any loss of information due to rasterization processes. While deep learning has recently proved its effectiveness for this particular task by encoding the information through Siamese networks, we investigate here the idea of also using change information in early steps of deep networks. To do this, we first propose to provide the Siamese KPConv State-of-The-Art (SoTA) network with hand-crafted features and especially a change-related one. This improves the mean of Intersection over Union (IoU) over classes of change by 4.70\%. Considering that the major improvement was obtained thanks to the change-related feature, we propose three new architectures to address 3D PCs change segmentation: OneConvFusion, Triplet KPConv, and Encoder Fusion SiamKPConv. All the three networks take into account change information in early steps and outperform SoTA methods. In particular, the last network, entitled Encoder Fusion SiamKPConv, overtakes SoTA with more than 5% of mean of IoU over classes of change emphasizing the value of having the network focus on change information for change detection task. | ['Sébastien Lefèvre', 'Thomas Corpetti', 'Iris de Gélis'] | 2023-04-25 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 1.90322503e-01 -1.77930892e-01 1.83628544e-01 -3.01919878e-01
-2.99398333e-01 -6.02905631e-01 1.07674611e+00 4.70190406e-01
-8.66595924e-01 6.78650081e-01 -5.98034337e-02 -2.43595153e-01
-3.53761226e-01 -1.17431915e+00 -9.00501907e-01 -6.43929183e-01
-3.70645911e-01 4.08016235e-01 6.22124970e-01 -6.03416264e-01
5.87329753e-02 1.16431606e+00 -1.79853487e+00 8.12996104e-02
1.04985464e+00 1.02566707e+00 1.43160775e-01 4.91950631e-01
-2.89059311e-01 1.71900287e-01 -3.31082582e-01 -1.74269751e-01
4.24167067e-01 7.51197431e-03 -6.74377263e-01 -2.62835026e-01
6.87418520e-01 -1.49153024e-01 -1.92239713e-02 1.03326666e+00
4.67443913e-01 1.02931634e-01 6.42456830e-01 -9.34819818e-01
7.06892237e-02 5.26353538e-01 -6.69994473e-01 6.33593261e-01
-3.97418030e-02 2.28063598e-01 9.42337930e-01 -6.94357395e-01
9.19095278e-01 1.03142083e+00 8.62996221e-01 -2.55288005e-01
-1.28807986e+00 -4.00933474e-01 5.87820888e-01 5.19420743e-01
-1.33235621e+00 -1.72550559e-01 9.36273694e-01 -6.30464017e-01
1.08214474e+00 4.00499344e-01 9.81186390e-01 7.60755539e-01
1.37521282e-01 8.26463938e-01 9.70403612e-01 -8.53979960e-02
2.13807255e-01 -6.45559505e-02 4.81221452e-02 3.02764207e-01
2.65240788e-01 -7.91823142e-04 4.94946241e-02 3.95237833e-01
6.23350978e-01 1.28048196e-01 -4.40984845e-01 -5.37636638e-01
-1.14308417e+00 6.48670077e-01 1.09959817e+00 8.77016962e-01
-5.46595097e-01 2.68734962e-01 3.46445084e-01 3.45101148e-01
6.33173823e-01 4.80195194e-01 -5.70281506e-01 -2.27397814e-01
-1.19039702e+00 2.32970491e-01 3.23795706e-01 4.20352519e-01
1.07184601e+00 -4.94378619e-03 -7.48829991e-02 6.55110240e-01
-1.61741763e-01 7.25423574e-01 2.44221464e-01 -5.16245425e-01
5.69472790e-01 8.91419470e-01 -6.93174032e-03 -1.40182948e+00
-8.21855307e-01 -8.79877746e-01 -1.14021635e+00 4.50845867e-01
4.08850372e-01 9.91680175e-02 -1.07000279e+00 1.57395554e+00
3.96258622e-01 -1.49600640e-01 -2.24151641e-01 7.97206342e-01
5.85149229e-01 6.81196034e-01 -3.19126070e-01 5.81107549e-02
1.13079107e+00 -4.81813431e-01 -3.47274125e-01 -6.52131513e-02
5.54270566e-01 -3.07681501e-01 9.15532351e-01 2.55525708e-01
-6.77554309e-01 -7.24160910e-01 -1.01044679e+00 2.39501148e-01
-9.04105604e-01 1.06095195e-01 5.76861680e-01 1.95162997e-01
-1.24866939e+00 1.05214417e+00 -9.32172775e-01 -6.73215985e-01
5.90338469e-01 3.97101372e-01 -4.64209169e-01 4.50142063e-02
-1.28611588e+00 7.81801522e-01 4.24278170e-01 3.40479970e-01
-5.30496120e-01 -7.11414814e-01 -6.31286800e-01 2.62180001e-01
4.07078683e-01 -4.68211949e-01 6.84214830e-01 -1.13313520e+00
-1.22906518e+00 6.39469743e-01 -6.87279273e-03 -7.66412497e-01
9.48084831e-01 -1.28490239e-01 -4.55543578e-01 2.76753843e-01
3.02025843e-02 9.16209579e-01 8.30912650e-01 -1.26664686e+00
-1.05776858e+00 -3.09386134e-01 3.47573638e-01 9.32952687e-02
-2.26226211e-01 -5.95223665e-01 -3.88144642e-01 -5.26269734e-01
2.36935988e-01 -8.76898646e-01 -2.06429407e-01 8.28643814e-02
-2.34005675e-01 -2.31932342e-01 7.84124315e-01 -5.51652968e-01
1.16042972e+00 -2.11158729e+00 7.13904724e-02 3.66642118e-01
1.77035451e-01 5.15679359e-01 -1.22625522e-01 3.39583665e-01
-2.54516155e-01 2.26421729e-01 -7.86183357e-01 -1.28168166e-01
-6.35671755e-03 1.53140798e-01 -5.96308596e-02 4.80606973e-01
5.17941177e-01 7.64015079e-01 -7.82227635e-01 -1.28103912e-01
5.27183652e-01 4.69065756e-01 -4.64250654e-01 -3.84810090e-01
-1.96971744e-01 3.45120847e-01 -1.46182671e-01 3.58867854e-01
1.08735847e+00 1.72252178e-01 -3.21053684e-01 -2.08143771e-01
-6.77685082e-01 -1.68195531e-01 -1.44075656e+00 1.53527391e+00
-3.17668676e-01 8.70587707e-01 -5.36167175e-02 -8.61620843e-01
9.89369333e-01 -1.72805458e-01 7.17578053e-01 -8.60214353e-01
-2.09107213e-02 3.00631702e-01 1.05631929e-02 -3.07695508e-01
5.66666663e-01 1.27577096e-01 1.25358418e-01 -1.37718081e-01
-2.83323079e-01 -2.68612564e-01 3.94256979e-01 -5.69083579e-02
1.07148051e+00 1.23024374e-01 2.09714696e-01 -3.86295885e-01
8.01823139e-01 1.41065091e-01 3.52525145e-01 6.99339807e-01
-1.76767290e-01 6.53449118e-01 5.27544498e-01 -6.76547289e-01
-7.50012338e-01 -8.85504007e-01 -2.84822285e-01 5.44418156e-01
1.29419699e-01 -6.41728118e-02 -3.87384027e-01 -7.25594759e-01
3.57442021e-01 6.58420026e-01 -7.26064324e-01 -1.79866944e-02
-7.10398972e-01 -7.65157044e-01 3.71441722e-01 3.98805201e-01
1.00220907e+00 -8.18866432e-01 -7.67353833e-01 3.33306700e-01
1.05836593e-01 -8.25662553e-01 2.83206794e-02 4.29929107e-01
-8.59575331e-01 -9.86613870e-01 -7.80465543e-01 -3.77431989e-01
3.99686545e-01 3.08201253e-01 8.29035223e-01 -5.00238478e-01
-7.06163347e-02 2.81108469e-01 -4.74033743e-01 -1.39427036e-01
-7.71089941e-02 4.88593698e-01 -2.22706035e-01 1.73937723e-01
1.79259494e-01 -9.19399798e-01 -5.49953878e-01 1.49459913e-01
-1.03003907e+00 -9.74189416e-02 7.47804821e-01 4.51678395e-01
6.03078902e-01 3.93766552e-01 1.20639399e-01 -6.11282885e-01
1.45819053e-01 -3.03201765e-01 -6.92118347e-01 -5.37595479e-03
-6.44661546e-01 6.05483763e-02 6.88887954e-01 -7.13605210e-02
-9.37925160e-01 1.91075549e-01 -4.91469145e-01 -3.09578896e-01
-4.64128822e-01 5.24088860e-01 -2.25791261e-01 -1.78365082e-01
6.16786659e-01 1.22583300e-01 -2.88365960e-01 -6.66072249e-01
3.86017978e-01 3.89360130e-01 4.23625618e-01 -5.55307744e-03
9.91561890e-01 8.77332985e-01 1.07209988e-01 -1.00298190e+00
-2.55630463e-01 -6.72453582e-01 -1.08970773e+00 -2.89010137e-01
7.81861126e-01 -7.79122412e-01 -5.00351667e-01 7.71386921e-01
-1.12656009e+00 -2.85253704e-01 -6.70448661e-01 2.87218630e-01
-3.99285764e-01 2.25221246e-01 -1.84802592e-01 -4.36391205e-01
-2.43572354e-01 -9.37150955e-01 9.63672519e-01 2.31396094e-01
2.32748374e-01 -9.38537002e-01 1.78094819e-01 -4.14478242e-01
6.30815804e-01 7.91735709e-01 8.34631383e-01 -3.38762313e-01
-6.49646759e-01 -1.48981899e-01 -3.12545806e-01 2.28250891e-01
2.73852140e-01 8.17373917e-02 -9.54607606e-01 -3.16038877e-01
-2.01601744e-01 4.69003618e-01 1.47821856e+00 4.90430832e-01
8.46064627e-01 8.62122923e-02 -4.32627261e-01 8.61329794e-01
1.72405362e+00 2.27435395e-01 6.09156370e-01 5.50475299e-01
7.31680095e-01 4.81274724e-01 4.32049870e-01 4.60969716e-01
4.25127029e-01 8.70809138e-01 9.11174715e-01 -4.02071804e-01
-3.16060662e-01 7.43506178e-02 2.63010353e-01 4.49642569e-01
-2.84233779e-01 -2.34323993e-01 -1.09499717e+00 7.00063765e-01
-1.81875026e+00 -9.77176845e-01 -6.06190026e-01 2.06155181e+00
4.94126558e-01 4.69875425e-01 5.85716078e-03 3.63366902e-01
6.19060636e-01 5.19145608e-01 -6.07249975e-01 -2.36359641e-01
-5.61949909e-01 1.67520180e-01 6.84358001e-01 4.45897818e-01
-1.47592866e+00 8.97539914e-01 4.44144773e+00 8.34256411e-01
-1.42487645e+00 8.36665556e-02 1.92766264e-01 -1.70533080e-02
-3.23953569e-01 -1.09683588e-01 -7.35572815e-01 4.20678526e-01
5.80412149e-01 2.90480614e-01 1.92825437e-01 5.29872775e-01
3.03294450e-01 -5.60386240e-01 -7.48753548e-01 9.06495810e-01
-7.67145455e-02 -1.16407371e+00 2.17946425e-01 -6.01059422e-02
7.73403883e-01 4.27283943e-01 -2.13344201e-01 2.96865612e-01
4.22302559e-02 -5.13862669e-01 8.12184393e-01 8.02049994e-01
6.40878141e-01 -7.52304852e-01 8.96799743e-01 7.98775256e-02
-1.48881400e+00 -1.21714473e-01 -2.66801268e-01 -1.06647991e-01
2.48503640e-01 1.05252349e+00 -7.93203294e-01 8.36258113e-01
8.16999972e-01 1.27050960e+00 -9.73578572e-01 1.32286525e+00
-1.93769380e-01 4.90042299e-01 -7.63659537e-01 1.74317747e-01
7.33455479e-01 -1.80575833e-01 9.01706874e-01 1.39214814e+00
4.56805944e-01 -3.13816488e-01 -9.28572044e-02 7.73581207e-01
1.41537726e-01 3.33174430e-02 -6.17768645e-01 2.05419078e-01
-5.56431571e-03 1.22016060e+00 -1.12132323e+00 -3.04427296e-01
-1.42540950e-02 1.15265715e+00 8.59523937e-02 2.61282623e-01
-7.29311883e-01 -6.86922371e-01 8.65159154e-01 2.67284244e-01
9.58519459e-01 -3.38938296e-01 -1.01952329e-01 -9.97817099e-01
1.48276806e-01 -3.00762087e-01 1.72124535e-01 -5.16061187e-01
-8.36398065e-01 6.19113684e-01 8.45362842e-02 -1.36274791e+00
-3.04196123e-02 -4.94195282e-01 -5.26320696e-01 6.89602911e-01
-2.10535145e+00 -1.06017804e+00 -5.76095760e-01 5.12153804e-01
5.00383079e-01 3.39498222e-01 2.62411237e-01 4.83437687e-01
-5.28602839e-01 2.11148873e-01 4.32566434e-01 -1.64856598e-01
5.80228627e-01 -1.46630776e+00 6.43582702e-01 1.11359859e+00
-1.90330911e-02 8.49223733e-02 6.15154088e-01 -6.90747619e-01
-8.85826349e-01 -1.28536129e+00 7.84732878e-01 -1.44488528e-01
5.22211790e-01 -1.98324680e-01 -9.57123876e-01 2.47926369e-01
-7.56774563e-03 -1.04876161e-01 -8.22476745e-02 3.62396017e-02
-1.23801865e-02 -6.53010845e-01 -1.01784742e+00 4.45941597e-01
1.33312881e+00 -2.66157448e-01 -3.39330494e-01 9.22488049e-02
7.30845273e-01 -2.67425239e-01 -8.96194935e-01 5.95290959e-01
2.33089328e-01 -1.23693240e+00 7.71354735e-01 -1.72791407e-01
2.83298463e-01 -6.54104769e-01 -1.33189643e-02 -1.53857040e+00
-5.96351862e-01 -1.41432002e-01 2.13551834e-01 1.26201153e+00
3.37239772e-01 -7.96910703e-01 5.67949355e-01 -1.45287350e-01
-5.12830257e-01 -5.32331705e-01 -1.18089628e+00 -8.59128475e-01
9.50311497e-02 -6.40280783e-01 8.29247177e-01 1.00344872e+00
-6.54732764e-01 4.81387116e-02 9.23338905e-03 3.32460731e-01
1.57452554e-01 2.69326657e-01 8.03760231e-01 -1.62782192e+00
1.37469426e-01 -9.08425808e-01 -7.84885287e-01 -8.22631359e-01
-2.66830802e-01 -1.03242910e+00 -1.57350883e-01 -1.79184294e+00
-2.98427373e-01 -4.70064640e-01 -2.81896859e-01 5.92259645e-01
-2.84767859e-02 1.45194024e-01 3.90888095e-01 2.53359616e-01
-2.56128252e-01 7.05383480e-01 1.12662137e+00 -4.19863939e-01
-5.42560816e-01 1.09104216e-01 -1.97072774e-01 6.33022308e-01
6.80113018e-01 -3.76912624e-01 -1.51150703e-01 -5.21491766e-01
3.71146768e-01 -5.48483849e-01 5.07490814e-01 -1.60959995e+00
1.91616580e-01 8.87817070e-02 3.92401695e-01 -1.01495218e+00
2.68220752e-01 -1.08637059e+00 4.58513439e-01 7.34006345e-01
1.41978294e-01 -9.32285041e-02 4.93362576e-01 4.62124705e-01
-3.33057314e-01 -3.18968333e-02 8.04649770e-01 -2.88346354e-02
-1.20118761e+00 2.48943210e-01 -4.23241407e-01 -2.88220674e-01
8.98407876e-01 -4.50838238e-01 -2.59811163e-01 -1.34385437e-01
-8.30437839e-01 2.66840428e-01 5.46722710e-01 3.46399665e-01
2.77656943e-01 -1.13201463e+00 -5.35411596e-01 3.23978096e-01
1.88969627e-01 2.00242490e-01 4.50039327e-01 1.05596244e+00
-6.15610242e-01 4.58565950e-01 -3.74530911e-01 -9.72419202e-01
-9.90840912e-01 3.49442124e-01 6.15605295e-01 -4.06342506e-01
-7.40612030e-01 7.85012662e-01 -8.87337327e-02 -4.57077682e-01
-9.15721729e-02 -1.03012550e+00 -5.38523138e-01 8.12064350e-01
2.17268944e-01 4.41109955e-01 4.78792280e-01 -5.86434186e-01
-5.61516523e-01 9.81890321e-01 1.99993744e-01 1.35034010e-01
1.64047325e+00 -9.88853350e-02 2.69026607e-02 4.61045206e-01
1.31943476e+00 -7.94861391e-02 -1.51252794e+00 -2.48299345e-01
7.04182684e-02 -3.00964057e-01 2.17874244e-01 -9.32898819e-01
-1.25059903e+00 9.54771578e-01 1.09481049e+00 2.08071083e-01
1.23188019e+00 -2.35845774e-01 5.70759177e-01 5.40180206e-01
4.22179222e-01 -1.03777814e+00 -3.57960969e-01 7.62416661e-01
9.72128868e-01 -1.26019132e+00 -5.91236958e-03 -1.96562141e-01
-3.59878600e-01 1.17094624e+00 2.77640611e-01 -1.79698780e-01
7.88438857e-01 -1.73426107e-01 -2.47194976e-01 -2.30048642e-01
-1.47056088e-01 -9.16814864e-01 1.99590668e-01 5.37369847e-01
-2.32392460e-01 2.06577644e-01 -2.22698092e-01 1.34532377e-01
-1.45830393e-01 -7.49147683e-02 3.28675359e-01 7.73304880e-01
-5.99583983e-01 -7.85375237e-01 -3.30018401e-01 5.67560434e-01
1.25056267e-01 9.73868836e-03 -3.43930334e-01 1.09671891e+00
6.49483323e-01 5.38686037e-01 4.90768284e-01 -4.24379379e-01
8.53715599e-01 -3.35211396e-01 1.48636788e-01 -1.40962660e-01
-8.55046570e-01 -1.41876370e-01 1.96624007e-02 -6.83261275e-01
-5.55634022e-01 -1.06284308e+00 -1.05234385e+00 -2.89381206e-01
-1.08196013e-01 -1.16960265e-01 8.14258575e-01 7.89073169e-01
4.52820331e-01 7.65571356e-01 6.92314446e-01 -1.17757118e+00
-3.02070267e-02 -9.56184745e-01 -5.70634365e-01 3.03922474e-01
5.54711640e-01 -8.54093492e-01 -4.87062454e-01 -3.12698156e-01] | [9.710628509521484, -1.600701093673706] |
bafd5637-8333-42b9-b1dc-9b80de3d68da | inverse-path-tracing-for-joint-material-and | 1903.07145 | null | http://arxiv.org/abs/1903.07145v1 | http://arxiv.org/pdf/1903.07145v1.pdf | Inverse Path Tracing for Joint Material and Lighting Estimation | Modern computer vision algorithms have brought significant advancement to 3D
geometry reconstruction. However, illumination and material reconstruction
remain less studied, with current approaches assuming very simplified models
for materials and illumination. We introduce Inverse Path Tracing, a novel
approach to jointly estimate the material properties of objects and light
sources in indoor scenes by using an invertible light transport simulation. We
assume a coarse geometry scan, along with corresponding images and camera
poses. The key contribution of this work is an accurate and simultaneous
retrieval of light sources and physically based material properties (e.g.,
diffuse reflectance, specular reflectance, roughness, etc.) for the purpose of
editing and re-rendering the scene under new conditions. To this end, we
introduce a novel optimization method using a differentiable Monte Carlo
renderer that computes derivatives with respect to the estimated unknown
illumination and material properties. This enables joint optimization for
physically correct light transport and material models using a tailored
stochastic gradient descent. | ['Matthias Nießner', 'Tzu-Mao Li', 'Dejan Azinović', 'Anton Kaplanyan'] | 2019-03-17 | null | null | null | null | ['lighting-estimation'] | ['computer-vision'] | [ 6.27840757e-01 -6.54314280e-01 8.01988542e-01 -3.89941752e-01
-4.45283502e-01 -4.66484487e-01 6.56253040e-01 1.05846375e-02
-1.88175544e-01 8.11500072e-01 -2.04455405e-01 1.94080211e-02
-2.14209035e-01 -8.84941876e-01 -6.39916003e-01 -8.21315825e-01
4.27099109e-01 6.78011894e-01 -7.37121701e-02 1.12557001e-02
3.43924791e-01 1.04418182e+00 -1.78167558e+00 -1.55924857e-01
8.89795542e-01 6.18032098e-01 2.60587662e-01 9.58205283e-01
-2.33255565e-01 4.42200363e-01 5.08146873e-03 -2.09710836e-01
2.87058383e-01 -2.56885946e-01 -4.43320036e-01 4.26046103e-01
5.35077512e-01 -5.71801782e-01 -5.97140454e-02 9.34623659e-01
3.42366427e-01 4.55490023e-01 8.70978236e-01 -7.01795995e-01
-3.82243067e-01 -3.13974977e-01 -6.75555348e-01 -4.38836843e-01
7.38472581e-01 2.00435266e-01 5.34786582e-01 -9.67486799e-01
5.28032124e-01 1.05100131e+00 6.65307581e-01 2.93275893e-01
-1.44832301e+00 -6.24327436e-02 2.72417217e-02 1.10213026e-01
-1.52777350e+00 -4.10075784e-01 1.12671912e+00 -4.51350480e-01
7.27182806e-01 4.06928033e-01 9.75642264e-01 6.07880950e-01
2.40756661e-01 8.87496024e-02 1.57480800e+00 -7.06225634e-01
3.97848934e-01 5.20538807e-01 -2.35877875e-02 8.43565941e-01
1.38060868e-01 2.13924751e-01 -4.99587208e-01 -2.53269881e-01
9.30000842e-01 5.99200875e-02 -3.64700645e-01 -5.70145071e-01
-1.13263798e+00 2.47976616e-01 1.64208189e-01 -1.86417639e-01
-7.52325237e-01 1.91976696e-01 -3.73553395e-01 3.90232056e-02
8.52574825e-01 1.57546267e-01 -2.78853565e-01 2.44754389e-01
-5.69409311e-01 2.57738739e-01 8.94944608e-01 6.94247067e-01
1.29059267e+00 2.00434029e-02 9.75905806e-02 7.51617908e-01
7.90831864e-01 1.30518591e+00 -6.26967490e-01 -1.34068751e+00
-5.03985360e-02 1.59417257e-01 6.36171699e-01 -9.09310341e-01
-1.71947017e-01 -3.85749727e-01 -4.63583559e-01 6.00084245e-01
3.67823094e-01 1.08665526e-01 -6.84468806e-01 1.34649253e+00
9.34372842e-01 4.19404715e-01 1.42802659e-03 8.25218320e-01
5.13612807e-01 6.53056085e-01 -2.94399798e-01 -3.42670053e-01
9.74335551e-01 -6.11263156e-01 -5.01223862e-01 8.49368423e-02
5.29398024e-02 -1.13708222e+00 7.94762611e-01 6.36882782e-01
-1.29219759e+00 -1.08921267e-01 -7.02818334e-01 -2.29891166e-01
-5.32820746e-02 1.96507171e-01 5.68293631e-01 6.57022595e-01
-8.35721433e-01 6.99534178e-01 -7.15560019e-01 -1.17362201e-01
1.15134805e-01 -3.86462696e-02 7.75834545e-02 -4.33634132e-01
-5.27336597e-01 7.90004611e-01 -4.78947639e-01 2.84726977e-01
-8.83517981e-01 -9.61663246e-01 -5.48519254e-01 -3.93513113e-01
2.23538563e-01 -1.11707747e+00 8.68005097e-01 -6.71254754e-01
-2.25254560e+00 9.42153871e-01 -3.25746477e-01 1.45414442e-01
6.43388987e-01 -3.54776591e-01 4.37631160e-02 9.00172666e-02
-1.85149550e-01 -5.07598510e-03 1.05230236e+00 -1.94568086e+00
-1.16945073e-01 -3.34724844e-01 4.05336395e-02 5.12043655e-01
3.57129663e-01 -2.54025817e-01 -3.74375850e-01 -3.14423703e-02
3.49095553e-01 -6.70242727e-01 -3.15212965e-01 6.44392073e-01
-4.17485863e-01 6.05526805e-01 2.88877934e-01 -7.78491318e-01
2.86176741e-01 -1.84900641e+00 1.87623993e-01 4.19673890e-01
-1.26113920e-02 -1.86741084e-01 -9.87143815e-02 4.73415285e-01
2.10041836e-01 -5.82981765e-01 -5.00111997e-01 -8.55491340e-01
-2.41896808e-01 1.09410077e-01 -2.59067059e-01 8.10019433e-01
-2.01287493e-01 5.34316063e-01 -9.45578754e-01 -1.99956909e-01
8.85933638e-01 1.29463756e+00 -4.43496794e-01 1.57312617e-01
-4.22729701e-01 1.08315241e+00 -5.17670512e-01 5.58703244e-01
1.19280648e+00 1.67136818e-01 -1.41483381e-01 -5.15933752e-01
-6.78756893e-01 -6.35560304e-02 -1.53384650e+00 1.70084119e+00
-1.15234327e+00 3.77659351e-01 7.24142313e-01 -3.67908001e-01
8.57347786e-01 9.29230303e-02 6.27029896e-01 -4.83871907e-01
2.79631317e-01 3.14478278e-01 -7.75661469e-01 -3.28527480e-01
4.45879042e-01 -3.90587628e-01 8.44007194e-01 6.07011139e-01
-6.69144511e-01 -1.14661837e+00 -4.22766328e-01 -3.62641141e-02
5.86505175e-01 7.10695148e-01 -7.75884762e-02 -1.61604553e-01
7.71754146e-01 1.80009287e-02 9.10235792e-02 5.88078082e-01
4.84761149e-01 5.77001989e-01 -3.82047445e-01 -3.44980896e-01
-1.09431791e+00 -1.21844673e+00 -2.23357394e-01 6.23528063e-01
2.76305079e-01 3.31138879e-01 -8.62302721e-01 1.68264925e-01
9.87309217e-02 1.06692851e+00 -2.21034542e-01 1.41623244e-01
-5.66097021e-01 -8.08553040e-01 -3.39032471e-01 -2.49957934e-01
5.15259027e-01 -5.97256899e-01 -6.52364492e-01 1.87128723e-01
-3.50564010e-02 -1.20107913e+00 2.93308403e-04 -4.48371768e-01
-8.77778888e-01 -1.11595118e+00 -6.88384831e-01 -1.07424095e-01
9.95676219e-01 6.15324974e-01 1.26831198e+00 1.35471389e-01
-7.81282961e-01 1.23261118e+00 -6.91633970e-02 -3.41345996e-01
-4.43837404e-01 -6.81209922e-01 -1.59001708e-01 6.86520815e-01
-3.97038698e-01 -6.90285921e-01 -7.43295610e-01 2.97644138e-01
-6.55669868e-01 6.35709167e-01 1.12287790e-01 1.26324892e-01
1.10348284e+00 2.04264577e-02 -4.33583170e-01 -9.03081298e-01
3.06893736e-01 -1.65205210e-01 -1.04862309e+00 2.78599709e-01
-5.15543342e-01 -1.67400375e-01 3.90466779e-01 -1.07412316e-01
-1.79298079e+00 5.44669293e-02 -7.68235028e-02 -2.30268270e-01
-3.23126256e-01 -9.60639492e-02 -1.66036218e-01 -5.74571967e-01
5.18861830e-01 2.38225520e-01 -3.72765839e-01 -6.82441890e-01
5.23424506e-01 1.73355177e-01 2.98498362e-01 -8.11143994e-01
9.39129531e-01 1.34096503e+00 7.32940495e-01 -1.29142344e+00
-8.02882731e-01 -4.40109611e-01 -5.76918185e-01 -5.13589680e-01
5.59333742e-01 -6.59423172e-01 -7.71814525e-01 7.38501966e-01
-1.16738319e+00 -4.92340863e-01 -5.45804620e-01 7.29603648e-01
-6.80610359e-01 4.99307245e-01 -2.44111151e-01 -1.12662482e+00
-2.28288502e-01 -1.01969326e+00 1.34177148e+00 3.89449671e-02
2.59749919e-01 -1.24392664e+00 3.30700725e-01 4.46527809e-01
3.95995975e-01 3.84135693e-01 6.13931596e-01 9.37948406e-01
-1.29185104e+00 1.86590292e-02 -4.39158976e-01 3.76542836e-01
3.24502438e-01 3.97554785e-01 -1.22225046e+00 -1.51435539e-01
3.82714331e-01 2.47941211e-01 5.91582298e-01 6.03899956e-01
8.94395769e-01 1.76651731e-01 -2.12692872e-01 1.10725284e+00
2.01447749e+00 -1.50696844e-01 3.94362748e-01 -2.08232906e-02
9.77772832e-01 7.99803376e-01 5.15348017e-01 7.60643601e-01
2.71172225e-01 7.31384337e-01 6.85899854e-01 -5.45070916e-02
-5.74074328e-01 2.18403205e-01 -1.51027907e-02 6.03353679e-01
-6.91582561e-01 -2.44923636e-01 -6.47579908e-01 2.03894511e-01
-1.17059910e+00 -6.59431875e-01 -7.85006106e-01 2.60071778e+00
5.81042767e-01 -4.88504916e-01 -6.10392570e-01 -1.38507470e-01
3.56097937e-01 -1.35599479e-01 -5.46728790e-01 -1.67677790e-01
-1.54458433e-01 3.60459685e-01 7.61635542e-01 1.22773778e+00
-4.33097869e-01 5.49868882e-01 6.04721165e+00 3.85710806e-01
-9.19637740e-01 1.96388707e-01 -1.42008914e-02 8.11088681e-02
-1.03819406e+00 2.69319922e-01 -6.49010837e-01 6.70803636e-02
4.41573173e-01 2.73933828e-01 1.14923525e+00 1.63737655e-01
6.13192320e-01 -5.74999332e-01 -8.66601944e-01 1.07996273e+00
1.30005538e-01 -9.97411370e-01 4.08382863e-02 -4.61707413e-02
1.05195498e+00 -2.89201178e-02 -2.95969122e-03 -5.95696092e-01
2.49293581e-01 -3.12630922e-01 7.21500576e-01 1.32994199e+00
7.21912622e-01 -3.98745477e-01 1.67705283e-01 2.91403055e-01
-7.82470703e-01 4.93045002e-01 -1.74440354e-01 7.97374099e-02
6.07271135e-01 1.17483366e+00 -5.23967445e-01 6.62503481e-01
3.39145213e-01 6.15776360e-01 -3.46963406e-02 1.03542793e+00
-4.77268040e-01 3.10161620e-01 -6.26763284e-01 2.37609163e-01
-3.59035462e-01 -1.02520978e+00 9.03512239e-01 8.55306208e-01
3.66161317e-01 1.49596378e-01 2.79935487e-02 1.27388489e+00
2.26664215e-01 2.03036487e-01 -3.48207146e-01 4.99860257e-01
-7.27728233e-02 1.35417736e+00 -7.49169469e-01 -2.57760976e-02
-3.16862941e-01 1.12156510e+00 -5.46058752e-02 8.71360719e-01
-6.23050272e-01 1.13821067e-01 5.58630645e-01 3.10983032e-01
-3.26583952e-01 -6.21973991e-01 -3.14001918e-01 -1.17503726e+00
1.06360791e-02 -7.77966008e-02 -4.89907980e-01 -1.15379655e+00
-1.18600190e+00 -1.01325482e-01 1.41213620e-02 -8.25511754e-01
2.89269179e-01 -6.78693533e-01 -4.04541820e-01 1.24075818e+00
-2.03058600e+00 -1.23142815e+00 -5.58916092e-01 7.24211335e-01
3.44173610e-01 4.74986345e-01 7.42375672e-01 2.12495327e-01
-1.94060177e-01 -3.46206814e-01 7.28706717e-01 -6.55939221e-01
3.90081853e-01 -1.07106543e+00 1.74008086e-01 5.88463128e-01
-1.73202798e-01 4.37535375e-01 9.08196270e-01 -5.94111145e-01
-1.87036228e+00 -8.46026599e-01 3.01485896e-01 -4.08490598e-01
2.71200895e-01 -1.49332121e-01 -6.28976583e-01 3.48101854e-01
-1.22185282e-01 5.38105592e-02 2.85423428e-01 -3.15132141e-01
8.03962201e-02 -2.61175632e-01 -1.39201117e+00 5.07546425e-01
1.08278322e+00 -6.20665193e-01 9.45901126e-02 7.07991242e-01
2.45411262e-01 -5.56398094e-01 -5.16607642e-01 2.35712171e-01
6.17247701e-01 -1.17269444e+00 1.42707181e+00 1.71747297e-01
-6.64497241e-02 -4.80761111e-01 -2.99434066e-01 -1.27905953e+00
-6.59556165e-02 -6.51539743e-01 -1.82085298e-02 1.06060004e+00
9.23518091e-03 -8.73850107e-01 6.41210616e-01 9.38980997e-01
-2.54579306e-01 -2.76058346e-01 -5.75516284e-01 -5.07372499e-01
-4.43312138e-01 -7.22237408e-01 5.01759768e-01 7.17663765e-01
-1.16536379e+00 -6.55555874e-02 -4.05816913e-01 5.70525825e-01
1.47037637e+00 4.60758120e-01 8.08379471e-01 -1.38523650e+00
-3.42622966e-01 -2.02226453e-02 3.51516932e-01 -9.68244255e-01
1.52984500e-01 -6.47267938e-01 1.73439190e-01 -1.76516449e+00
6.93767294e-02 -8.48462462e-01 2.95329630e-01 -3.52048576e-01
1.52568728e-01 2.64274806e-01 -3.13508421e-01 1.75139681e-01
-6.10655956e-02 6.34631991e-01 1.57707822e+00 3.98772024e-02
-4.67824548e-01 3.00732017e-01 -5.72285280e-02 9.21055257e-01
6.92289650e-01 -2.94411927e-01 -4.00022179e-01 -9.47561324e-01
7.33568132e-01 -1.70669779e-01 8.55524182e-01 -8.40837538e-01
-5.88314943e-02 -4.51487094e-01 1.56455100e-01 -5.59150577e-01
7.91808248e-01 -1.17386937e+00 6.98166132e-01 1.43703058e-01
-4.36892882e-02 -6.68613553e-01 9.96436551e-02 6.47203982e-01
4.51521188e-01 -4.92336303e-01 8.15031469e-01 -3.82882565e-01
-4.31020111e-01 3.91829699e-01 -1.97858065e-01 -3.06357175e-01
7.09048331e-01 -3.99239689e-01 6.49500787e-02 -1.83922023e-01
-5.93376994e-01 -3.41550887e-01 9.14347291e-01 -3.23797315e-01
7.97863483e-01 -1.06219411e+00 -7.86241889e-01 2.91515976e-01
-2.30978832e-01 7.65812993e-02 4.53751951e-01 7.31524944e-01
-1.17613685e+00 -7.89861307e-02 2.28135303e-01 -5.68518162e-01
-1.25361538e+00 1.53921142e-01 5.86331427e-01 2.04007477e-01
-6.15648091e-01 7.67691791e-01 1.90294415e-01 -5.83525360e-01
-3.20989430e-01 -2.37087354e-01 2.30235949e-01 -3.66661429e-01
2.78073221e-01 9.46048439e-01 1.92822590e-01 -7.74612188e-01
-1.84557706e-01 1.42564094e+00 5.20913422e-01 -2.06470221e-01
1.35030270e+00 -7.16569364e-01 -4.59855109e-01 5.48435569e-01
9.55400586e-01 5.28617918e-01 -1.41494119e+00 -3.45045120e-01
-9.58486617e-01 -9.64239359e-01 6.00141823e-01 -6.03476405e-01
-8.27419043e-01 8.35201025e-01 2.83368021e-01 -2.32618541e-01
1.00221264e+00 -2.35540912e-01 5.61398983e-01 3.62835437e-01
7.46579587e-01 -1.03516424e+00 -4.52910304e-01 3.71321976e-01
7.43588030e-01 -1.02078438e+00 5.65074503e-01 -9.75247085e-01
1.30603582e-01 1.06009936e+00 -9.70803052e-02 -6.29476160e-02
8.35379481e-01 6.77830502e-02 3.91687602e-02 -3.50265503e-01
-1.45513266e-01 -6.00885004e-02 3.01600605e-01 6.77886188e-01
2.04144359e-01 8.05287361e-02 9.02868286e-02 -7.26152003e-01
2.19708756e-01 -4.74516153e-02 3.35630864e-01 7.44714260e-01
-2.84792483e-01 -9.26464736e-01 -7.13175893e-01 9.40840170e-02
1.10591454e-02 -1.51725575e-01 -3.00804973e-02 1.24868169e-01
1.46254554e-01 8.18208575e-01 -8.09842497e-02 4.45722252e-01
4.19334531e-01 -3.62370461e-01 9.64066863e-01 -7.39459932e-01
-6.87214062e-02 1.16811149e-01 -5.45908622e-02 -5.46064258e-01
-8.19265664e-01 -9.17519987e-01 -1.03146374e+00 -1.54704198e-01
-4.51571405e-01 -1.62342072e-01 1.36871338e+00 6.85925424e-01
-2.53977384e-02 3.51355582e-01 7.66878188e-01 -1.23084474e+00
-3.93161736e-02 -1.98524460e-01 -8.33716571e-01 3.33729625e-01
4.75172877e-01 -7.91073263e-01 -5.54332137e-01 1.93066567e-01] | [9.736343383789062, -3.0645880699157715] |
f85fcf0e-bcf3-4d46-85ad-28f35e1250da | panoramic-image-reflection-removal | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Hong_Panoramic_Image_Reflection_Removal_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Hong_Panoramic_Image_Reflection_Removal_CVPR_2021_paper.pdf | Panoramic Image Reflection Removal | This paper studies the problem of panoramic image reflection removal, aiming at reliving the content ambiguity between reflection and transmission scenes. Although a partial view of the reflection scene is included in the panoramic image, it cannot be utilized directly due to its misalignment with the reflection-contaminated image. We propose a two-step approach to solve this problem, by first accomplishing geometric and photometric alignment for the reflection scene via a coarse-to-fine strategy, and then restoring the transmission scene via a recovery network. The proposed method is trained with a synthetic dataset and verified quantitatively with a real panoramic image dataset. The effectiveness of the proposed method is validated by the significant performance advantage over single image-based reflection removal methods and generalization capacity to limited-FoV scenarios captured by conventional camera or mobile phone users. | ['Boxin Shi', 'Alex C. Kot', 'Xudong Jiang', 'Lingran Zhao', 'Qian Zheng', 'Yuchen Hong'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['reflection-removal'] | ['computer-vision'] | [ 1.11995411e+00 -2.04042464e-01 4.27955300e-01 -1.08646020e-01
-6.38516128e-01 -2.76443332e-01 6.76742256e-01 -9.03585374e-01
-1.90139338e-01 5.08654058e-01 3.22745919e-01 -2.45344311e-01
-3.24073911e-01 -8.69416595e-01 -6.50527954e-01 -9.08952773e-01
6.57085001e-01 -1.05782673e-01 9.58145782e-03 -2.66489685e-01
2.77075022e-01 4.26898897e-01 -1.44825745e+00 -1.55854169e-02
8.28121305e-01 6.81353927e-01 5.88968337e-01 5.45356512e-01
3.19696218e-01 5.93039155e-01 -4.15619910e-01 -1.85182970e-02
7.81657457e-01 -5.88538706e-01 -3.52058709e-01 5.89235842e-01
8.10420394e-01 -7.94225514e-01 -5.18779337e-01 1.19244838e+00
2.38647640e-01 3.77856903e-02 3.17322195e-01 -7.57418275e-01
-3.43876094e-01 -4.69070449e-02 -8.21909130e-01 -1.86428219e-01
5.38593113e-01 -2.30047673e-01 4.58723873e-01 -9.43930805e-01
4.11732435e-01 8.67859721e-01 5.91377914e-01 7.78882951e-02
-7.86722064e-01 -5.03363431e-01 -4.94374394e-01 -7.89276212e-02
-1.25495315e+00 -5.76552689e-01 1.02627373e+00 -8.92247632e-02
4.24402177e-01 4.35426563e-01 6.07376397e-01 5.83775997e-01
3.56356502e-01 2.19801426e-01 1.49663997e+00 -8.00096810e-01
-2.74229705e-01 1.96173519e-01 -9.70128272e-03 5.41347444e-01
3.90748084e-01 4.06933218e-01 -2.11993679e-01 7.91882128e-02
1.05842924e+00 4.53660309e-01 -7.97446251e-01 -3.50007117e-01
-1.05388451e+00 1.67170838e-01 3.02978188e-01 1.29068062e-01
-5.40902257e-01 -1.71273574e-01 -3.47599208e-01 3.43433261e-01
3.42800289e-01 4.40563768e-01 1.83984727e-01 4.15886700e-01
-1.07156610e+00 -1.46874890e-01 4.99077767e-01 9.72259521e-01
7.60493398e-01 2.55179495e-01 3.37759078e-01 1.06158698e+00
3.53367060e-01 1.04275286e+00 1.37546703e-01 -8.28870654e-01
5.12301028e-01 2.57615924e-01 1.89023435e-01 -8.75438094e-01
-3.07533115e-01 -3.41645300e-01 -8.61414313e-01 3.90891850e-01
3.54308099e-01 -4.36512530e-02 -7.75942445e-01 1.35112917e+00
3.00813228e-01 3.97326946e-02 3.42972428e-01 1.14549637e+00
5.17480195e-01 8.22029829e-01 -8.08249891e-01 -4.95305181e-01
1.19074988e+00 -1.03153658e+00 -6.91435099e-01 -2.84822971e-01
-2.15087190e-01 -1.40200996e+00 7.75664568e-01 6.97474957e-01
-1.12984836e+00 -4.87041503e-01 -1.20858693e+00 6.59140870e-02
3.07570577e-01 3.47383320e-01 2.82247633e-01 8.46934855e-01
-8.29948366e-01 -7.06158252e-03 -3.55944961e-01 -5.29429913e-01
-1.60318509e-01 -7.09496764e-03 -4.22538698e-01 -6.72903001e-01
-8.09702098e-01 9.13195133e-01 1.28709748e-02 3.12147379e-01
-7.14075506e-01 -6.01600289e-01 -6.40123844e-01 1.00599714e-01
3.41490924e-01 -6.07764781e-01 8.53111506e-01 -1.10259795e+00
-1.68518376e+00 5.69519520e-01 -1.01732939e-01 -6.24382123e-02
5.17583847e-01 -5.44031024e-01 -7.82695711e-01 5.14409065e-01
-2.06753448e-01 -7.53920153e-02 1.24752617e+00 -1.63852096e+00
-4.40413356e-01 -2.85754502e-01 1.13988936e-01 7.02100217e-01
-4.80094776e-02 -1.46532238e-01 -6.37643456e-01 -4.46500778e-01
7.09025025e-01 -8.91364276e-01 -7.03289500e-03 -1.53861418e-01
-5.82549870e-01 1.00114584e+00 1.13858712e+00 -8.09625983e-01
7.06383049e-01 -1.97807372e+00 -4.23155457e-01 4.70801890e-01
-1.04229562e-01 3.56243372e-01 -3.33354771e-01 7.09148586e-01
-2.91024446e-01 -7.36546993e-01 -3.56952906e-01 -1.14509985e-01
-7.10630357e-01 -2.34012067e-01 -4.86877203e-01 8.82510126e-01
-3.94036263e-01 4.62977380e-01 -5.13622701e-01 -1.21292725e-01
5.77578366e-01 7.81529486e-01 -3.04583937e-01 3.28739107e-01
3.49647790e-01 5.55746913e-01 -3.11631709e-01 6.02830410e-01
1.30541718e+00 -2.92686429e-02 4.15108055e-01 -4.78489786e-01
-3.69363695e-01 1.00908697e-01 -1.19074619e+00 1.29775262e+00
-8.52208912e-01 6.64635658e-01 2.97966748e-01 -7.04996943e-01
1.12957680e+00 1.85508341e-01 6.62393749e-01 -1.19525921e+00
-4.50528041e-02 1.42183363e-01 -2.99790084e-01 -4.81211215e-01
9.29439127e-01 -3.72981250e-01 2.02007934e-01 8.08243513e-01
-3.69001508e-01 -4.22359914e-01 -3.51378232e-01 8.80656391e-03
6.54577196e-01 3.01822245e-01 3.91538352e-01 -1.57370433e-01
7.43678629e-01 -8.34794566e-02 2.39945248e-01 7.95079470e-01
4.00877893e-01 1.11909997e+00 -2.31333181e-01 -2.29455560e-01
-1.29916024e+00 -1.20644045e+00 -2.49974981e-01 3.88360053e-01
7.57951379e-01 -3.53525057e-02 -7.38078058e-01 -1.30505860e-01
-5.80746233e-01 6.31426811e-01 -1.33839965e-01 -9.81922522e-02
-7.20334053e-01 -8.67860854e-01 7.36769363e-02 -2.42094323e-01
1.04123092e+00 -7.01729953e-01 -6.86430335e-01 -1.42534360e-01
-5.48182726e-01 -1.22122359e+00 -2.06234649e-01 -4.11260754e-01
-7.89610505e-01 -1.51023638e+00 -7.47165561e-01 -5.41005015e-01
8.23404551e-01 1.42446017e+00 8.09201121e-01 1.25006124e-01
-2.39793807e-01 6.59029424e-01 -2.20394984e-01 3.80801335e-02
-4.04195487e-01 -6.69575155e-01 -1.25455722e-01 2.52104729e-01
-6.24356084e-02 -5.46695948e-01 -9.18368340e-01 6.69083953e-01
-1.08516753e+00 4.87940550e-01 8.37584078e-01 7.21224725e-01
2.76484877e-01 4.16795343e-01 6.19142838e-02 -9.72699344e-01
3.37403983e-01 -1.00482307e-01 -8.59596789e-01 3.03616434e-01
-6.59998596e-01 -5.37591934e-01 5.75773597e-01 1.07790537e-01
-1.88521779e+00 2.01178491e-02 5.30651249e-02 -1.15588494e-01
-3.79148349e-02 3.90897170e-02 -2.84041107e-01 -3.99524748e-01
4.73573893e-01 5.08920729e-01 2.23295078e-01 -4.38141346e-01
2.71691293e-01 6.55057132e-01 8.80596161e-01 -1.45688087e-01
1.17121053e+00 1.04257596e+00 9.04713124e-02 -1.34182799e+00
-6.87839150e-01 -6.83867097e-01 -5.75889170e-01 -3.46958429e-01
6.91447437e-01 -1.08741546e+00 -2.15537384e-01 8.22215438e-01
-9.13461685e-01 -9.70403925e-02 2.54555382e-02 8.59130740e-01
-4.58104968e-01 9.27564561e-01 -3.35947245e-01 -6.13704801e-01
-3.69415969e-01 -9.88516688e-01 8.87246251e-01 2.39286989e-01
4.57874447e-01 -7.11387753e-01 2.19601601e-01 6.88885570e-01
4.65959579e-01 -3.05381000e-01 6.88553870e-01 2.68884093e-01
-9.63344693e-01 -2.56602049e-01 -5.94425797e-01 5.04698217e-01
3.87317359e-01 -2.27381602e-01 -1.13102603e+00 -3.60035986e-01
5.94933510e-01 -3.54035832e-02 7.89924383e-01 4.65213984e-01
6.06366217e-01 -1.06328443e-01 -1.06803752e-01 8.75311494e-01
1.90708041e+00 1.41787738e-01 1.25090420e+00 4.85132068e-01
7.83309639e-01 4.95703727e-01 8.50176632e-01 1.69747442e-01
9.79969576e-02 7.08899677e-01 5.53561032e-01 -5.16006291e-01
-4.40611601e-01 -2.49911502e-01 2.42617294e-01 6.48395717e-01
-2.22567558e-01 -5.24419248e-01 -3.27670753e-01 2.28376284e-01
-1.45910168e+00 -1.02853251e+00 -3.46765578e-01 2.61293101e+00
1.26619235e-01 -3.31914812e-01 -4.52451199e-01 1.26097217e-01
7.07513452e-01 3.87133360e-01 -2.95252427e-02 -1.32516876e-01
-3.51970941e-01 -1.89178251e-02 6.69083655e-01 8.23398471e-01
-7.91844845e-01 7.29797959e-01 6.91312647e+00 5.76320052e-01
-1.43931806e+00 -1.47004485e-01 1.85625449e-01 3.33749264e-01
-5.05382717e-01 2.10364088e-01 -4.86219734e-01 1.84387974e-02
2.94139951e-01 3.20257664e-01 5.36931694e-01 3.35344285e-01
4.39085215e-01 -6.94165885e-01 -6.07347965e-01 1.10518396e+00
5.70418775e-01 -8.98549795e-01 7.63984621e-02 5.76547487e-03
9.49212372e-01 -1.34979919e-01 8.60320032e-02 -4.03080523e-01
-6.93782344e-02 -8.65043521e-01 3.33744615e-01 7.09440708e-01
8.90515208e-01 -4.57398146e-01 3.78009915e-01 2.82852530e-01
-8.63340080e-01 4.51606624e-02 -4.52744722e-01 2.58526839e-02
4.71636325e-01 7.22218454e-01 -7.52439022e-01 9.67557132e-01
6.51625216e-01 5.56923568e-01 -2.07554117e-01 1.10031998e+00
-2.71135837e-01 3.56705189e-01 -3.75432014e-01 7.62876987e-01
-3.92536893e-02 -1.05904508e+00 8.17169726e-01 8.65211368e-01
7.12752938e-01 1.28099531e-01 -1.24518044e-01 7.12837458e-01
2.52307534e-01 -1.05222508e-01 -8.73599708e-01 4.99680132e-01
1.59216523e-01 1.43597150e+00 -4.16747689e-01 -1.88699499e-01
-7.60318458e-01 9.85461056e-01 -3.96514148e-01 7.77479351e-01
-6.50989234e-01 -1.74002782e-01 1.51594788e-01 1.09166443e-01
1.82386249e-01 -1.15876883e-01 -1.57696724e-01 -1.23680127e+00
-6.51280675e-03 -8.60292912e-01 -2.43433546e-02 -1.44783783e+00
-7.79802382e-01 5.33225238e-01 6.98172152e-02 -1.68135560e+00
1.26757547e-01 -3.49854857e-01 -7.63567746e-01 1.00779474e+00
-1.82402420e+00 -1.37871516e+00 -7.05000520e-01 9.22199965e-01
4.19090688e-01 -1.24264792e-01 4.96388525e-01 4.33334112e-01
-1.80159554e-01 8.82863104e-02 5.28891027e-01 -3.44954848e-01
8.47406447e-01 -6.12021685e-01 -2.85431802e-01 1.33487737e+00
-1.18431441e-01 5.86707771e-01 7.83220112e-01 -4.16225076e-01
-1.51405656e+00 -9.34848309e-01 5.02940476e-01 1.00723356e-01
2.23414022e-02 -1.24551035e-01 -6.13369405e-01 7.02384114e-01
4.60981607e-01 -3.62144202e-01 2.10134998e-01 -3.63092571e-01
-3.68579835e-01 -3.35733712e-01 -1.03681254e+00 5.21925569e-01
7.00425744e-01 -5.85603654e-01 -6.29387915e-01 2.61822402e-01
1.74154967e-01 -3.74877930e-01 -4.47337836e-01 4.80585098e-01
8.22840452e-01 -1.36178994e+00 1.31760967e+00 3.55548829e-01
6.28683448e-01 -4.13152248e-01 -4.04131144e-01 -1.30019915e+00
-2.18342226e-02 -6.14822626e-01 6.18370116e-01 9.19973314e-01
1.82934761e-01 -7.38902509e-01 5.78230500e-01 2.55091283e-02
-1.74582839e-01 4.02425975e-02 -4.21141416e-01 -4.36105639e-01
-3.69753778e-01 -2.16430366e-01 3.22127432e-01 9.15035427e-01
-4.23888505e-01 2.85212517e-01 -1.11370993e+00 5.04365563e-01
9.44930375e-01 6.70227170e-01 1.14278936e+00 -7.84557641e-01
-4.48215544e-01 1.79577455e-01 1.36708409e-01 -1.37429106e+00
-3.61913502e-01 -3.39422077e-01 1.79729849e-01 -1.44713247e+00
3.41015071e-01 -5.80579281e-01 5.46924360e-02 -1.42647699e-01
7.82971233e-02 7.24338055e-01 6.57025501e-02 5.10022044e-01
-2.06097916e-01 5.40187120e-01 1.53580928e+00 1.78200275e-01
-1.39439017e-01 2.49994770e-01 -4.43452418e-01 8.03005874e-01
5.57011485e-01 -1.78903952e-01 -7.27701724e-01 -4.93487865e-01
2.73736447e-01 6.01649284e-01 5.93984425e-01 -1.12103093e+00
5.06841019e-02 -9.20714140e-02 3.36331844e-01 -7.92718768e-01
7.44101524e-01 -1.23611414e+00 5.18823147e-01 2.41715193e-01
1.18439071e-01 -2.57495075e-01 -1.80940181e-01 6.64582908e-01
-2.38857612e-01 -2.93291509e-01 1.00089681e+00 -1.31219551e-01
-5.95377326e-01 -9.77147669e-02 -2.97932982e-01 -3.96685332e-01
7.28236794e-01 -5.00445664e-01 -8.28405678e-01 -7.00270534e-01
-1.77349031e-01 -3.93519461e-01 7.88115084e-01 1.75251976e-01
8.23726237e-01 -1.04137886e+00 -5.33847690e-01 5.51076889e-01
2.23133471e-02 -2.98606038e-01 6.27170265e-01 1.06038237e+00
-8.73239040e-01 3.85125011e-01 -4.75642443e-01 -5.72565258e-01
-1.67143869e+00 3.78905267e-01 4.69998538e-01 -6.68139830e-02
-1.16662836e+00 6.83377013e-02 6.99320018e-01 -4.22717839e-01
-3.12932014e-01 1.73475400e-01 -1.89866319e-01 -6.73018396e-01
5.55433095e-01 4.44228858e-01 9.47595760e-02 -9.27977383e-01
2.02175155e-02 1.16001427e+00 1.79923117e-01 -2.97175884e-01
1.15721130e+00 -7.89019048e-01 -1.14661641e-01 -4.34715636e-02
1.09468782e+00 5.97074747e-01 -1.14529645e+00 -5.41177273e-01
-7.77164876e-01 -9.45949674e-01 1.91190869e-01 -6.02304280e-01
-1.16886187e+00 7.24833846e-01 6.53359771e-01 -8.17704573e-02
1.35997927e+00 -4.80627894e-01 7.27744341e-01 4.83493596e-01
2.61794686e-01 -9.41943109e-01 1.25070632e-01 2.73379743e-01
7.76375353e-01 -1.12082219e+00 4.56650048e-01 -7.55979717e-01
-4.07960176e-01 1.29601991e+00 2.91309148e-01 -1.56957865e-01
5.42369008e-01 -3.59288119e-02 4.40450490e-01 -2.51317024e-01
-1.61908820e-01 1.11103743e-01 1.90883920e-01 5.99753380e-01
1.18391372e-01 -3.10537428e-01 -1.99612871e-01 -2.90804029e-01
-1.36521578e-01 -8.50684345e-02 9.97492373e-01 7.39463389e-01
-5.18224061e-01 -7.68502355e-01 -9.76525903e-01 4.31220382e-02
-3.20009679e-01 -1.78081021e-01 -1.12244580e-02 9.92949724e-01
-3.28631818e-01 1.12513888e+00 4.08624411e-02 -9.84713435e-02
1.76940009e-01 -4.40771490e-01 5.79307377e-01 -1.14065208e-01
-5.64899966e-02 6.53788149e-01 2.27567315e-01 -5.20256937e-01
-7.47026682e-01 -4.75624681e-01 -6.50345862e-01 -1.31898850e-01
-4.26698297e-01 -1.01349898e-01 7.80792952e-01 8.13085556e-01
-9.84485894e-02 3.29625785e-01 1.03223860e+00 -7.68987179e-01
-2.40038916e-01 -7.96164751e-01 -8.05833638e-01 4.55710083e-01
4.78363782e-01 -3.11259687e-01 -4.70069200e-01 4.67459857e-02] | [10.156574249267578, -2.813112735748291] |
26282afa-61bd-4aed-90a5-fb4bdb2871b0 | stereoscene-bev-assisted-stereo-matching | 2303.13959 | null | https://arxiv.org/abs/2303.13959v2 | https://arxiv.org/pdf/2303.13959v2.pdf | StereoScene: BEV-Assisted Stereo Matching Empowers 3D Semantic Scene Completion | 3D semantic scene completion (SSC) is an ill-posed task that requires inferring a dense 3D scene from incomplete observations. Previous methods either explicitly incorporate 3D geometric input or rely on learnt 3D prior behind monocular RGB images. However, 3D sensors such as LiDAR are expensive and intrusive while monocular cameras face challenges in modeling precise geometry due to the inherent ambiguity. In this work, we propose StereoScene for 3D Semantic Scene Completion (SSC), which explores taking full advantage of light-weight camera inputs without resorting to any external 3D sensors. Our key insight is to leverage stereo matching to resolve geometric ambiguity. To improve its robustness in unmatched areas, we introduce bird's-eye-view (BEV) representation to inspire hallucination ability with rich context information. On top of the stereo and BEV representations, a mutual interactive aggregation (MIA) module is carefully devised to fully unleash their power. Specifically, a Bi-directional Interaction Transformer (BIT) augmented with confidence re-weighting is used to encourage reliable prediction through mutual guidance while a Dual Volume Aggregation (DVA) module is designed to facilitate complementary aggregation. Experimental results on SemanticKITTI demonstrate that the proposed StereoScene outperforms the state-of-the-art camera-based methods by a large margin with a relative improvement of 26.9% in geometry and 38.6% in semantic. | ['Dalong Du', 'Hang Xiao', 'James Okae', 'Yunpeng Zhang', 'Xiaoefeng Wang', 'Zheng Zhu', 'Wenjun Zeng', 'Xin Jin', 'Yasheng Sun', 'Bohan Li'] | 2023-03-24 | null | null | null | null | ['3d-semantic-scene-completion', 'stereo-matching-1'] | ['computer-vision', 'computer-vision'] | [ 2.55650729e-01 1.44281015e-01 1.57156155e-01 -4.91346538e-01
-6.05938256e-01 -4.23573852e-01 5.38603127e-01 -3.12373608e-01
-1.27255693e-01 3.30500931e-01 3.44553769e-01 -1.80280685e-01
7.82052428e-02 -6.89704895e-01 -7.67463863e-01 -5.92861414e-01
4.20712203e-01 2.19885543e-01 3.18428010e-01 -2.26118460e-01
2.85548925e-01 4.03068423e-01 -1.62592137e+00 1.23884998e-01
1.11287510e+00 1.09973621e+00 7.91491568e-01 2.23070398e-01
-1.61562070e-01 5.41299284e-01 -6.84464797e-02 -1.04409702e-01
6.30845010e-01 4.81516682e-02 -2.98774362e-01 4.54584628e-01
5.23365617e-01 -6.13326788e-01 -4.28544819e-01 1.04001307e+00
3.08463395e-01 1.44539043e-01 4.03924763e-01 -1.15113151e+00
-3.07820946e-01 -2.50349343e-01 -8.48201334e-01 -1.60428450e-01
9.31488156e-01 2.46329919e-01 8.52058589e-01 -1.33175349e+00
5.26378155e-01 1.23477757e+00 5.05512834e-01 2.45035052e-01
-1.17025328e+00 -8.09878647e-01 3.73663813e-01 3.48648839e-02
-1.47309303e+00 -4.25569832e-01 1.17190695e+00 -2.48909444e-01
8.85411382e-01 2.28693500e-01 6.95865035e-01 9.29328263e-01
-1.69835463e-01 6.98066711e-01 1.32986355e+00 -2.95570493e-01
3.00123841e-01 2.93260247e-01 -2.45841578e-01 6.83763325e-01
1.97751999e-01 2.74038196e-01 -8.16954613e-01 -2.75960378e-02
1.08475506e+00 4.42862332e-01 -3.60272050e-01 -7.62527406e-01
-1.05324674e+00 6.27660930e-01 7.88247108e-01 -2.47729823e-01
-4.27166104e-01 -6.60254955e-02 -1.64830506e-01 -8.68956074e-02
4.52863753e-01 2.44890049e-01 -2.20507473e-01 2.06569389e-01
-7.43319631e-01 1.39481902e-01 2.74109274e-01 1.27172661e+00
1.06833088e+00 1.63739026e-01 2.60633945e-01 7.09485471e-01
5.09655058e-01 6.77370250e-01 6.61843121e-02 -1.23009634e+00
7.42358863e-01 9.45444524e-01 4.85276431e-02 -9.91708100e-01
-2.22813934e-01 -4.03042108e-01 -7.70735204e-01 4.50796843e-01
5.97178303e-02 3.98628205e-01 -1.00607491e+00 1.31346011e+00
6.46676123e-01 2.62308359e-01 6.25251373e-03 1.25372362e+00
6.80370510e-01 3.50904107e-01 -2.85434037e-01 1.25904337e-01
1.12534618e+00 -7.04801977e-01 -2.78724432e-01 -5.32534480e-01
1.49225309e-01 -7.87556767e-01 1.09204078e+00 3.38259995e-01
-9.06613052e-01 -5.13756931e-01 -1.07238436e+00 -3.81266445e-01
-1.73297361e-01 -1.99846089e-01 6.37655914e-01 4.80726242e-01
-8.73930335e-01 8.93684998e-02 -7.82972693e-01 -3.17890823e-01
4.22059149e-01 9.94419008e-02 -5.75458884e-01 -5.75208604e-01
-7.73434699e-01 7.19833970e-01 2.14914456e-01 -7.19044730e-02
-8.04460466e-01 -7.02762783e-01 -1.21653879e+00 -2.01763108e-01
6.83953106e-01 -9.64804053e-01 8.27866495e-01 -4.32238102e-01
-1.41118264e+00 8.25301826e-01 -3.17385703e-01 -1.42929301e-01
5.09917974e-01 -3.62919748e-01 -4.44654152e-02 4.69496638e-01
1.71032235e-01 8.91500056e-01 7.78824151e-01 -1.65296292e+00
-5.89731693e-01 -7.85410941e-01 4.20636296e-01 8.24575663e-01
-2.27182917e-02 -5.75542212e-01 -6.54151320e-01 -5.74439943e-01
8.38600278e-01 -8.58489931e-01 -3.72802168e-01 3.54876548e-01
-4.39180642e-01 2.17304140e-01 8.43371391e-01 -6.21967018e-01
6.99876070e-01 -2.04309082e+00 6.56611323e-02 1.27941877e-01
3.07454407e-01 -2.12021284e-02 1.00006886e-01 2.73061365e-01
1.97981343e-01 -3.11606586e-01 -2.96446264e-01 -7.71525025e-01
-1.45685598e-01 3.24950486e-01 -4.29170847e-01 5.25802553e-01
1.98349312e-01 6.90097034e-01 -9.39952075e-01 -3.28042239e-01
9.13069129e-01 7.73979664e-01 -9.48066890e-01 3.79564643e-01
-2.07988203e-01 6.14902079e-01 -6.61313593e-01 9.63368356e-01
1.02091634e+00 -3.24144393e-01 -1.12655640e-01 -3.74039859e-01
-1.68162972e-01 2.69627601e-01 -1.36341858e+00 2.38765693e+00
-5.59531331e-01 1.04561955e-01 1.65533334e-01 -5.99366009e-01
1.04583442e+00 4.58529517e-02 2.84046680e-01 -7.95121551e-01
1.08397147e-02 8.23793486e-02 -6.84392691e-01 -2.52155513e-01
5.81154346e-01 -1.36505753e-01 5.22751212e-02 1.01566270e-01
-2.31420055e-01 -7.50330627e-01 -6.11770749e-01 3.70755136e-01
9.12822783e-01 5.48966467e-01 3.83502483e-01 -4.56995070e-02
5.20023108e-01 -5.00884764e-02 7.12965488e-01 3.63470256e-01
-4.20012474e-02 1.07768714e+00 -1.01083830e-01 -2.47607216e-01
-9.74259198e-01 -1.31201327e+00 5.47988340e-02 3.94329309e-01
6.64815426e-01 -3.30438524e-01 -3.37614030e-01 -4.59981412e-01
6.54561147e-02 8.19005072e-01 -3.90352547e-01 8.00409261e-03
-2.40247041e-01 -2.86343187e-01 -9.63701680e-02 6.53502405e-01
8.47558796e-01 -4.85704571e-01 -8.79307389e-01 -1.17277682e-01
-2.84776270e-01 -1.39402401e+00 -4.01589811e-01 2.81907804e-02
-1.01765764e+00 -1.06380177e+00 -5.70078194e-01 -3.16960782e-01
8.14845443e-01 1.13701880e+00 7.43542492e-01 -2.30724111e-01
-1.73983857e-01 4.83650595e-01 -4.33725864e-01 -2.07678586e-01
1.53505564e-01 -3.12337816e-01 1.84986606e-01 -2.19087768e-02
2.52232194e-01 -1.04134166e+00 -9.78846133e-01 3.97890866e-01
-8.48649263e-01 7.58178711e-01 7.06007302e-01 6.85128629e-01
6.07665300e-01 -2.52782732e-01 3.27644497e-02 -5.66863716e-01
-2.12561652e-01 -4.83132958e-01 -6.70761108e-01 -1.32345214e-01
-5.39932489e-01 -2.95625087e-02 3.50973696e-01 -7.62791783e-02
-1.37232316e+00 3.82831037e-01 4.33112197e-02 -9.84654725e-01
-2.10685134e-01 1.36419356e-01 -4.55440581e-01 -7.80932382e-02
4.06945348e-01 3.27716589e-01 1.06501710e-02 -5.24580777e-01
4.67776388e-01 5.71343541e-01 5.48813462e-01 -4.16440874e-01
9.45875347e-01 1.02420712e+00 -4.94834222e-02 -7.23962605e-01
-1.06166768e+00 -7.36112118e-01 -6.79518521e-01 -2.27610022e-01
8.32854986e-01 -1.50915551e+00 -4.64142650e-01 2.08663300e-01
-9.83116806e-01 -2.99785305e-02 -4.11170013e-02 5.23251951e-01
-5.85930526e-01 5.81665754e-01 -2.61451364e-01 -9.56356585e-01
-1.33328959e-01 -1.08178699e+00 1.52416778e+00 2.37648994e-01
-3.54138538e-02 -5.63901365e-01 -3.61162931e-01 9.55712557e-01
1.23950049e-01 3.47686976e-01 4.47683960e-01 5.57497032e-02
-1.16158414e+00 2.03159600e-02 -5.66177249e-01 2.21425608e-01
1.52626321e-01 -6.13684237e-01 -1.43857038e+00 -1.31391540e-01
2.03136757e-01 -2.29371652e-01 6.87038422e-01 1.37031838e-01
9.87912714e-01 1.05028741e-01 -1.90578118e-01 9.77028847e-01
1.56331789e+00 1.20424498e-02 5.23171186e-01 2.64856458e-01
1.05129933e+00 7.12022305e-01 7.72372842e-01 6.91703618e-01
8.30446661e-01 7.16366231e-01 8.87634397e-01 -8.50617066e-02
-2.36804575e-01 -6.64705157e-01 2.18194976e-01 6.37247026e-01
-5.90629131e-02 1.32070228e-01 -7.68875837e-01 2.56274939e-01
-1.65579426e+00 -6.14718139e-01 8.79434124e-03 2.28397894e+00
5.82501888e-01 1.89343944e-01 -3.33562613e-01 2.01584995e-01
4.90918845e-01 2.90192842e-01 -7.27954507e-01 2.93438435e-01
-2.60670394e-01 -3.00431196e-02 4.65257168e-01 6.91349804e-01
-6.58243299e-01 9.93904412e-01 4.56542301e+00 6.99209809e-01
-8.53869855e-01 -3.33204120e-02 4.35759455e-01 -9.91302282e-02
-6.51120424e-01 2.95532137e-01 -7.16211140e-01 2.25179866e-01
-2.09900085e-02 2.65181541e-01 4.47095722e-01 7.74706125e-01
2.43185386e-01 -5.23358881e-01 -9.17755008e-01 1.39463687e+00
1.95425466e-01 -1.13150907e+00 4.01982889e-02 2.83646673e-01
7.46506691e-01 6.67965859e-02 3.92115489e-03 1.16697736e-02
2.90893435e-01 -6.25877082e-01 8.62098575e-01 5.31399548e-01
9.46949244e-01 -5.29819191e-01 3.77046674e-01 5.32480836e-01
-1.29995406e+00 -5.72906546e-02 -3.61918896e-01 -3.21081340e-01
3.94741058e-01 8.72249484e-01 -7.55587339e-01 8.24346423e-01
8.59404445e-01 7.59031534e-01 -3.83001268e-01 8.06000590e-01
-3.53139490e-01 3.06028482e-02 -5.39129019e-01 4.46113467e-01
1.94861382e-01 -3.40379685e-01 7.33626783e-01 5.95247328e-01
4.25337225e-01 5.77223539e-01 2.53147960e-01 1.03424954e+00
2.05259264e-01 -2.76768744e-01 -8.06432545e-01 5.76703787e-01
6.03626847e-01 1.17990577e+00 -5.40523231e-01 -1.86356753e-01
-5.58334410e-01 1.26012492e+00 2.48748332e-01 4.43100542e-01
-6.15073740e-01 5.88138886e-02 7.74580777e-01 3.13265681e-01
3.57967436e-01 -4.58338439e-01 -6.36292636e-01 -1.39375615e+00
3.14458877e-01 -3.55817527e-01 1.58583850e-01 -1.39262009e+00
-1.18071377e+00 4.55889434e-01 -5.10438830e-02 -1.49898326e+00
-3.54695059e-02 -2.67705202e-01 -3.41283768e-01 9.86690938e-01
-1.81807172e+00 -1.35268188e+00 -8.86072099e-01 8.67687881e-01
7.11894214e-01 2.21755818e-01 5.55651784e-01 2.88488232e-02
-2.35383138e-01 9.31120738e-02 -3.96726519e-01 -3.14892769e-01
5.27931690e-01 -1.05425537e+00 2.82646745e-01 8.39539409e-01
5.53639084e-02 4.72288042e-01 6.70802951e-01 -7.21693754e-01
-1.72405875e+00 -1.09065139e+00 6.21477008e-01 -6.43894911e-01
1.93329602e-01 -5.44203222e-01 -7.10172772e-01 4.56851512e-01
-1.66192219e-01 1.89805344e-01 3.36075872e-01 -1.57049298e-01
-6.41841054e-01 -1.70353726e-01 -1.22184503e+00 6.75579667e-01
1.49400485e+00 -7.13970006e-01 -6.76594317e-01 -1.45394579e-02
9.98276114e-01 -5.80002308e-01 -5.68417788e-01 5.93969107e-01
3.97214472e-01 -1.37131703e+00 1.30145764e+00 1.44584477e-01
4.71876413e-01 -5.78566909e-01 -7.40546763e-01 -1.01426232e+00
1.71565171e-02 -5.22872388e-01 -9.52034891e-02 9.57339466e-01
-1.26303047e-01 -5.48238218e-01 1.01844203e+00 8.18989396e-01
-3.69885653e-01 -7.03946888e-01 -8.36510479e-01 -6.17622137e-01
-5.85217357e-01 -8.45722854e-01 5.27619541e-01 9.47683394e-01
-1.21670440e-01 3.42909127e-01 -3.43095511e-01 5.62505782e-01
8.93974960e-01 3.44069213e-01 1.03850257e+00 -1.01664102e+00
-1.58158436e-01 -5.77988140e-02 -4.41036910e-01 -1.51212978e+00
-2.46310323e-01 -6.24226570e-01 -1.42653868e-01 -1.42392826e+00
1.53237581e-01 -5.50245464e-01 -2.69324798e-02 2.10000411e-01
-1.20418347e-01 3.92045557e-01 3.38366210e-01 1.33209363e-01
-5.53766489e-01 1.04440629e+00 1.23016143e+00 1.54539421e-01
-2.42924839e-01 -1.94892868e-01 -7.42369533e-01 9.57041621e-01
4.31560338e-01 -1.36820525e-01 -6.42806470e-01 -6.37646973e-01
1.56052545e-01 3.21820468e-01 7.40384519e-01 -1.01131475e+00
3.39683622e-01 -1.01796918e-01 4.63220447e-01 -9.88437414e-01
9.36513960e-01 -1.02458537e+00 1.95880368e-01 4.71238457e-02
2.46166766e-01 -1.45877331e-01 -3.62096447e-03 8.12436581e-01
-1.96602449e-01 3.00592750e-01 5.09748280e-01 -3.12065244e-01
-9.20009434e-01 4.72866386e-01 1.65257588e-01 -1.04538903e-01
9.03088212e-01 -7.25002944e-01 4.44850549e-02 -4.89038497e-01
-4.49096322e-01 2.89165050e-01 9.38240647e-01 3.78659725e-01
1.18193924e+00 -1.35021925e+00 -3.15678537e-01 5.29141426e-01
4.66958970e-01 6.00192666e-01 4.98113513e-01 7.28213668e-01
-3.42341781e-01 4.36537683e-01 -5.47170117e-02 -9.84109938e-01
-1.01475775e+00 3.70708853e-01 -8.24289098e-02 1.28196985e-01
-9.60334539e-01 8.61221910e-01 6.85333490e-01 -5.91386080e-01
3.32680106e-01 -3.45728636e-01 1.86659813e-01 -3.09317201e-01
4.29406077e-01 2.73391843e-01 -7.70847946e-02 -5.94875574e-01
-3.34472686e-01 8.38372886e-01 8.22431445e-02 -1.98417500e-01
1.27014542e+00 -6.49042845e-01 3.15434366e-01 2.05914408e-01
1.01959610e+00 7.07394502e-04 -1.89202523e+00 -6.12472236e-01
-4.34085965e-01 -1.00179684e+00 3.10605794e-01 -6.36097729e-01
-9.07602608e-01 1.04630673e+00 3.29747856e-01 -3.79386961e-01
1.31089580e+00 4.02937345e-02 7.93656886e-01 2.22763434e-01
9.30774927e-01 -7.54375458e-01 2.11128622e-01 3.91047955e-01
9.63160753e-01 -1.49696100e+00 1.19395696e-01 -9.35131431e-01
-7.64879704e-01 8.61875296e-01 7.38985300e-01 -9.80221108e-02
5.35654724e-01 9.36503988e-03 -1.20053567e-01 -3.84644389e-01
-4.85151350e-01 -3.41468453e-01 3.08487505e-01 6.12044692e-01
-1.11470789e-01 -1.14812963e-01 4.51669991e-01 4.11707461e-01
-1.67219505e-01 -1.96180388e-01 2.99782306e-01 8.71684909e-01
-4.47471261e-01 -6.09432042e-01 -5.12953281e-01 1.69249296e-01
1.55506834e-01 -2.03069851e-01 -1.77932661e-02 6.62956357e-01
1.44392461e-01 1.04428768e+00 1.67787224e-02 -5.54490149e-01
4.22519803e-01 -3.44568223e-01 5.46073496e-01 -7.25220144e-01
4.01928201e-02 3.10075492e-01 -9.54000726e-02 -1.11021328e+00
-4.28900480e-01 -6.78462684e-01 -1.15453231e+00 -1.58846453e-01
-3.10267299e-01 -3.76831532e-01 7.26717293e-01 8.97480607e-01
3.91097754e-01 9.92742032e-02 7.23398089e-01 -1.30151010e+00
-2.16462269e-01 -6.91975236e-01 -5.63624799e-01 2.97334284e-01
3.25962961e-01 -9.85914528e-01 -4.70047474e-01 -7.00832978e-02] | [8.607855796813965, -2.801826000213623] |
3c73c41a-591b-4e4d-ba73-73ba0ee70cf4 | astra-a-novel-algorithm-level-approach-to | 2209.01685 | null | https://arxiv.org/abs/2209.01685v1 | https://arxiv.org/pdf/2209.01685v1.pdf | ASTra: A Novel Algorithm-Level Approach to Imbalanced Classification | We propose a novel output layer activation function, which we name ASTra (Asymmetric Sigmoid Transfer function), which makes the classification of minority examples, in scenarios of high imbalance, more tractable. We combine this with a loss function that helps to effectively target minority misclassification. These two methods can be used together or separately, with their combination recommended for the most severely imbalanced cases. The proposed approach is tested on datasets with IRs from 588.24 to 4000 and very few minority examples (in some datasets, as few as five). Results using neural networks with from two to 12 hidden units are demonstrated to be comparable to, or better than, equivalent results obtained in a recent study that deployed a wide range of complex, hybrid data-level ensemble classifiers. | ['Denise Gorse', 'David Twomey'] | 2022-09-04 | null | null | null | null | ['imbalanced-classification'] | ['miscellaneous'] | [ 3.76573503e-01 2.80594677e-01 -2.51552820e-01 -4.97540861e-01
-3.94388944e-01 -3.51673067e-01 5.19442856e-01 5.18824220e-01
-7.24764526e-01 1.17219841e+00 -2.57701278e-01 -5.93287766e-01
-3.45723897e-01 -8.55395555e-01 -4.64302808e-01 -7.55677640e-01
-6.70036748e-02 4.13757920e-01 2.07942780e-02 -2.39861324e-01
3.04319084e-01 5.96233428e-01 -1.95427561e+00 6.23478949e-01
1.12775791e+00 1.08578193e+00 -5.73587775e-01 4.98651296e-01
-6.82732183e-03 6.96696818e-01 -1.15446770e+00 -3.71431142e-01
2.58297354e-01 -1.00680396e-01 -3.64488512e-01 -2.72444129e-01
5.68744302e-01 -6.51808307e-02 2.96056420e-01 3.93549383e-01
7.79192746e-01 8.19735005e-02 7.81823516e-01 -1.24671853e+00
-2.45608002e-01 5.27850091e-01 -3.98053199e-01 4.35725212e-01
3.30096595e-02 -4.50632237e-02 5.06960511e-01 -7.91002572e-01
2.64773935e-01 9.94154274e-01 1.23894811e+00 2.97341883e-01
-1.09233809e+00 -8.45688879e-01 -5.02986088e-03 8.10927451e-02
-1.29088962e+00 -4.09631789e-01 4.76575494e-01 -3.93456608e-01
1.33517075e+00 4.59169567e-01 4.18496907e-01 8.05437267e-01
2.22779170e-01 1.93190664e-01 1.23194492e+00 -5.79063714e-01
1.39272571e-01 6.77220821e-01 4.18940455e-01 9.17664915e-02
8.15106869e-01 -2.31988411e-02 -2.28559017e-01 -2.63650090e-01
-8.80680420e-03 -1.77093334e-02 -4.46346998e-02 1.37685552e-01
-7.79617667e-01 9.58978534e-01 5.02033949e-01 5.29233992e-01
-5.42445481e-01 -2.67921656e-01 5.60770512e-01 8.61683428e-01
8.61487031e-01 4.73179668e-01 -3.33716691e-01 1.19000919e-01
-8.61823738e-01 3.13857973e-01 9.74564075e-01 3.33409309e-01
4.44156885e-01 3.25119019e-01 -1.34917095e-01 8.83771718e-01
-6.61321729e-03 1.24770500e-01 5.83934665e-01 -4.06444252e-01
6.90692186e-01 9.97986257e-01 2.27844432e-01 -8.62786651e-01
-7.91188657e-01 -8.15720379e-01 -1.02424133e+00 6.32588863e-01
5.17799258e-01 -3.00837219e-01 -9.68600810e-01 1.44167101e+00
2.35919386e-01 -1.19239027e-02 3.26202303e-01 6.10759497e-01
7.16034174e-01 4.91640151e-01 1.29708827e-01 -1.04117706e-01
1.08247137e+00 -5.93380690e-01 -6.30029678e-01 -1.15147159e-02
5.70789218e-01 -6.57588124e-01 8.63650143e-01 6.73106432e-01
-1.04011810e+00 -6.36961579e-01 -1.41903496e+00 2.18072116e-01
-9.75337267e-01 1.87090695e-01 4.38991010e-01 1.16242433e+00
-1.01164901e+00 8.29154253e-01 -4.84477788e-01 -1.78241208e-01
5.12541652e-01 8.18295836e-01 -2.18214333e-01 1.96899548e-01
-1.43737936e+00 1.27081573e+00 6.41331732e-01 3.23288769e-01
-1.99938521e-01 -6.52832747e-01 -7.05641687e-01 8.39875918e-03
-7.99534544e-02 -5.85794747e-01 7.04374909e-01 -1.26597500e+00
-1.20086598e+00 9.01471794e-01 3.21852207e-01 -7.63032377e-01
8.16012919e-01 -3.43181849e-01 -5.74785948e-01 -3.37293506e-01
-3.38071406e-01 4.67933446e-01 6.38963163e-01 -1.13285208e+00
-5.29242992e-01 -5.81861615e-01 -4.89846058e-03 1.96493492e-01
-5.97899377e-01 -6.53957054e-02 6.10560715e-01 -7.04899251e-01
-1.85751885e-01 -5.33407331e-01 -5.71238399e-02 -3.43549669e-01
-2.90164590e-01 -2.89502263e-01 8.28254104e-01 -8.45881999e-01
1.39263535e+00 -1.77737880e+00 -1.35909557e-01 4.48726118e-01
1.48221515e-02 5.75873137e-01 1.83870956e-01 4.29590344e-01
-4.73907530e-01 1.33932605e-01 -4.92860019e-01 -4.05051529e-01
-1.09895311e-01 1.05398871e-01 4.20873333e-03 5.55593789e-01
5.62698364e-01 3.83640498e-01 -3.44016641e-01 1.11783177e-01
1.87859595e-01 4.96343195e-01 -3.30162287e-01 1.62846163e-01
3.09987992e-01 1.43067688e-01 6.24436587e-02 7.35368133e-01
8.88138592e-01 1.49352431e-01 2.28200033e-02 6.49027526e-03
-1.89717337e-01 1.71019301e-01 -1.40860701e+00 6.18785143e-01
-6.05253041e-01 5.51309168e-01 -2.31010109e-01 -1.37800503e+00
1.27586675e+00 2.70259947e-01 7.08517656e-02 -4.90883917e-01
1.96535513e-01 5.33740580e-01 4.42096263e-01 -4.14095283e-01
2.92334259e-01 -4.19111222e-01 -2.49768067e-02 2.97887534e-01
8.88242349e-02 2.40465641e-01 2.37017155e-01 -4.71763313e-01
7.44691491e-01 -1.38706848e-01 4.60863084e-01 -2.63756007e-01
7.41956413e-01 -2.00907886e-01 5.03808141e-01 8.44261289e-01
-1.91691518e-01 5.28243423e-01 5.55629194e-01 -6.71567619e-01
-1.26835084e+00 -7.62333870e-01 -4.72303808e-01 9.77828801e-01
-1.30209416e-01 1.31095782e-01 -6.40253603e-01 -4.65646774e-01
3.32692713e-01 6.90722406e-01 -6.61716521e-01 -4.16712165e-01
-6.71023309e-01 -1.38514233e+00 8.13969195e-01 5.67870796e-01
4.77509528e-01 -1.27327728e+00 -6.44189119e-01 2.08329901e-01
4.15960521e-01 -7.70085454e-01 4.21871424e-01 7.10877836e-01
-1.18725705e+00 -1.00583005e+00 -7.36936867e-01 -5.56395829e-01
5.58426619e-01 -4.39608186e-01 1.37939715e+00 3.52665305e-01
-2.23172128e-01 -3.31476003e-01 -2.55380511e-01 -7.40455747e-01
-4.11269635e-01 4.32208657e-01 1.05018444e-01 3.72976996e-02
5.52876770e-01 -6.20981276e-01 -3.68665874e-01 7.58544132e-02
-9.82186079e-01 -3.17498207e-01 7.10369229e-01 1.04906535e+00
-8.13515261e-02 -1.34851094e-02 1.15593815e+00 -1.23312056e+00
6.90245748e-01 -8.30588937e-01 -2.71763474e-01 2.67434679e-02
-8.49324822e-01 -3.77035618e-01 1.03305924e+00 -5.43931723e-01
-7.97576308e-01 -5.31816006e-01 -3.39797229e-01 -1.12086132e-01
-4.26047295e-01 2.39991948e-01 5.86259626e-02 -1.89505994e-01
8.21118832e-01 -9.92597640e-02 1.63961366e-01 -5.00096440e-01
-2.74668097e-01 1.08646131e+00 1.33440718e-01 -4.86981571e-02
4.49510634e-01 1.92364186e-01 -1.96757346e-01 -6.16514266e-01
-4.81963724e-01 -2.05446348e-01 -4.60691273e-01 5.70565276e-02
3.15277249e-01 -8.30654860e-01 -6.40563905e-01 9.16845143e-01
-6.72408342e-01 -3.08318347e-01 -2.29408011e-01 4.05069411e-01
-2.26258114e-01 -1.65416047e-01 -5.99120975e-01 -1.20669830e+00
-5.27704716e-01 -7.54996657e-01 6.75471842e-01 4.04721797e-01
-2.88106114e-01 -1.03991222e+00 -2.43125647e-01 2.53634304e-01
7.72100866e-01 7.24145889e-01 1.02380610e+00 -1.15100825e+00
2.81927168e-01 -3.79682362e-01 -9.04687941e-02 7.59819865e-01
2.12711282e-02 1.50307208e-01 -1.24312794e+00 -5.03235281e-01
-3.93843800e-02 -3.99683207e-01 1.00358725e+00 1.87500566e-01
1.05093956e+00 -3.06980669e-01 -2.26656735e-01 4.11888808e-01
1.50497711e+00 3.44084859e-01 8.66941869e-01 6.86787784e-01
3.44883829e-01 7.36910164e-01 5.37892520e-01 4.44247961e-01
2.23941088e-01 6.35082901e-01 4.48916048e-01 -4.73976523e-01
3.76197957e-02 4.30967808e-01 1.35148019e-01 5.75695813e-01
-3.93810719e-01 -3.70230883e-01 -9.19601262e-01 4.43941474e-01
-1.41068184e+00 -8.26956213e-01 -2.30969369e-01 2.34020638e+00
7.29624152e-01 6.87867761e-01 3.54625791e-01 7.75488436e-01
9.34508681e-01 -4.97006103e-02 -5.12172222e-01 -1.12449574e+00
-3.16310585e-01 4.90958482e-01 4.96056885e-01 4.52865213e-01
-1.30886137e+00 3.23771149e-01 7.07010317e+00 6.10586941e-01
-1.22815919e+00 1.26110256e-01 1.08115113e+00 -1.77107245e-01
-1.12469435e-01 -4.62772161e-01 -7.81506300e-01 7.78902829e-01
1.36428988e+00 1.89591095e-01 4.55839075e-02 6.07579887e-01
4.02424447e-02 -1.72298759e-01 -6.83857679e-01 5.92875779e-01
2.01466441e-01 -8.80614340e-01 -9.85506624e-02 -2.84634471e-01
8.77714992e-01 -2.30709419e-01 1.48193508e-01 5.74723303e-01
-2.24809542e-01 -1.42000639e+00 4.16791588e-01 4.06639457e-01
6.59900784e-01 -1.12665999e+00 1.49438322e+00 3.94658715e-01
-5.27795911e-01 -5.59776425e-01 -2.32685760e-01 -5.87205231e-01
-2.70338982e-01 8.66610050e-01 -8.97499979e-01 6.07606709e-01
9.28874850e-01 2.98711270e-01 -5.64263999e-01 9.64140356e-01
3.05981338e-01 5.84478021e-01 -5.36474884e-01 -1.64936662e-01
2.02313662e-01 -1.31384395e-02 2.09112659e-01 1.15283477e+00
4.66136128e-01 -2.33002752e-01 -3.19917321e-01 4.91886497e-01
-1.50886076e-02 1.93999723e-01 -6.83708191e-01 4.71070439e-01
4.84514445e-01 1.23233771e+00 -5.98014593e-01 -5.02896369e-01
-4.17435803e-02 5.20551085e-01 2.78546304e-01 2.17457712e-01
-7.55578816e-01 -7.64863372e-01 5.11762619e-01 8.60790387e-02
1.23046480e-01 3.62994641e-01 -5.95966756e-01 -7.77710259e-01
2.34823346e-01 -1.02637279e+00 7.11605251e-01 -2.96935856e-01
-1.29304910e+00 7.95551181e-01 1.49858743e-02 -1.28864217e+00
-3.13212901e-01 -8.79643619e-01 -9.38807309e-01 1.05340421e+00
-1.55813873e+00 -8.93622339e-01 -5.19009352e-01 1.18563779e-01
1.59373313e-01 -2.86158174e-01 8.17582130e-01 7.47058749e-01
-5.54816782e-01 9.15358305e-01 2.03916594e-01 -3.15056801e-01
6.58266544e-01 -1.34067857e+00 2.29244363e-02 2.63956010e-01
-5.15269518e-01 4.64979976e-01 7.10938811e-01 -3.25839639e-01
-5.72824478e-01 -1.03064811e+00 1.04500782e+00 -2.21502677e-01
1.22945281e-02 -4.14465904e-01 -1.05552220e+00 3.87333602e-01
1.76414043e-01 -1.49150655e-01 9.07564819e-01 1.05037332e-01
1.08298860e-01 -2.92122275e-01 -1.75970018e+00 1.62041336e-01
6.20260060e-01 -3.60300615e-02 -6.86299860e-01 1.06031552e-01
2.34062046e-01 -4.65568602e-01 -1.25080180e+00 9.62395012e-01
6.82751119e-01 -1.33869886e+00 9.00871158e-01 -7.25598574e-01
2.40427881e-01 -1.81803152e-01 1.68287978e-01 -1.46293199e+00
1.49647877e-01 -7.94220716e-02 -2.81787872e-01 1.15560913e+00
5.40428579e-01 -1.16139627e+00 7.24815488e-01 9.16585252e-02
-7.36030638e-02 -1.15349019e+00 -1.06695521e+00 -7.43324816e-01
2.57148594e-01 -1.17039271e-01 5.91781020e-01 9.95924890e-01
-3.63831371e-01 -1.32513836e-01 -2.98678637e-01 -1.01594515e-01
2.71356374e-01 -8.97961929e-02 4.50790018e-01 -1.35209095e+00
-3.10785528e-02 -5.95670342e-01 -7.95157135e-01 2.57131867e-02
3.80547382e-02 -7.78701127e-01 -3.83036882e-01 -9.59683895e-01
-2.32984945e-01 -8.17169726e-01 -7.52023458e-01 5.28135061e-01
-3.58454853e-01 9.08644855e-01 3.39363911e-03 5.35463681e-03
3.42175327e-02 2.36077100e-01 7.07367420e-01 -2.29810923e-02
-1.85273781e-01 2.97846317e-01 -8.49752128e-01 7.13464022e-01
1.00384665e+00 -4.58127618e-01 -1.02577638e-02 -2.98778955e-02
6.50448576e-02 -3.98590893e-01 1.46640643e-01 -1.41019034e+00
-1.87494248e-01 2.59050488e-01 8.42803657e-01 -5.56334853e-01
2.60443151e-01 -7.70151377e-01 2.42392510e-01 8.68969381e-01
-3.04653823e-01 1.79905981e-01 3.97790641e-01 -4.82511185e-02
-5.03401995e-01 -3.86349767e-01 7.52223551e-01 1.68447465e-01
-4.09608334e-01 -2.62236446e-01 -1.46237478e-01 -2.22417951e-01
1.28302753e+00 -5.77373087e-01 -4.26711291e-01 -3.53786826e-01
-8.46408784e-01 1.73789352e-01 2.85828680e-01 4.42617446e-01
2.45482355e-01 -1.45644236e+00 -9.26268756e-01 2.28868410e-01
7.24420026e-02 -1.76913649e-01 2.19106719e-01 1.11564839e+00
-6.66016102e-01 2.98457652e-01 -5.23688555e-01 -4.64389563e-01
-1.26858783e+00 2.42186263e-01 6.09345615e-01 -4.79187220e-01
-2.90065855e-01 5.91102302e-01 -5.19323468e-01 -9.77449417e-01
2.34208167e-01 -2.70339906e-01 -6.51188135e-01 4.25583988e-01
5.99391222e-01 8.75229359e-01 6.82773650e-01 -3.90490353e-01
-3.93003970e-01 3.50180149e-01 2.01695010e-01 4.72623885e-01
1.48430014e+00 2.64321834e-01 -4.08754759e-02 7.16193855e-01
1.07391393e+00 -2.27152735e-01 -7.86168277e-01 3.09780717e-01
1.01400211e-01 -2.47690022e-01 -2.26585239e-01 -9.96211469e-01
-8.25329959e-01 8.86255443e-01 1.05042052e+00 7.02148557e-01
1.33795846e+00 -5.88784873e-01 3.39576602e-01 3.21469933e-01
-8.34058300e-02 -1.02758539e+00 -3.19717705e-01 3.97434831e-01
7.63508737e-01 -1.28404665e+00 3.86117212e-02 -1.12656087e-01
-2.57099003e-01 1.38095164e+00 9.19398844e-01 -4.41202819e-01
3.54009837e-01 5.69770992e-01 2.02775076e-01 8.78096893e-02
-8.20134342e-01 2.09098771e-01 2.39451170e-01 5.31798124e-01
6.73353493e-01 -4.35145348e-02 -7.77071595e-01 5.80156863e-01
-9.43203494e-02 1.56677827e-01 4.52026874e-01 8.86669576e-01
-4.07990932e-01 -1.02773547e+00 -7.34559596e-01 1.10134304e+00
-8.63879204e-01 -5.07401079e-02 -1.56173065e-01 1.12940788e+00
6.86376572e-01 7.95217931e-01 6.91019773e-01 -2.07543150e-01
5.50965428e-01 3.28809291e-01 1.16509847e-01 -2.92579412e-01
-1.23976135e+00 -4.94093567e-01 3.49958986e-01 -2.38108426e-01
-5.35266042e-01 -3.64184082e-01 -7.17554152e-01 -4.65426207e-01
-4.36292976e-01 -6.96274266e-02 5.76715827e-01 8.39499414e-01
2.06560120e-01 7.25092888e-01 6.71870589e-01 -9.00769711e-01
-7.89440870e-01 -1.56922054e+00 -5.95083952e-01 5.81001401e-01
3.58261496e-01 -7.65566409e-01 -7.39331305e-01 -3.38839233e-01] | [8.69035530090332, 4.244235992431641] |
e038945d-23e2-41a7-8315-103cb613fa60 | a-bayesian-treatment-of-real-to-sim-for | 2112.05068 | null | https://arxiv.org/abs/2112.05068v1 | https://arxiv.org/pdf/2112.05068v1.pdf | A Bayesian Treatment of Real-to-Sim for Deformable Object Manipulation | Deformable object manipulation remains a challenging task in robotics research. Conventional techniques for parameter inference and state estimation typically rely on a precise definition of the state space and its dynamics. While this is appropriate for rigid objects and robot states, it is challenging to define the state space of a deformable object and how it evolves in time. In this work, we pose the problem of inferring physical parameters of deformable objects as a probabilistic inference task defined with a simulator. We propose a novel methodology for extracting state information from image sequences via a technique to represent the state of a deformable object as a distribution embedding. This allows to incorporate noisy state observations directly into modern Bayesian simulation-based inference tools in a principled manner. Our experiments confirm that we can estimate posterior distributions of physical properties, such as elasticity, friction and scale of highly deformable objects, such as cloth and ropes. Overall, our method addresses the real-to-sim problem probabilistically and helps to better represent the evolution of the state of deformable objects. | ['Jeannette Bohg', 'Fabio Ramos', 'Dieter Fox', 'Priya Sundaresan', 'Jingyun Yang', 'Rika Antonova'] | 2021-12-09 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [ 1.24113886e-02 -1.17881931e-01 -9.80062559e-02 -6.22876473e-02
-2.29403615e-01 -7.53927410e-01 6.47255003e-01 -2.11838841e-01
-3.15847486e-01 7.75799572e-01 -1.83827907e-01 6.23492673e-02
-3.53195250e-01 -6.66437685e-01 -9.09297585e-01 -1.01030421e+00
-1.41687049e-02 1.06753981e+00 6.47619605e-01 3.21126916e-03
1.80204943e-01 1.00192869e+00 -1.32454813e+00 -5.56449354e-01
2.99989969e-01 5.82163751e-01 3.69720519e-01 1.07587707e+00
3.31592590e-01 2.23955512e-01 -3.20938766e-01 -5.21374196e-02
1.02496952e-01 2.14338303e-02 -6.35694206e-01 9.57753435e-02
1.56955048e-01 -6.54826701e-01 -5.56706011e-01 9.27306533e-01
9.40938890e-02 3.66284698e-01 1.12201500e+00 -1.19582665e+00
-3.25343817e-01 3.59676391e-01 -1.14133596e-01 -1.69114888e-01
3.18560332e-01 2.59182751e-01 6.53524101e-01 -2.77922302e-01
7.37719834e-01 1.56186461e+00 3.43085498e-01 5.31755149e-01
-1.44512367e+00 -1.89693511e-01 1.93259418e-01 -6.34387881e-02
-1.18726122e+00 -1.03765212e-01 8.13546360e-01 -7.98056543e-01
3.81626427e-01 1.87676087e-01 8.45983863e-01 1.20445538e+00
6.02777779e-01 6.75050437e-01 9.15197253e-01 -2.47852266e-01
6.12319529e-01 -2.52604157e-01 6.02206495e-03 7.67250299e-01
5.26508212e-01 2.00866953e-01 -2.38985136e-01 -4.30519998e-01
1.31235254e+00 2.97354590e-02 -6.48704469e-02 -9.02038455e-01
-1.30745399e+00 6.05253518e-01 -8.93590506e-03 -2.86971867e-01
-3.48011047e-01 8.26588035e-01 1.08368047e-01 -3.26822847e-01
1.24296799e-01 1.80873051e-01 -5.19362986e-01 -2.87447721e-01
-4.35610294e-01 8.18973482e-01 1.35177815e+00 9.05405343e-01
5.70220292e-01 -1.28831148e-01 -5.35354353e-02 2.58050352e-01
8.44807029e-01 1.00552893e+00 -1.58026606e-01 -1.36619163e+00
-1.34217218e-01 4.81722280e-02 7.11247146e-01 -8.09111357e-01
-1.32222369e-01 3.95718277e-01 -2.69937605e-01 6.35980248e-01
5.81148267e-01 -1.98931605e-01 -1.24619734e+00 1.71095467e+00
6.92965090e-01 2.25782916e-01 -2.10991442e-01 8.90663683e-01
1.26256034e-01 6.01708651e-01 8.11795518e-02 -5.81631400e-02
1.11924386e+00 -1.30325243e-01 -7.49543786e-01 2.83040069e-02
-1.94804847e-01 -6.76707268e-01 5.31544566e-01 2.64240980e-01
-1.04168499e+00 -3.09945345e-01 -8.62665176e-01 1.48537740e-01
2.01381426e-02 -1.59857348e-02 3.99178207e-01 2.76104093e-01
-6.18553579e-01 9.08813477e-01 -1.81463706e+00 -3.00314158e-01
8.19776133e-02 5.01515925e-01 -1.72691017e-01 2.59025961e-01
-6.91813707e-01 1.32312155e+00 4.53846306e-01 3.10743153e-01
-1.32582521e+00 -3.84016603e-01 -8.51624072e-01 -3.74930829e-01
4.34051007e-01 -8.17869782e-01 1.38558793e+00 6.99592084e-02
-2.19231677e+00 3.83833498e-01 -7.40980133e-02 -1.14742875e-01
6.47656024e-01 -4.89922076e-01 2.23300532e-01 8.22665393e-02
-3.99299026e-01 2.93783844e-01 1.10110462e+00 -1.54192698e+00
-4.69165258e-02 -4.03954268e-01 2.06668407e-01 1.66045293e-01
3.90744746e-01 -3.57433617e-01 -3.27703536e-01 -4.79206413e-01
3.18638057e-01 -1.61550951e+00 -4.19884920e-01 5.14600158e-01
-4.38209265e-01 -1.92875579e-01 1.10315406e+00 -4.61947113e-01
4.25932020e-01 -1.71642709e+00 7.12055266e-01 1.38397828e-01
8.21789652e-02 1.43191323e-01 3.10189784e-01 4.44934815e-01
4.97935772e-01 -2.11609930e-01 -2.95206785e-01 -1.08335875e-01
3.14267159e-01 7.96101749e-01 -4.55353200e-01 9.86959040e-01
3.50859791e-01 8.88916016e-01 -1.11047840e+00 -5.24284363e-01
5.14227808e-01 7.35752165e-01 -4.23185706e-01 3.69795650e-01
-5.90295613e-01 6.12643242e-01 -9.97655928e-01 4.85724181e-01
6.01364911e-01 1.37310401e-01 2.42891490e-01 -4.59585011e-01
-2.36978605e-01 1.28437085e-02 -1.49185860e+00 1.41555929e+00
-3.12962055e-01 2.62020141e-01 2.28222057e-01 -7.92669117e-01
8.88748765e-01 3.13173771e-01 5.26042759e-01 4.67295527e-01
4.38728034e-01 -9.35175270e-02 -1.32435430e-02 -7.82006860e-01
5.98410308e-01 -4.38901037e-01 -1.86475173e-01 4.81161773e-01
-1.69295189e-03 -1.10645723e+00 -1.33528158e-01 -8.43593329e-02
1.09489584e+00 8.03484559e-01 6.26583919e-02 -2.70766616e-01
-4.30937558e-02 -1.75470516e-01 3.26007664e-01 5.92370868e-01
-1.84173416e-02 3.41637403e-01 1.83844998e-01 -1.14979357e-01
-1.27215099e+00 -1.76983762e+00 -3.72432142e-01 4.20138031e-01
4.61206466e-01 9.82404649e-02 -5.95016420e-01 -1.46689326e-01
5.52915871e-01 4.52148438e-01 -7.07120299e-01 -2.43817270e-01
-7.90607393e-01 -6.10895276e-01 5.42335771e-02 5.36624134e-01
-1.10548086e-01 -8.45633566e-01 -8.12731564e-01 4.49318260e-01
9.37752612e-03 -1.22712481e+00 -1.23155020e-01 1.39426291e-02
-9.63967800e-01 -8.79971981e-01 -4.48929727e-01 -3.10929716e-01
6.96943104e-01 -2.18763143e-01 6.75215185e-01 -3.00414324e-01
-6.64980769e-01 8.09556663e-01 -6.71487749e-02 -3.83040130e-01
-8.01371574e-01 -3.43316555e-01 4.00353730e-01 -2.39906713e-01
-4.44132119e-01 -5.75521767e-01 -4.46805120e-01 6.17656291e-01
-8.50545049e-01 -3.25713545e-01 2.35370010e-01 3.80054146e-01
7.87408352e-01 5.01799621e-02 -1.31536126e-01 -4.74190861e-01
4.27684993e-01 -3.24571371e-01 -9.74551976e-01 2.07246557e-01
-2.16256268e-02 4.61454630e-01 2.03763559e-01 -1.12178779e+00
-1.12130976e+00 5.62776506e-01 2.18651533e-01 -5.16931593e-01
-2.14143977e-01 3.20700586e-01 -2.84723174e-02 -3.92840840e-02
3.24129969e-01 -1.03956081e-01 2.03771979e-01 -4.34049875e-01
4.48501021e-01 2.39984125e-01 6.55839562e-01 -1.36378551e+00
1.00819564e+00 9.11848724e-01 5.97269356e-01 -7.56032288e-01
-4.90421325e-01 -2.33339369e-01 -8.75460267e-01 -3.70179623e-01
8.02511990e-01 -5.14202893e-01 -1.28530133e+00 6.16844594e-01
-1.07970047e+00 -6.43482387e-01 -5.12907147e-01 7.59389043e-01
-9.97802794e-01 5.06663620e-01 -5.26077032e-01 -1.17274225e+00
1.72547683e-01 -1.15107191e+00 1.46450698e+00 9.40866545e-02
-3.12164247e-01 -1.03069878e+00 4.74004865e-01 -2.54823379e-02
2.28816614e-01 7.45205760e-01 5.11170626e-01 6.46371245e-02
-8.57229710e-01 -4.69470561e-01 2.66360372e-01 2.22843707e-01
2.52234071e-01 6.18934989e-01 -4.73944873e-01 -3.18726808e-01
1.88348860e-01 -7.58157894e-02 4.50299978e-01 6.70355022e-01
9.04096782e-01 -1.91613868e-01 -5.50697446e-01 1.32197976e-01
1.30174494e+00 -4.83280607e-02 4.45573479e-01 -2.24796817e-01
7.54652143e-01 5.47277153e-01 7.46965528e-01 7.10709810e-01
3.21905971e-01 8.23913455e-01 6.40978754e-01 6.68915689e-01
5.08846231e-02 -1.06713183e-01 4.58033144e-01 6.31667972e-01
-3.93894941e-01 -3.50333750e-01 -8.23204935e-01 4.75649744e-01
-1.85753858e+00 -7.64839292e-01 -1.12862505e-01 2.26023078e+00
1.02949810e+00 4.07957006e-03 -7.72743151e-02 -1.85792848e-01
7.25528002e-01 -2.45253280e-01 -8.49397361e-01 2.80580558e-02
4.85078126e-01 1.70359343e-01 6.54876351e-01 7.07005858e-01
-8.61803591e-01 8.25905144e-01 7.25072908e+00 2.45147958e-01
-1.07073259e+00 -2.60542095e-01 -2.90976912e-01 1.11272864e-01
-1.45097867e-01 2.32644454e-01 -1.13029766e+00 4.05249029e-01
8.90486956e-01 -1.56416088e-01 3.95601988e-01 6.57177210e-01
2.13447198e-01 -3.97328347e-01 -1.18975747e+00 4.28115964e-01
-2.28952348e-01 -8.40192914e-01 -2.15765573e-02 1.67820975e-01
4.95801091e-01 4.83999848e-02 -6.49511218e-02 -1.42754987e-01
9.30000305e-01 -6.18778765e-01 1.05989635e+00 1.04405522e+00
2.95117170e-01 -2.42079645e-01 3.09250027e-01 4.50348914e-01
-9.27567422e-01 2.67518699e-01 -3.28341812e-01 -5.36181107e-02
7.46240556e-01 7.19840765e-01 -9.63121593e-01 1.55854538e-01
3.83310795e-01 5.38647652e-01 1.15817532e-01 9.97625232e-01
-2.17165843e-01 6.88278437e-01 -9.77633774e-01 -1.50445566e-01
-3.85146797e-01 -3.26725006e-01 1.03223717e+00 8.26856613e-01
2.04337701e-01 1.30783737e-01 4.13683176e-01 1.12897873e+00
3.23580116e-01 -7.35172510e-01 -4.23601180e-01 -1.60784006e-01
3.22005808e-01 1.14431810e+00 -8.97096038e-01 -9.95584354e-02
3.12706709e-01 5.98979712e-01 -8.90413823e-04 2.83890575e-01
-1.11216295e+00 3.85216400e-02 8.61930013e-01 1.86491489e-01
5.20571351e-01 -1.20310867e+00 2.96668500e-01 -1.23390055e+00
1.91421546e-02 -2.43775606e-01 -3.44643891e-01 -6.66215181e-01
-1.18151569e+00 -4.27637361e-02 7.16480315e-01 -1.12539887e+00
-3.93525243e-01 -8.08987141e-01 -1.19166903e-01 6.25295997e-01
-1.13887095e+00 -1.11618996e+00 -1.93937674e-01 4.34984386e-01
2.95242280e-01 5.12969971e-01 6.80046618e-01 -3.45109433e-01
-3.16074938e-01 -2.52300590e-01 2.03608692e-01 6.07930236e-02
3.95055354e-01 -1.26965249e+00 2.55073547e-01 5.84998667e-01
-1.80635124e-01 7.05573916e-01 1.39520133e+00 -1.06755471e+00
-2.01004124e+00 -7.66315043e-01 -2.78235465e-01 -8.17827880e-01
1.01775742e+00 -3.21137816e-01 -8.96611273e-01 7.21078753e-01
-5.24972677e-01 4.60847139e-01 3.95247079e-02 -3.10263693e-01
6.11362532e-02 3.35757643e-01 -1.04443312e+00 5.02560496e-01
8.52425337e-01 -5.11238277e-01 -6.35087311e-01 2.87660986e-01
4.64055896e-01 -8.68957102e-01 -1.26815605e+00 6.13398373e-01
8.75517547e-01 -1.07198395e-01 1.11580813e+00 -4.92246866e-01
2.54017748e-02 -6.55194402e-01 -2.01653033e-01 -1.17583621e+00
-8.50555003e-02 -7.26316452e-01 -6.75839245e-01 9.86432135e-01
-2.41435796e-01 -4.48756099e-01 7.76584804e-01 8.59716475e-01
1.22242816e-01 -4.54368889e-01 -1.04792988e+00 -1.08665264e+00
5.51966429e-02 -2.52745539e-01 1.18534200e-01 5.11228263e-01
-5.14835894e-01 -2.11453781e-01 -1.38240516e-01 7.10348189e-01
1.04915762e+00 1.17266960e-01 8.45823646e-01 -1.30999243e+00
-4.57120955e-01 5.79313897e-02 -6.97102547e-01 -9.67065454e-01
4.80716556e-01 -2.90646762e-01 7.69639254e-01 -1.47368479e+00
2.58564711e-01 -6.28009379e-01 1.94223151e-01 6.85450286e-02
1.40568633e-02 -5.47359325e-02 -1.86456740e-02 1.64462671e-01
-2.61689156e-01 5.97621858e-01 1.47432792e+00 4.84917574e-02
1.85403787e-02 1.85399875e-01 3.22382063e-01 9.67628062e-01
5.75521111e-01 -6.55899167e-01 -2.12213963e-01 -3.67704213e-01
1.03754543e-01 2.08841965e-01 7.77347445e-01 -8.21481049e-01
-7.10555837e-02 -7.37134159e-01 1.69300046e-02 -5.50395012e-01
6.76588476e-01 -9.56206083e-01 6.65209413e-01 5.62928557e-01
-1.50023893e-01 -2.97165126e-01 1.40198052e-01 9.86616433e-01
2.95449674e-01 -3.85735303e-01 8.47936749e-01 1.14119761e-02
-3.56306255e-01 4.03584182e-01 -5.84765613e-01 -1.30081713e-01
9.59036648e-01 -1.06866723e-02 -2.17842106e-02 -1.00351036e-01
-1.03217947e+00 -2.24893391e-02 7.84832120e-01 3.28484535e-01
5.41798294e-01 -9.54120994e-01 -5.71230829e-01 -1.27214953e-01
-2.18551725e-01 3.63441288e-01 -6.69971779e-02 4.72239286e-01
-7.29356289e-01 -1.51023835e-01 -2.18145981e-01 -9.24167693e-01
-1.12051070e+00 1.85900673e-01 2.45577097e-01 8.36078543e-03
-5.51265776e-01 5.67234457e-01 -1.45433038e-01 -4.01406556e-01
-1.30877018e-01 -8.56592596e-01 2.50591427e-01 -3.43425930e-01
-1.77616067e-02 4.85993385e-01 -3.63643795e-01 -7.54033804e-01
-2.31757566e-01 8.70758057e-01 1.48120940e-01 -3.86607349e-01
1.40170729e+00 -1.16802141e-01 -2.30782300e-01 9.32875156e-01
8.62321436e-01 -1.65356949e-01 -1.88061607e+00 3.27326395e-02
-2.81123042e-01 -3.68594855e-01 2.41165422e-02 -4.45642859e-01
-6.34218156e-01 5.39304316e-01 3.79007518e-01 1.26312360e-01
1.66206464e-01 3.51321816e-01 7.31270134e-01 5.51054776e-01
6.78612769e-01 -8.87455404e-01 1.90017551e-01 5.63093483e-01
9.53969300e-01 -9.15682316e-01 4.73062009e-01 -4.87181515e-01
-8.38932097e-02 1.23442352e+00 1.71899945e-01 -4.84218270e-01
1.04572463e+00 7.46133327e-01 -3.03201050e-01 -1.69819575e-02
-4.96872008e-01 9.07719806e-02 3.54179889e-01 5.42482495e-01
-1.76029559e-02 3.65260005e-01 2.00980663e-01 -1.70015544e-01
-1.67604685e-01 1.48141414e-01 6.02855623e-01 1.41497803e+00
-4.85501766e-01 -1.19094920e+00 -5.45835972e-01 5.12624756e-02
-2.69914597e-01 5.71649611e-01 4.89317365e-02 7.48748064e-01
-2.24654302e-01 4.83163595e-01 4.84774373e-02 9.42499489e-02
2.65649378e-01 -2.00712353e-01 1.12433660e+00 -6.56574607e-01
2.03153312e-01 -3.10185459e-02 -1.58063203e-01 -4.88336951e-01
-5.89178085e-01 -1.04801106e+00 -1.41289163e+00 -1.05850592e-01
-7.64157653e-01 -2.40687042e-01 1.21350884e+00 1.05206072e+00
-3.17045748e-02 4.30493325e-01 2.21837133e-01 -1.49169016e+00
-1.02649534e+00 -6.72568262e-01 -6.86137378e-01 2.86320299e-01
4.96573955e-01 -1.24325013e+00 -4.73225743e-01 4.83657539e-01] | [5.601956367492676, -0.4974064528942108] |
fbcd5ea8-4540-4fd6-8de2-b000beb6a9ff | knowledge-driven-answer-generation-for | 2104.06892 | null | https://arxiv.org/abs/2104.06892v1 | https://arxiv.org/pdf/2104.06892v1.pdf | Knowledge-driven Answer Generation for Conversational Search | The conversational search paradigm introduces a step change over the traditional search paradigm by allowing users to interact with search agents in a multi-turn and natural fashion. The conversation flows naturally and is usually centered around a target field of knowledge. In this work, we propose a knowledge-driven answer generation approach for open-domain conversational search, where a conversation-wide entities' knowledge graph is used to bias search-answer generation. First, a conversation-specific knowledge graph is extracted from the top passages retrieved with a Transformer-based re-ranker. The entities knowledge-graph is then used to bias a search-answer generator Transformer towards information rich and concise answers. This conversation specific bias is computed by identifying the most relevant passages according to the most salient entities of that particular conversation. Experiments show that the proposed approach successfully exploits entities knowledge along the conversation, and outperforms a set of baselines on the search-answer generation task. | ['João Magalhães', 'David Semedo', 'Rafael Ferreira', 'Mariana Leite'] | 2021-04-14 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [ 2.59398311e-01 6.05747759e-01 -4.84520316e-01 -1.50809869e-01
-1.17097712e+00 -7.45409131e-01 1.14935458e+00 2.12417096e-01
-3.81369978e-01 8.95622194e-01 9.23567235e-01 -1.72620475e-01
-2.32374325e-01 -8.45449030e-01 -2.38437667e-01 -1.59072146e-01
2.09104344e-01 1.13896215e+00 5.04060328e-01 -8.30682456e-01
6.73762202e-01 -1.06951915e-01 -1.41272295e+00 6.36084735e-01
1.17694759e+00 7.62299418e-01 3.49723518e-01 7.82666862e-01
-7.75541723e-01 9.73197103e-01 -7.22274363e-01 -7.03683138e-01
-2.80088753e-01 -7.92170227e-01 -1.73489952e+00 -2.39188492e-01
8.01076517e-02 5.39809950e-02 -1.87599406e-01 6.68711782e-01
5.81030607e-01 6.81561947e-01 6.59097970e-01 -1.03854561e+00
-4.41693872e-01 9.45027649e-01 1.46399468e-01 5.31129062e-01
1.13105977e+00 -1.30119011e-01 1.64628947e+00 -1.07922065e+00
1.02981758e+00 1.47319925e+00 1.08044937e-01 7.13442385e-01
-1.02493382e+00 -4.13581938e-01 2.06833541e-01 3.66462141e-01
-9.92800057e-01 -2.17082307e-01 9.45227802e-01 -2.90118694e-01
1.09679055e+00 6.18296623e-01 7.14112937e-01 1.12887967e+00
-4.87310857e-01 9.38928366e-01 6.29898906e-01 -5.77674508e-01
1.84871573e-02 4.68113422e-01 3.91863465e-01 4.70200151e-01
-2.87744194e-01 -1.95977241e-01 -1.02323949e+00 -7.63695121e-01
1.28196925e-01 -2.76511401e-01 -4.77012515e-01 -2.69793868e-01
-1.20105684e+00 1.06442916e+00 6.17312610e-01 3.25440913e-01
-5.40005267e-01 -4.99004185e-01 3.99023443e-01 4.10126358e-01
5.96884906e-01 1.15270007e+00 -3.80699247e-01 -3.22432935e-01
-6.17880881e-01 8.73635828e-01 1.63038778e+00 9.48235512e-01
8.26027691e-01 -8.56271207e-01 -1.07743406e+00 1.23052728e+00
1.35049760e-01 4.50675189e-01 6.93843901e-01 -8.71620893e-01
7.12388337e-01 9.13347065e-01 5.14280558e-01 -1.29348338e+00
-2.35382970e-02 -2.55988210e-01 -2.73489445e-01 -6.84744596e-01
8.20649713e-02 -3.22188437e-01 -2.49867335e-01 1.57813084e+00
4.48805094e-01 -3.17221820e-01 3.07828784e-01 8.40017974e-01
1.34772384e+00 5.58027506e-01 -1.35355294e-01 -1.65416121e-01
1.50085759e+00 -1.44530940e+00 -7.58504272e-01 -2.36295894e-01
2.63815314e-01 -7.76362300e-01 1.28220344e+00 -2.44017437e-01
-1.00651479e+00 -1.87884524e-01 -5.56257010e-01 -7.14154914e-02
-4.92336661e-01 -2.49879971e-01 1.80530727e-01 1.37695223e-01
-8.21884871e-01 1.23622917e-01 1.75328687e-01 -5.24428487e-01
2.82131042e-03 -1.50134131e-01 3.06353658e-01 1.52513847e-01
-1.82489061e+00 9.24018323e-01 1.45320952e-01 -5.01444995e-01
-4.53314394e-01 -6.53031290e-01 -7.06252158e-01 1.98613659e-01
4.10387099e-01 -1.14281070e+00 1.91937220e+00 -7.01350093e-01
-1.61253202e+00 7.84504473e-01 -6.90263629e-01 -4.68957245e-01
9.38505977e-02 -1.78510368e-01 -2.62571543e-01 3.82747382e-01
4.45567548e-01 6.42745674e-01 6.80725753e-01 -1.19661009e+00
-9.40629423e-01 1.80543482e-01 4.62331444e-01 8.48984718e-01
-2.06081152e-01 1.43626437e-01 -5.33403575e-01 -4.13577646e-01
-3.13350141e-01 -9.03811514e-01 -1.73063144e-01 -8.74508798e-01
-6.62900329e-01 -9.00405467e-01 5.64478219e-01 -4.17156845e-01
1.74956870e+00 -1.43464184e+00 2.48122215e-01 3.33667994e-01
2.10460842e-01 9.27352682e-02 -1.43136725e-01 9.76260006e-01
3.04273725e-01 1.57910898e-01 2.77650177e-01 3.31187621e-04
1.63339376e-01 -2.49286428e-01 -5.60124338e-01 -5.38948774e-01
-1.66909665e-01 1.20582604e+00 -1.54492366e+00 -6.72064126e-01
-4.20938760e-01 -1.12666659e-01 -5.32438576e-01 5.29856980e-01
-7.84473419e-01 4.02390420e-01 -9.18669999e-01 3.65663022e-01
-1.53717667e-01 -5.41699708e-01 -6.02812245e-02 1.26898848e-02
1.89066291e-01 1.11993074e+00 -5.66946983e-01 1.55274022e+00
-8.77757609e-01 5.69433391e-01 -1.40892372e-01 -4.51232225e-01
8.52148890e-01 3.96550536e-01 1.29372373e-01 -7.92803586e-01
-2.24959984e-01 1.44385740e-01 -3.66833895e-01 -6.15622222e-01
8.16901088e-01 1.89935207e-01 -3.58991951e-01 8.73005331e-01
-4.84243184e-02 -4.96508360e-01 6.16721690e-01 8.98954153e-01
1.19083667e+00 -4.83677983e-01 2.06750900e-01 -1.36153623e-01
9.05885935e-01 4.31976557e-01 -2.72578734e-04 1.07274985e+00
1.80687234e-01 2.04805955e-01 3.11909527e-01 -1.96684077e-01
-3.95782173e-01 -8.17860305e-01 4.67106938e-01 1.42152572e+00
2.94494838e-01 -7.80603588e-01 -6.65699720e-01 -9.99808311e-01
3.95317748e-02 8.16480577e-01 -5.19411683e-01 -2.95819610e-01
-4.60757911e-01 1.36600919e-02 3.42850924e-01 -1.17867172e-01
3.59755903e-01 -1.54181445e+00 -3.16394776e-01 1.97534129e-01
-1.14019799e+00 -9.62226093e-01 -1.06167817e+00 -2.16499671e-01
-4.63528544e-01 -1.18401515e+00 -8.47535908e-01 -9.63060200e-01
5.13565004e-01 4.12147075e-01 1.83652985e+00 9.90405753e-02
1.01776540e-01 6.71000898e-01 -6.96772099e-01 -3.76804620e-01
-4.86979425e-01 5.76356769e-01 -4.17835414e-01 1.63548999e-02
5.98693788e-01 -4.29322422e-01 -9.65488374e-01 3.38496476e-01
-4.87122983e-01 -1.00709394e-01 2.56473482e-01 1.02756810e+00
8.81065726e-02 -4.53190148e-01 9.65172291e-01 -9.84888077e-01
1.87528265e+00 -7.49909401e-01 -6.73340037e-02 4.30203110e-01
-7.12692022e-01 3.33321571e-01 2.79929131e-01 -4.00510937e-01
-1.37707460e+00 -4.85481620e-01 2.19109595e-01 2.55464733e-01
1.90430015e-01 7.02696443e-01 2.38618776e-01 1.48984879e-01
1.08716190e+00 3.06983203e-01 -3.28979850e-01 -2.60744870e-01
7.24230766e-01 8.73182595e-01 1.99552849e-01 -6.12633586e-01
6.58725917e-01 1.37756750e-01 -5.75213373e-01 -7.23333359e-01
-1.04969752e+00 -1.06505823e+00 1.49225704e-02 -4.64969486e-01
6.00660264e-01 -5.92280746e-01 -7.86831141e-01 -6.09039068e-02
-1.46677530e+00 -1.47472590e-01 -4.01564240e-01 7.14016482e-02
-4.11964864e-01 8.83942544e-02 -4.24113870e-01 -7.43566692e-01
-8.06380570e-01 -7.40869999e-01 1.01348734e+00 4.96603519e-01
-8.57728839e-01 -9.56919849e-01 6.00592315e-01 7.96863556e-01
5.04289091e-01 -2.49911442e-01 9.88835275e-01 -1.30667675e+00
-6.26478970e-01 -2.60433823e-01 -8.15896690e-02 -1.97038397e-01
3.00404370e-01 -5.15638232e-01 -6.62189841e-01 2.24672452e-01
-3.75688106e-01 -5.68911552e-01 6.96224272e-01 -1.60139352e-01
5.39802074e-01 -7.40879953e-01 -6.57409728e-01 -2.43392870e-01
7.27709591e-01 6.43727183e-02 9.73956585e-02 3.13382834e-01
1.26479253e-01 8.83022070e-01 7.45995820e-01 3.39313626e-01
8.60280573e-01 7.56130755e-01 -1.65421572e-02 2.37690896e-01
-1.83161855e-01 -6.38873935e-01 -2.87858043e-02 7.35868633e-01
2.63204545e-01 -3.37218374e-01 -7.56381750e-01 9.44107652e-01
-1.83153772e+00 -1.29844403e+00 2.31951341e-01 1.85977805e+00
1.45275164e+00 -1.38518764e-02 2.98238963e-01 -3.90481889e-01
6.72870755e-01 3.53463084e-01 -3.43351811e-01 -3.23407829e-01
8.20481330e-02 2.06320375e-01 -2.46812731e-01 9.37092066e-01
-5.70519030e-01 1.11233854e+00 5.97840118e+00 7.66626537e-01
-5.30665159e-01 -6.87047616e-02 2.10233375e-01 2.91443896e-02
-6.65820181e-01 1.05166279e-01 -8.64085555e-01 2.82973051e-01
5.89535296e-01 -1.05734766e+00 3.18957537e-01 8.80614281e-01
6.56177700e-02 -1.11259423e-01 -1.16725588e+00 8.61182213e-01
2.80754209e-01 -1.58839941e+00 3.86057556e-01 -3.29426676e-01
8.14083338e-01 -1.97100759e-01 -2.88054526e-01 6.99106932e-01
7.53058851e-01 -6.59953475e-01 1.50318339e-01 6.91216111e-01
1.20974287e-01 -6.61212683e-01 3.81775469e-01 6.13152444e-01
-1.22620153e+00 -9.55852941e-02 1.31784126e-01 1.72600672e-01
4.14266527e-01 4.51588869e-01 -1.56483793e+00 3.68747443e-01
6.20348930e-01 1.48131862e-01 -2.92658597e-01 9.36581850e-01
-3.74851257e-01 4.89330232e-01 -4.21914496e-02 -9.25186098e-01
3.27190131e-01 -7.53686354e-02 1.05934989e+00 1.59872484e+00
6.99980359e-04 2.09451318e-01 3.18594635e-01 8.13897014e-01
-4.64962631e-01 4.17205185e-01 -6.54172957e-01 4.12449054e-02
8.58889878e-01 1.23090911e+00 -3.02093506e-01 -5.83174407e-01
-2.14914694e-01 1.14869356e+00 4.39287275e-01 4.90155727e-01
-1.29390553e-01 -6.52089357e-01 4.89200354e-01 -3.41335572e-02
2.03851059e-01 4.36513960e-01 3.24775100e-01 -1.09498632e+00
1.50846884e-01 -1.23000681e+00 6.95117950e-01 -7.79745638e-01
-1.43629241e+00 8.31406116e-01 4.92437519e-02 -9.66065288e-01
-1.05669868e+00 2.99370557e-01 -8.69088590e-01 1.08766675e+00
-1.63876128e+00 -7.24062204e-01 -2.95880437e-01 5.87711573e-01
1.08704269e+00 -1.94180861e-01 1.02751803e+00 -9.88094136e-02
1.57080665e-01 3.75797451e-01 -3.22781414e-01 1.12623058e-01
6.89383924e-01 -1.37763381e+00 3.92770857e-01 3.71750683e-01
3.37048650e-01 9.53188479e-01 7.82276571e-01 -7.48043776e-01
-9.04175222e-01 -6.32621288e-01 1.72884881e+00 -6.58874273e-01
7.44357646e-01 -8.13538656e-02 -7.10232556e-01 9.53916535e-02
7.85089850e-01 -7.29924023e-01 8.99249911e-01 4.38034296e-01
-2.88369000e-01 -4.98939818e-03 -8.02086949e-01 9.30645227e-01
9.25964952e-01 -8.46517384e-01 -1.25774086e+00 6.94442451e-01
1.02819872e+00 -4.74747777e-01 -3.18428725e-01 1.96236268e-01
3.01177502e-01 -6.08078718e-01 1.14418268e+00 -7.99595475e-01
1.15746766e-01 -4.80649658e-02 2.74068534e-01 -1.71341026e+00
-4.53820676e-02 -1.29864454e+00 -3.84224802e-01 1.23354650e+00
1.02135336e+00 -3.95206302e-01 6.32192731e-01 5.72491348e-01
1.05582580e-01 -7.04983234e-01 -6.01779401e-01 -4.03270960e-01
-3.18747550e-01 2.14800507e-01 5.81586599e-01 7.67692208e-01
8.38347256e-01 1.19592822e+00 -2.28588805e-02 -2.24055290e-01
1.88688025e-01 5.87036371e-01 8.73490512e-01 -1.40183854e+00
-9.66534317e-02 -7.04833508e-01 4.42860156e-01 -1.51273370e+00
3.36142480e-01 -1.04866469e+00 2.66760647e-01 -1.76557195e+00
2.87613839e-01 -1.62855506e-01 3.06899864e-02 -1.95048735e-01
-6.28359854e-01 -4.59375381e-01 -7.67108286e-03 1.46362141e-01
-1.01776826e+00 5.91919065e-01 1.45266676e+00 -2.13314518e-01
-5.74308574e-01 5.42015851e-01 -9.98639941e-01 4.20603782e-01
5.59548497e-01 -3.70843828e-01 -8.73091042e-01 6.44864291e-02
6.23896360e-01 3.04231256e-01 8.69352147e-02 -3.76136929e-01
8.05896997e-01 -1.32802874e-01 -4.02154863e-01 -5.65792978e-01
3.39475632e-01 -3.39728296e-01 -3.89873475e-01 1.00140885e-01
-1.03387523e+00 5.59989028e-02 -1.76907405e-01 7.32999563e-01
-6.68063641e-01 -2.69281387e-01 6.13998771e-02 -2.94206172e-01
-3.76617193e-01 1.48016065e-01 -4.55099225e-01 8.20803702e-01
4.03699338e-01 4.53244634e-02 -2.41827548e-01 -1.16268134e+00
-5.61125636e-01 6.62949681e-01 -1.59144595e-01 7.99145520e-01
6.19369805e-01 -1.33022463e+00 -7.90129006e-01 -3.41854244e-01
5.51297784e-01 -7.46129677e-02 -5.09499945e-02 4.44311708e-01
1.59740269e-01 9.14857924e-01 5.73671639e-01 -2.06652924e-01
-1.27408564e+00 2.92674720e-01 2.86817163e-01 -8.72486472e-01
-2.84427613e-01 1.20985639e+00 -5.05807996e-02 -5.85177124e-01
5.12409806e-01 -1.90752327e-01 -7.69476116e-01 4.55758572e-01
7.62227297e-01 2.39412889e-01 7.71078328e-03 -4.02080297e-01
-2.17322364e-01 2.01449394e-01 -2.01932281e-01 -6.56778276e-01
8.22661936e-01 -4.33430970e-01 -1.69070885e-01 1.71877816e-01
9.90459442e-01 3.14872056e-01 -4.71628308e-01 -8.43417466e-01
5.00714540e-01 -3.28825384e-01 -4.16329980e-01 -1.08757889e+00
-4.36137617e-01 1.61896035e-01 -2.34902039e-01 7.20420361e-01
7.80006588e-01 5.74089885e-01 9.75513816e-01 8.90229881e-01
2.13687912e-01 -1.11828196e+00 5.93334019e-01 1.00528455e+00
1.30760956e+00 -1.17529702e+00 -4.35653389e-01 -4.17649508e-01
-9.39265549e-01 9.38103497e-01 7.21317649e-01 4.38477218e-01
5.15115082e-01 -3.49205226e-01 2.38059789e-01 -6.65379047e-01
-1.17344534e+00 -6.31694973e-01 6.60542190e-01 5.56178510e-01
4.65464950e-01 -2.65847474e-01 -5.11800945e-01 6.24769390e-01
-5.17807186e-01 -1.60642505e-01 7.17776269e-02 8.33844662e-01
-5.84115565e-01 -1.23263848e+00 -2.74763163e-02 3.49430352e-01
-1.41693309e-01 -4.95634347e-01 -1.23915517e+00 2.19392821e-01
-5.42122662e-01 1.51743209e+00 -2.26744235e-01 -3.08607101e-01
5.55742085e-01 5.32157183e-01 -1.03510134e-01 -8.32618535e-01
-1.04870307e+00 -4.74333405e-01 7.55944729e-01 -4.71640110e-01
-4.11496490e-01 -5.50868750e-01 -8.87092471e-01 8.85009244e-02
-6.15907252e-01 1.15723026e+00 3.16151589e-01 8.08420837e-01
8.04464579e-01 2.65993923e-01 8.41080368e-01 -4.82083529e-01
-4.97301072e-01 -1.22777379e+00 1.74465358e-01 6.69145882e-01
4.63986158e-01 -5.22416115e-01 -4.99548018e-01 -7.83356056e-02] | [12.106024742126465, 7.9031596183776855] |
20fae2f8-bd94-4e72-a1fc-8a172d1e79f0 | cross-attention-is-not-enough-incongruity | 2305.13583 | null | https://arxiv.org/abs/2305.13583v2 | https://arxiv.org/pdf/2305.13583v2.pdf | Cross-Attention is Not Enough: Incongruity-Aware Hierarchical Multimodal Sentiment Analysis and Emotion Recognition | Fusing multiple modalities for affective computing tasks has proven effective for performance improvement. However, how multimodal fusion works is not well understood, and its use in the real world usually results in large model sizes. In this work, on sentiment and emotion analysis, we first analyze how the salient affective information in one modality can be affected by the other in crossmodal attention. We find that inter-modal incongruity exists at the latent level due to crossmodal attention. Based on this finding, we propose a lightweight model via Hierarchical Crossmodal Transformer with Modality Gating (HCT-MG), which determines a primary modality according to its contribution to the target task and then hierarchically incorporates auxiliary modalities to alleviate inter-modal incongruity and reduce information redundancy. The experimental evaluation on three benchmark datasets: CMU-MOSI, CMU-MOSEI, and IEMOCAP verifies the efficacy of our approach, showing that it: 1) achieves better performance than prior work as well as manual selection of the primary modality; 2) can recognize hard samples whose emotions are hard to tell; 3) mitigates the inter-modal incongruity at the latent level when modalities have mismatched affective tendencies; 4) reduces model size to less than 1M parameters while outperforming existing models of similar sizes. | ['Catherine Lai', 'Peter Bell', 'Yuanchao Li', 'Yaoting Wang'] | 2023-05-23 | null | null | null | null | ['multimodal-sentiment-analysis', 'sentiment-analysis', 'multimodal-sentiment-analysis'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing'] | [ 2.25527465e-01 -9.24005955e-02 -6.10041954e-02 -3.72160822e-01
-6.95297062e-01 -3.96200776e-01 5.22680521e-01 2.57044703e-01
-3.24523926e-01 5.37455142e-01 5.05497754e-01 1.48770258e-01
6.39774799e-02 -3.86102796e-01 -3.87904376e-01 -7.70271122e-01
3.67923737e-01 2.48404264e-01 -1.19798847e-01 -4.89130735e-01
1.35640547e-01 -1.87803328e-01 -1.73447049e+00 7.38265216e-01
1.04978716e+00 1.34855378e+00 -1.39768990e-02 1.49262041e-01
-1.58973694e-01 8.55683506e-01 -3.41816694e-01 -6.96809173e-01
-1.67715669e-01 -4.85252649e-01 -8.76094699e-01 -1.12816446e-01
4.83734906e-02 8.40215161e-02 9.73445699e-02 8.95959854e-01
6.72347486e-01 1.11908622e-01 5.82569897e-01 -1.41859925e+00
-4.65979725e-01 6.73147619e-01 -6.99966013e-01 -5.63149229e-02
3.03311527e-01 -8.55774656e-02 1.16515386e+00 -1.04142082e+00
2.87087947e-01 1.32028282e+00 6.63650215e-01 5.47091365e-01
-1.12492681e+00 -7.71639466e-01 3.85014713e-01 4.19899404e-01
-1.22966528e+00 -5.07492244e-01 1.02641380e+00 -1.80495113e-01
9.76329029e-01 2.67124176e-01 4.77647960e-01 1.29461026e+00
1.82976335e-01 9.63166118e-01 1.31261945e+00 -4.01262522e-01
9.52318162e-02 2.52945781e-01 2.06967190e-01 3.94105107e-01
-4.12123799e-01 -4.23748791e-01 -9.43675995e-01 -1.03850514e-01
1.04011141e-01 -7.87020549e-02 -2.26788610e-01 3.10475398e-02
-1.21800661e+00 7.61381030e-01 3.28933358e-01 4.24429864e-01
-5.72417378e-01 -2.68619061e-01 7.35005617e-01 2.89901376e-01
4.09015775e-01 2.76910871e-01 -5.04795194e-01 -2.75304943e-01
-4.91363347e-01 -2.26123586e-01 4.32957292e-01 6.11675918e-01
6.39537632e-01 -8.98843482e-02 -3.60011280e-01 1.20694923e+00
2.04326585e-01 4.57869172e-01 5.75166941e-01 -7.14225948e-01
4.97912705e-01 8.85088682e-01 -6.93206638e-02 -9.68904495e-01
-6.72963738e-01 -9.46169421e-02 -1.07952785e+00 -2.87525386e-01
-1.22855529e-01 -2.02837691e-01 -7.14454472e-01 2.26406741e+00
1.15110911e-01 -2.08225593e-01 3.04185450e-01 1.02352357e+00
1.15282512e+00 5.47831655e-01 5.47005832e-01 -3.27523112e-01
1.68177414e+00 -8.00178528e-01 -1.06215894e+00 -5.82332790e-01
4.22108084e-01 -8.17727983e-01 1.18235159e+00 2.88135529e-01
-1.00994623e+00 -5.41112304e-01 -9.22242522e-01 -2.93532815e-02
-4.52324718e-01 1.18450776e-01 8.88341308e-01 5.02828717e-01
-8.26009631e-01 -4.01278697e-02 -4.34207857e-01 -3.36930931e-01
3.01925451e-01 5.32836258e-01 -5.25984406e-01 1.03864066e-01
-1.66703498e+00 9.86139774e-01 4.97609735e-01 3.37410301e-01
-2.76420325e-01 -4.69764262e-01 -9.62755203e-01 2.08941355e-01
3.52597982e-01 -7.09556699e-01 8.83114874e-01 -1.46684325e+00
-1.46447158e+00 6.75505161e-01 -4.34537262e-01 7.15308189e-02
-1.43291011e-01 -1.74349174e-01 -6.20992482e-01 1.27696395e-01
-1.48129597e-01 1.04742897e+00 8.07937741e-01 -1.22522974e+00
-6.03444278e-01 -5.80480218e-01 -2.55977344e-02 6.67237401e-01
-9.02542651e-01 5.28559685e-02 -6.52067482e-01 -4.80326384e-01
1.25438645e-01 -9.46101785e-01 2.48337444e-02 -7.38612652e-01
-3.15929264e-01 -3.27165872e-01 6.39553666e-01 -4.17686731e-01
1.36907744e+00 -2.13974857e+00 4.23922181e-01 1.62132338e-01
1.51538104e-01 4.56399731e-02 -1.75983831e-01 4.46028650e-01
-7.28783384e-02 4.04149033e-02 -1.08546853e-01 -6.09616876e-01
1.75781876e-01 9.33863372e-02 -2.44191200e-01 3.27733122e-02
2.52808660e-01 1.03370500e+00 -5.82300544e-01 -6.25315487e-01
2.09407866e-01 6.26215696e-01 -3.99576455e-01 2.64870286e-01
2.89288998e-01 3.04567039e-01 -8.73933062e-02 8.99778545e-01
5.71126759e-01 -3.25745523e-01 4.13529277e-01 -7.30841160e-01
1.61960736e-01 5.29023334e-02 -9.67644453e-01 1.54734087e+00
-5.74772835e-01 4.22984660e-01 1.43278047e-01 -7.94123590e-01
8.94694388e-01 4.64529276e-01 4.57764506e-01 -1.11320031e+00
3.86326462e-01 -5.88402785e-02 -4.15344574e-02 -5.16681969e-01
5.64873099e-01 -4.17229682e-01 -4.54590708e-01 1.86217591e-01
2.07647339e-01 2.72015423e-01 1.52101338e-01 2.06878141e-01
7.71764398e-01 -1.70173511e-01 1.27312064e-01 -7.11463466e-02
6.55424297e-01 -4.26888496e-01 7.49375880e-01 4.81852204e-01
-3.59840780e-01 2.97891170e-01 5.91499925e-01 3.16723026e-02
-4.00154710e-01 -7.60112584e-01 6.35642260e-02 1.68504333e+00
4.39221084e-01 -4.46413934e-01 -5.14348507e-01 -5.86147666e-01
-3.99086237e-01 6.03135109e-01 -8.46803427e-01 -4.73569483e-01
-3.39547880e-02 -1.18748140e+00 4.31170583e-01 7.33885705e-01
5.07786870e-01 -1.20623398e+00 -4.89925116e-01 -9.41328034e-02
-8.35030973e-01 -1.26676631e+00 -2.55188286e-01 4.02473241e-01
-5.68608522e-01 -6.32255018e-01 -3.09946388e-01 -6.62729800e-01
5.02878606e-01 1.16103739e-01 1.24437332e+00 -2.82298863e-01
3.11051369e-01 4.57127750e-01 -4.35148478e-01 -3.03675324e-01
-9.67663452e-02 5.03721349e-02 -4.81494255e-02 3.98475140e-01
6.12252772e-01 -3.72196525e-01 -5.06588817e-01 3.33271921e-01
-1.00873184e+00 3.19166541e-01 6.93676174e-01 1.00129354e+00
4.85576808e-01 4.28941585e-02 8.57379377e-01 -7.44774461e-01
6.86120868e-01 -5.90238333e-01 1.03927970e-01 3.59208375e-01
-6.40669048e-01 -2.19539732e-01 4.85870481e-01 -4.60136443e-01
-1.48673213e+00 -2.85800193e-02 -1.53746203e-01 -4.36981380e-01
-1.32019371e-01 7.50795305e-01 -4.55582172e-01 2.29871288e-01
1.17534406e-01 1.00556863e-02 -1.40817598e-01 -1.16112560e-01
2.67488182e-01 7.32160807e-01 3.97505015e-01 -4.56107706e-01
1.83434188e-01 5.52800417e-01 -2.62672931e-01 -5.28658986e-01
-8.73953521e-01 -3.27095598e-01 -4.21182364e-01 -4.62949514e-01
9.12491143e-01 -1.07245386e+00 -1.06571400e+00 4.63589311e-01
-8.83615792e-01 -7.97632709e-02 1.59583971e-01 4.59117919e-01
-1.35494024e-01 2.09582493e-01 -8.48747075e-01 -9.03824270e-01
-5.78397632e-01 -1.22055018e+00 1.12508380e+00 3.60463649e-01
-4.88657206e-01 -1.11012304e+00 -1.16707608e-01 7.38676429e-01
4.27866608e-01 -7.70224780e-02 1.15546334e+00 -5.65656304e-01
1.97935000e-01 -4.00379635e-02 -2.66441494e-01 2.51027972e-01
-1.28512457e-01 -1.65468588e-01 -1.27322710e+00 -8.80051851e-02
-9.27308351e-02 -7.12035537e-01 9.25601125e-01 2.15702936e-01
9.64197338e-01 5.35702333e-02 -2.62013096e-02 1.05891660e-01
1.16410768e+00 5.52712940e-02 7.43039727e-01 8.76363590e-02
7.97979712e-01 8.73815656e-01 6.97206378e-01 4.34126198e-01
9.09010649e-01 6.57702863e-01 5.92068076e-01 -3.92617613e-01
2.12519228e-01 1.35419406e-02 6.03820860e-01 1.25864005e+00
1.33441776e-01 -2.21929327e-01 -7.05586016e-01 5.92837334e-01
-2.09121752e+00 -8.08796763e-01 -7.43717477e-02 2.03666139e+00
8.79940033e-01 -4.30296324e-02 4.92877290e-02 2.59623695e-02
5.48467219e-01 5.84050529e-02 -4.32425052e-01 -5.90102196e-01
-5.97581565e-01 -5.99366911e-02 -9.45192501e-02 2.68721223e-01
-1.22567797e+00 8.12977374e-01 5.41219950e+00 8.34383249e-01
-1.27685857e+00 2.60287642e-01 8.82415891e-01 -1.59900859e-01
-4.41535383e-01 -1.83407351e-01 -4.90585566e-01 6.03450716e-01
7.23277271e-01 3.25624436e-01 3.15742850e-01 3.16990584e-01
-8.35450739e-02 -2.97670484e-01 -9.60249662e-01 1.24559534e+00
3.00866336e-01 -6.41366184e-01 -2.53615938e-02 -2.86041319e-01
6.17775440e-01 -1.54742613e-01 2.04268306e-01 7.38753438e-01
-1.01701692e-01 -9.76183355e-01 7.46166885e-01 6.20399773e-01
3.84905547e-01 -1.03312087e+00 1.16821456e+00 2.12047532e-01
-1.12857449e+00 -1.52499586e-01 -5.54113016e-02 -6.52975664e-02
4.13196608e-02 5.00647902e-01 -3.48472327e-01 5.80832958e-01
8.26003730e-01 4.83341813e-01 -8.10309768e-01 2.94696391e-01
-7.38557577e-02 6.25087321e-01 -2.24544868e-01 1.20828949e-01
6.12753183e-02 -1.67089880e-01 1.71480581e-01 1.33021975e+00
1.01024531e-01 -7.72018880e-02 -4.85134870e-03 5.31011760e-01
-9.94244516e-02 3.89079094e-01 -2.25902289e-01 -1.06269814e-01
4.19350654e-01 1.52934670e+00 -6.23638391e-01 -2.54712224e-01
-6.08081043e-01 1.08641601e+00 3.45159382e-01 3.21091652e-01
-1.00496972e+00 -1.62559479e-01 3.26501876e-01 -6.15538001e-01
1.55930638e-01 4.57205325e-01 -5.95332384e-01 -1.15984082e+00
3.05198636e-02 -9.59551275e-01 7.27708995e-01 -9.80280876e-01
-1.48742950e+00 7.47942030e-01 -2.82436341e-01 -1.10887110e+00
-1.48710117e-01 -2.98473269e-01 -3.78977329e-01 8.13724637e-01
-1.26789916e+00 -1.50527918e+00 -2.64983416e-01 7.25644708e-01
3.02563041e-01 9.18491110e-02 1.03583360e+00 5.73726892e-01
-8.10982049e-01 7.45401621e-01 -4.54807729e-01 -1.77886710e-01
9.80928421e-01 -1.09439480e+00 -5.74071944e-01 5.89760602e-01
-1.85658112e-01 5.61842084e-01 5.25170445e-01 -3.93644601e-01
-1.43339407e+00 -6.66662097e-01 1.01732183e+00 -3.62466782e-01
6.06392086e-01 -2.53936410e-01 -1.03233409e+00 3.61816257e-01
6.91810668e-01 -3.91574800e-01 1.23062360e+00 6.19504392e-01
-4.65736687e-01 -2.61321634e-01 -9.53023732e-01 6.43293679e-01
5.34124911e-01 -7.27684438e-01 -4.63191301e-01 -5.84291033e-02
3.59765351e-01 -1.52693376e-01 -1.06307197e+00 8.27062964e-01
6.45941854e-01 -1.01878619e+00 6.95245147e-01 -4.28577423e-01
7.52311170e-01 -9.02917162e-02 -3.55097651e-01 -1.33303797e+00
-2.75942504e-01 -2.99638778e-01 -9.91109237e-02 1.54523766e+00
4.35030490e-01 -4.05189931e-01 2.06470624e-01 7.48823225e-01
-2.74681356e-02 -1.02899051e+00 -8.69913518e-01 -9.62651148e-02
-2.21023113e-01 -4.99357462e-01 4.77858126e-01 1.19277489e+00
4.61815745e-01 9.75335717e-01 -5.92473507e-01 3.99994440e-02
1.79176465e-01 4.24586564e-01 3.69566768e-01 -1.04302788e+00
-2.87586153e-02 -7.06395030e-01 1.95013154e-02 -5.72286129e-01
4.18919057e-01 -6.30173266e-01 8.54928270e-02 -1.32130980e+00
5.79298437e-01 -1.47699580e-01 -8.34449053e-01 9.29597199e-01
-5.65761805e-01 5.84640861e-01 2.79530704e-01 1.31797101e-02
-1.04899967e+00 8.58110845e-01 8.82900476e-01 -7.58248568e-02
-2.21015394e-01 -4.29205030e-01 -1.01394975e+00 8.75441670e-01
6.62731886e-01 -2.68668518e-05 -4.20022368e-01 -1.32118285e-01
6.66825354e-01 5.32831661e-02 1.56440973e-01 -7.77146041e-01
-2.67597567e-02 -1.32263247e-02 3.47298145e-01 -5.19202292e-01
6.94757402e-01 -9.59102511e-01 -6.42506182e-02 -8.51161778e-02
-3.86617810e-01 4.36920151e-02 4.60479707e-01 2.72067547e-01
-4.85414326e-01 -3.45690101e-02 6.15635037e-01 3.52175593e-01
-9.77457166e-01 -7.10122809e-02 -4.54550058e-01 -9.37123597e-03
7.33282864e-01 1.38787746e-01 -3.26629519e-01 -5.62664747e-01
-6.85629368e-01 3.94572526e-01 2.10472137e-01 7.20800102e-01
4.94574994e-01 -1.60814953e+00 -4.78420943e-01 4.98579023e-03
4.47231889e-01 -7.12722719e-01 7.98398614e-01 1.43552482e+00
3.34000587e-01 3.04911911e-01 -2.12215632e-01 -5.95006168e-01
-1.50790143e+00 4.45964724e-01 2.22353369e-01 -3.60391885e-01
1.13021992e-01 8.31028700e-01 4.06829029e-01 -3.17984790e-01
2.21136719e-01 1.12194330e-01 -6.32221997e-01 7.32426763e-01
3.03839952e-01 1.63532868e-01 2.14512184e-01 -1.12937987e+00
-5.89728355e-01 4.20854330e-01 -8.49579275e-02 -1.75165415e-01
1.03611720e+00 -4.84741867e-01 -4.60680068e-01 7.89341092e-01
1.10147822e+00 -3.31623964e-02 -8.55233252e-01 -2.93213487e-01
-3.64145994e-01 -5.13812825e-02 1.57767668e-01 -1.04059029e+00
-1.08567703e+00 1.00216687e+00 6.99782193e-01 1.85038075e-01
1.70311952e+00 7.55528733e-02 8.15058351e-01 1.49966985e-01
2.36849077e-02 -1.53094852e+00 1.69462472e-01 6.14767313e-01
7.07809806e-01 -1.46272600e+00 -2.67334253e-01 -2.77648389e-01
-1.42228401e+00 7.83843160e-01 8.88066292e-01 5.69512725e-01
5.41924953e-01 1.58551738e-01 2.52697676e-01 -3.72449458e-01
-1.00006139e+00 -4.07595038e-01 4.96461391e-01 8.43381062e-02
5.72180748e-01 1.76317498e-01 -2.19144568e-01 1.15142870e+00
2.02924937e-01 -2.74902046e-01 -7.11895004e-02 7.58657277e-01
-3.97436060e-02 -7.49255896e-01 -3.97833973e-01 2.94642895e-01
-4.59849834e-01 -2.74208486e-01 -6.29965782e-01 4.52392995e-01
3.50368381e-01 1.20044446e+00 9.35562998e-02 -6.25173032e-01
3.05718154e-01 4.00817156e-01 3.50714535e-01 -6.47257715e-02
-8.67376268e-01 4.06719476e-01 1.08990975e-01 -5.10361493e-01
-7.19857216e-01 -4.81719524e-01 -1.24362421e+00 -1.02185443e-01
-2.31871545e-01 1.88093111e-02 4.51503307e-01 1.11137199e+00
6.98094964e-01 6.73053324e-01 6.01917088e-01 -7.95848787e-01
-6.11802936e-02 -1.14039648e+00 -3.02017808e-01 8.31469476e-01
-3.43555920e-02 -8.28239918e-01 -3.52858484e-01 -7.53432661e-02] | [13.183351516723633, 5.165999412536621] |
bdd726f9-e8db-4bbf-be91-963d3daa6837 | an-order-invariant-and-interpretable | 2302.06243 | null | https://arxiv.org/abs/2302.06243v1 | https://arxiv.org/pdf/2302.06243v1.pdf | An Order-Invariant and Interpretable Hierarchical Dilated Convolution Neural Network for Chemical Fault Detection and Diagnosis | Fault detection and diagnosis is significant for reducing maintenance costs and improving health and safety in chemical processes. Convolution neural network (CNN) is a popular deep learning algorithm with many successful applications in chemical fault detection and diagnosis tasks. However, convolution layers in CNN are very sensitive to the order of features, which can lead to instability in the processing of tabular data. Optimal order of features result in better performance of CNN models but it is expensive to seek such optimal order. In addition, because of the encapsulation mechanism of feature extraction, most CNN models are opaque and have poor interpretability, thus failing to identify root-cause features without human supervision. These difficulties inevitably limit the performance and credibility of CNN methods. In this paper, we propose an order-invariant and interpretable hierarchical dilated convolution neural network (HDLCNN), which is composed by feature clustering, dilated convolution and the shapley additive explanations (SHAP) method. The novelty of HDLCNN lies in its capability of processing tabular data with features of arbitrary order without seeking the optimal order, due to the ability to agglomerate correlated features of feature clustering and the large receptive field of dilated convolution. Then, the proposed method provides interpretability by including the SHAP values to quantify feature contribution. Therefore, the root-cause features can be identified as the features with the highest contribution. Computational experiments are conducted on the Tennessee Eastman chemical process benchmark dataset. Compared with the other methods, the proposed HDLCNN-SHAP method achieves better performance on processing tabular data with features of arbitrary order, detecting faults, and identifying the root-cause features. | ['Hongwei Wang', 'Min Wang', 'Peng Peng', 'Mengxuan Li'] | 2023-02-13 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 2.19089314e-01 -1.30437806e-01 3.17489058e-01 -1.57341763e-01
2.69593596e-01 -3.53466958e-01 1.47068188e-01 2.76501924e-01
2.88713863e-03 5.65034926e-01 -8.48395228e-02 -3.57136935e-01
-7.96636105e-01 -8.65736604e-01 -5.19276917e-01 -9.14242566e-01
-8.22912231e-02 2.65841544e-01 -1.37436688e-01 -3.74936238e-02
4.17710960e-01 7.36083210e-01 -1.57962513e+00 7.10054159e-01
1.08252740e+00 1.30450988e+00 2.35622928e-01 1.76860407e-01
-1.44240022e-01 8.21934283e-01 -6.42549753e-01 -9.25019942e-03
1.45267472e-01 -4.62942392e-01 -9.41334844e-01 1.35135576e-01
-2.53656894e-01 -1.20106354e-01 -7.40271509e-02 1.15941954e+00
3.42723340e-01 -7.23367848e-04 7.52350211e-01 -1.35395324e+00
-8.05925846e-01 5.26693225e-01 -2.40269601e-01 1.63755998e-01
1.55068608e-02 7.99453035e-02 7.99760699e-01 -9.69464421e-01
9.12321582e-02 1.43316901e+00 6.81995153e-01 2.35457003e-01
-8.76815140e-01 -5.42691827e-01 -1.98373795e-02 2.81403393e-01
-1.37631738e+00 4.00504582e-02 6.98813438e-01 -5.22072792e-01
1.10659754e+00 3.78805220e-01 4.52541292e-01 5.08119643e-01
6.45554423e-01 5.30044496e-01 7.65386105e-01 -1.04141667e-01
3.33178103e-01 -1.59838527e-01 9.65694785e-02 8.45267951e-01
5.63104331e-01 4.43017052e-05 -3.39547187e-01 -5.70094362e-02
8.91852677e-01 5.99372506e-01 -4.34710175e-01 9.20724422e-02
-1.04564345e+00 7.78764069e-01 8.83210719e-01 3.45934182e-01
-5.59349000e-01 8.23970661e-02 4.15201783e-01 3.49006921e-01
1.84710220e-01 8.27954590e-01 -6.54505491e-01 3.18228275e-01
-4.57866162e-01 -3.71180587e-02 3.91009927e-01 5.74442625e-01
6.00169599e-01 1.82954282e-01 -2.02390462e-01 5.17984450e-01
3.07817876e-01 7.22528026e-02 5.62818408e-01 -6.70559287e-01
2.23980889e-01 1.31902075e+00 -3.38248312e-01 -1.16493785e+00
-4.68226761e-01 -5.61151326e-01 -1.26066446e+00 2.51418918e-01
1.20127723e-01 2.69092657e-02 -1.04465401e+00 1.05555832e+00
-6.19075932e-02 -2.57122487e-01 2.94710904e-01 8.44222069e-01
9.05455291e-01 6.34354591e-01 -3.91655304e-02 -4.16407809e-02
1.63709342e+00 -6.00692034e-01 -9.82403994e-01 2.16345981e-01
5.13071060e-01 -6.52059674e-01 7.30094731e-01 6.53830051e-01
-5.55604696e-01 -6.33345246e-01 -1.16844928e+00 9.35723633e-02
-7.38205791e-01 3.42520833e-01 8.75154734e-01 2.97969937e-01
-6.76389575e-01 1.03638244e+00 -5.16026139e-01 -9.66821387e-02
7.10832298e-01 7.76805520e-01 -3.17025721e-01 -2.03046516e-01
-1.26742709e+00 6.63982511e-01 7.53391147e-01 5.77800870e-01
-6.78688526e-01 -6.43677592e-01 -7.46287107e-01 4.09463853e-01
2.54112720e-01 -5.54792702e-01 9.22912776e-01 -9.88970637e-01
-9.63569105e-01 -2.13392511e-01 -5.91047481e-02 -2.98632503e-01
2.46085152e-01 -5.14486320e-02 -4.43022877e-01 2.59758353e-01
8.49948451e-02 4.69422519e-01 6.28544390e-01 -1.07474267e+00
-6.13971591e-01 -3.40058565e-01 -1.30935803e-01 7.20194802e-02
-4.67326134e-01 -2.36633226e-01 3.53850015e-02 -4.14894760e-01
5.87160468e-01 -4.73913074e-01 -1.71515554e-01 -1.40384004e-01
-5.79820633e-01 -5.35964787e-01 1.26675820e+00 -5.46751916e-01
1.03161597e+00 -2.20292377e+00 -1.91094592e-01 3.38060081e-01
5.79194248e-01 7.51499683e-02 2.13899687e-01 2.84811527e-01
-4.37521756e-01 2.52123237e-01 -3.90235782e-01 3.54707152e-01
-1.63997814e-01 2.49421269e-01 1.35348976e-01 3.14834774e-01
8.72323692e-01 7.25497484e-01 -8.68303835e-01 -2.24223390e-01
3.08754414e-01 4.27568883e-01 -3.26233178e-01 -8.10173377e-02
-7.68239871e-02 1.83282480e-01 -6.63568258e-01 9.60179627e-01
6.94009840e-01 -4.00276005e-01 -5.41424043e-02 -6.91004455e-01
-1.39567435e-01 -7.11528659e-02 -1.20841491e+00 8.02271128e-01
3.63848954e-02 4.42603886e-01 -2.70706266e-01 -1.02359796e+00
1.04286778e+00 5.73933661e-01 5.82947552e-01 -6.14711761e-01
3.94469380e-01 3.58737350e-01 4.04038996e-01 -7.53294289e-01
2.28911832e-01 -8.54700804e-02 2.55550772e-01 1.69380113e-01
-1.17600061e-01 2.23003104e-01 1.85721695e-01 -1.04182772e-02
1.16498744e+00 -4.41609889e-01 1.52166188e-01 -4.75446433e-01
6.60848141e-01 -2.75537698e-03 6.16702378e-01 2.97217041e-01
1.04159787e-01 5.51614344e-01 7.09494889e-01 -8.94264758e-01
-8.99355471e-01 -6.43654704e-01 -3.72605622e-01 3.42925638e-01
1.27191380e-01 -1.65733755e-01 -6.38300300e-01 -7.78683245e-01
2.90880591e-01 3.25234294e-01 -7.14997053e-01 -5.00411808e-01
-1.53671026e-01 -9.16891754e-01 5.94722450e-01 7.97848344e-01
8.80508542e-01 -1.29856575e+00 -4.69325960e-01 3.14363867e-01
-1.07958101e-01 -7.24111855e-01 3.91836949e-02 7.79429615e-01
-9.31783020e-01 -1.50569701e+00 -2.70536751e-01 -8.08689833e-01
1.25916374e+00 1.48706853e-01 7.16591716e-01 5.13007343e-01
-3.87829363e-01 -3.36485445e-01 -3.06002051e-01 -5.45106947e-01
-3.19068432e-01 -2.57172227e-01 1.15556726e-02 1.14498390e-02
4.31202680e-01 -8.02809745e-03 -5.86162984e-01 2.79631674e-01
-1.26011324e+00 -1.25499174e-01 8.81689072e-01 1.07654226e+00
2.96038717e-01 1.06707668e+00 7.02380896e-01 -8.90821457e-01
8.85989726e-01 -4.22397107e-01 -3.34991872e-01 2.71653116e-01
-9.07565057e-01 3.14379364e-01 1.04892492e+00 -1.33072287e-01
-8.65019262e-01 1.67807385e-01 1.76522151e-01 -4.28416580e-01
-1.58694118e-01 7.87832320e-01 -2.82746166e-01 -1.17856599e-01
5.61427474e-01 4.89099249e-02 2.48495579e-01 -3.39708924e-01
-2.83592165e-01 5.97957551e-01 2.60264933e-01 -4.24002647e-01
5.95634758e-01 1.38917655e-01 3.98207903e-01 -5.15899241e-01
-2.78303087e-01 -3.24472547e-01 -6.30866230e-01 -1.46814093e-01
9.54707026e-01 -6.14275634e-01 -1.06069505e+00 5.28646469e-01
-1.26985478e+00 3.70276123e-01 4.36038449e-02 3.75184059e-01
6.09128214e-02 4.96763200e-01 -8.35076690e-01 -7.80669391e-01
-5.13850152e-01 -1.43326628e+00 8.21676373e-01 2.10111424e-01
-2.09388688e-01 -8.15377295e-01 -8.67160201e-01 1.62385702e-01
2.97460288e-01 4.34837341e-01 1.48652124e+00 -5.57029068e-01
-4.97546315e-01 -3.26148927e-01 -4.66427833e-01 5.18225193e-01
5.55601776e-01 1.77185535e-01 -9.37312722e-01 -1.63626164e-01
1.19483829e-01 -5.95470145e-03 8.76132905e-01 4.53606129e-01
1.42329764e+00 -4.51684088e-01 -2.32972458e-01 3.73995095e-01
1.49259293e+00 7.48814940e-01 6.95698857e-01 3.83793384e-01
8.41708720e-01 7.05682516e-01 4.53289747e-01 4.50720906e-01
-2.10911661e-01 -1.05302244e-01 8.34841907e-01 -5.43605328e-01
2.66195118e-01 1.76650479e-01 1.45149365e-01 5.91564953e-01
-2.01625571e-01 -2.05664456e-01 -7.57567346e-01 3.82015496e-01
-1.82541704e+00 -7.04098821e-01 -5.53384542e-01 1.70619297e+00
3.83138478e-01 4.72007282e-02 -4.50040460e-01 7.87337899e-01
8.06143701e-01 -5.59729695e-01 -6.59241199e-01 -4.15012985e-01
-1.60631672e-01 1.10176355e-01 5.04217625e-01 1.96206737e-02
-1.05465853e+00 3.32610816e-01 5.40411568e+00 6.82498038e-01
-1.09533441e+00 -3.76327574e-01 8.95706713e-01 3.35359514e-01
-9.13266093e-02 -2.38917291e-01 -3.68952751e-01 4.47481871e-01
5.17287672e-01 1.66855246e-01 3.60325277e-01 7.33939886e-01
3.29753816e-01 -8.76455083e-02 -1.25090826e+00 7.64810324e-01
-1.83566734e-01 -1.31400228e+00 3.67190808e-01 1.41810611e-01
6.98690772e-01 -5.46237350e-01 7.80581236e-02 -1.22625656e-01
2.76040821e-03 -1.45866287e+00 5.69519341e-01 3.03409100e-01
5.29070735e-01 -1.09150362e+00 1.36779559e+00 -3.91373644e-03
-1.18713045e+00 -5.85151494e-01 -5.99962294e-01 -2.60888249e-01
-3.60656589e-01 7.82118559e-01 -9.25570667e-01 9.50525522e-01
9.05025542e-01 6.51417971e-01 -5.43961942e-01 9.17001307e-01
4.93522733e-02 2.77540445e-01 -9.21201482e-02 -1.39705569e-01
3.95252973e-01 -2.61022244e-02 -4.78779376e-02 8.16400409e-01
5.10146558e-01 6.90042228e-02 -5.96273108e-04 1.09457028e+00
-7.85466935e-03 -7.17672035e-02 -5.90785980e-01 -1.62023455e-01
3.68661821e-01 1.21440208e+00 -1.13460815e+00 -2.47951150e-01
-1.88356161e-01 7.42106259e-01 -1.50661632e-01 3.03792119e-01
-6.04405642e-01 -6.07366860e-01 5.28059661e-01 1.04585774e-01
2.61427462e-01 1.66171789e-01 -7.03230798e-01 -4.02189314e-01
-3.91741581e-02 -8.93477619e-01 3.95977408e-01 -6.10145926e-01
-1.39966035e+00 8.16860795e-01 -2.72746742e-01 -1.29470372e+00
2.90518522e-01 -1.11210525e+00 -6.59574926e-01 8.94445181e-01
-1.30335259e+00 -7.36153901e-01 -4.41809863e-01 6.25793219e-01
5.78757167e-01 -2.36319408e-01 6.68438673e-01 3.21532190e-01
-8.25142086e-01 2.33499780e-01 2.03795075e-01 2.18575373e-01
2.99094766e-01 -1.25762832e+00 -9.55448747e-02 6.28461003e-01
-5.24422765e-01 8.07097852e-01 5.06037891e-01 -7.76620448e-01
-1.25244141e+00 -1.33466792e+00 7.40129113e-01 -8.48335028e-03
3.12555254e-01 -5.50034195e-02 -9.72929776e-01 1.52540579e-01
9.71410573e-02 -6.92373216e-02 5.64860523e-01 -1.78596705e-01
4.90782633e-02 -1.52801424e-01 -1.43628764e+00 4.22136903e-01
5.46266496e-01 -2.56719768e-01 -2.71579057e-01 3.91573071e-01
7.01178193e-01 -5.23588285e-02 -9.65904534e-01 7.15854883e-01
2.53803194e-01 -7.30471075e-01 7.62216330e-01 -3.42760026e-01
8.05121541e-01 -8.36832106e-01 1.99428752e-01 -1.07477367e+00
-7.01828003e-01 2.67727226e-02 1.32890612e-01 1.03831184e+00
5.23604393e-01 -8.07624698e-01 4.97737139e-01 5.78822315e-01
-5.67281961e-01 -9.41604912e-01 -7.06566155e-01 -6.86892152e-01
-3.45313936e-01 2.50596292e-02 8.95320356e-01 1.08095014e+00
-1.07120752e-01 1.66246220e-01 -7.82217830e-03 4.07330066e-01
4.58074868e-01 -6.56216219e-02 -1.56725980e-02 -1.54629529e+00
8.71983021e-02 -4.85014737e-01 -4.95393157e-01 -3.21361959e-01
-3.06659132e-01 -7.81123102e-01 2.12880716e-01 -1.60267091e+00
7.99274594e-02 -2.68338650e-01 -6.11970007e-01 8.15232158e-01
-1.61755428e-01 1.12526990e-01 -1.07035570e-01 2.62800515e-01
-2.55996853e-01 4.15266484e-01 1.45973253e+00 -4.84462500e-01
2.66875904e-02 -2.98806250e-01 -7.21823871e-01 4.96818721e-01
9.09986138e-01 -3.51515442e-01 -4.67566311e-01 -2.94780314e-01
8.07618424e-02 -2.44825974e-01 4.33178276e-01 -1.13433838e+00
3.22068781e-01 -2.64849607e-02 1.14201641e+00 -6.90028429e-01
-1.60723284e-01 -9.94749665e-01 5.16149521e-01 9.39803302e-01
-1.33240059e-01 3.63968074e-01 1.67834759e-01 3.73649985e-01
-4.13091302e-01 -5.07838607e-01 3.87110263e-01 -3.58210623e-01
-6.73610747e-01 1.05072595e-01 -6.07400000e-01 -8.42524052e-01
8.73682499e-01 -5.32333255e-01 -4.21062797e-01 1.56276986e-01
-3.87670636e-01 1.94659606e-01 9.17614251e-02 3.32238019e-01
9.24676776e-01 -1.37046480e+00 -4.64115083e-01 4.60627049e-01
3.48349810e-02 5.43892682e-01 2.52022296e-01 7.67524958e-01
-8.55613589e-01 4.70561266e-01 -2.45208487e-01 -5.23139596e-01
-1.11595821e+00 6.03106022e-01 2.97448158e-01 -2.28372261e-01
-3.94145280e-01 6.08077049e-01 3.66291374e-01 -1.86498180e-01
8.23640823e-02 -7.49485254e-01 -4.32822615e-01 -1.40345424e-01
2.73362398e-01 3.28467041e-01 4.62456495e-01 -2.54216522e-01
-3.11307192e-01 2.39800841e-01 -9.35881734e-02 7.07192898e-01
1.26194656e+00 3.08131754e-01 -5.41074574e-01 -3.76779623e-02
1.05010366e+00 -6.89923525e-01 -1.05355728e+00 1.71616867e-01
6.30072132e-02 -1.80169865e-01 1.23078533e-01 -8.38962018e-01
-1.32564008e+00 8.15911949e-01 6.12323165e-01 4.76648122e-01
1.31480575e+00 -3.46114755e-01 6.05927169e-01 4.97174621e-01
-4.83795851e-02 -9.96774197e-01 2.83128917e-01 5.42540073e-01
7.83441901e-01 -9.96793270e-01 3.59560065e-02 -7.39570320e-01
-3.94756734e-01 1.63311708e+00 8.58828127e-01 7.44782090e-02
4.59505528e-01 2.54768580e-01 -1.35569707e-01 -7.47721434e-01
-4.50050086e-01 9.06316862e-02 2.07355633e-01 4.80959952e-01
4.58945960e-01 2.57929601e-02 -1.51076630e-01 7.84106672e-01
-1.40644889e-02 -1.86475903e-01 2.70650417e-01 1.02635348e+00
-5.27435601e-01 -7.31576860e-01 -7.21127570e-01 7.89868832e-01
-4.92824644e-01 -2.17423990e-01 -3.98818225e-01 6.28833652e-01
5.15289664e-01 1.11940336e+00 1.12684600e-01 -5.71869254e-01
3.58514935e-01 -2.56086979e-02 6.55041784e-02 -3.00564378e-01
-8.11651647e-01 1.89327318e-02 -2.28906557e-01 -1.93512037e-01
-8.17105323e-02 -3.84619474e-01 -1.79091299e+00 -4.43140715e-01
-5.57156086e-01 3.59378904e-01 4.76519972e-01 1.14786136e+00
3.59867752e-01 1.17303336e+00 7.15557516e-01 -4.54639018e-01
-3.97370577e-01 -1.00830948e+00 -6.17081821e-01 5.32458544e-01
1.76642284e-01 -8.18399906e-01 -2.89812118e-01 5.95309287e-02] | [7.244388103485107, 2.2036118507385254] |
5acc523e-314f-4202-aa92-03ffceb166ba | a-multimodal-dataset-for-deception-detection | null | null | https://aclanthology.org/L14-1673 | https://aclanthology.org/L14-1673.pdf | A Multimodal Dataset for Deception Detection | This paper presents the construction of a multimodal dataset for deception detection, including physiological, thermal, and visual responses of human subjects under three deceptive scenarios. We present the experimental protocol, as well as the data acquisition process. To evaluate the usefulness of the dataset for the task of deception detection, we present a statistical analysis of the physiological and thermal modalities associated with the deceptive and truthful conditions. Initial results show that physiological and thermal responses can differentiate between deceptive and truthful states. | ["Ver{\\'o}nica P{\\'e}rez-Rosas", 'Mihai Burzo', 'Alexis Narvaez', 'Rada Mihalcea'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['deception-detection'] | ['miscellaneous'] | [-5.19216731e-02 -4.57418144e-01 3.17204982e-01 -7.92507946e-01
-4.21847731e-01 -8.37349892e-01 7.77193189e-01 -3.25635560e-02
-3.69089335e-01 7.80353487e-01 2.20658854e-01 2.93544456e-02
2.03143939e-01 2.66638130e-01 -5.92270680e-03 -7.27503419e-01
-4.02194215e-03 -3.46981019e-01 -6.34679794e-01 1.65193319e-01
8.81028295e-01 6.97803676e-01 -1.39779305e+00 6.41808510e-01
7.13305175e-01 1.60794699e+00 -7.94602633e-01 1.01115358e+00
5.60516834e-01 4.51290607e-01 -1.08167231e+00 -4.08034623e-01
2.34758966e-02 -5.88232994e-01 -6.85561061e-01 3.99445817e-02
7.64254928e-01 -7.14433551e-01 -5.47432721e-01 7.14931726e-01
7.78509259e-01 1.12727530e-01 7.05705225e-01 -1.59946799e+00
-8.78914118e-01 -4.54678386e-01 7.17837736e-02 4.11286771e-01
1.12166464e+00 5.60574889e-01 2.18140289e-01 -9.44671273e-01
-5.46486769e-03 1.20067978e+00 2.22765833e-01 1.21547067e+00
-1.15905523e+00 -5.15268922e-01 -3.54502976e-01 6.75357461e-01
-1.25756741e+00 -1.33662248e+00 8.20928752e-01 -4.84705925e-01
5.04985869e-01 6.50389791e-01 7.79141784e-01 2.09507799e+00
5.77724993e-01 8.13043356e-01 1.78551006e+00 3.54179516e-02
4.76692647e-01 3.59736681e-01 4.15641636e-01 2.89226353e-01
1.75406277e-01 6.83956206e-01 -8.90544295e-01 -6.99555278e-01
2.68498868e-01 -6.15100503e-01 -5.84895909e-01 -1.76752463e-01
-7.88200557e-01 4.74571764e-01 -4.39352058e-02 -2.75476612e-02
-2.99025118e-01 3.15526910e-02 6.02962375e-01 3.66389900e-01
3.25640202e-01 5.48462272e-01 9.25457254e-02 -4.59336787e-01
-7.77889073e-01 -1.60670519e-01 8.20347428e-01 2.74183333e-01
-7.06340298e-02 3.68994921e-01 -3.33894342e-01 5.73514342e-01
2.09179342e-01 8.71050835e-01 5.67763627e-01 -8.84051859e-01
-1.21698871e-01 2.42065862e-01 6.29320681e-01 -1.36292744e+00
-3.47029686e-01 3.25065494e-01 -4.26226407e-01 2.12083042e-01
3.66771817e-01 -2.45146349e-01 -8.87903512e-01 1.12627637e+00
-1.06610149e-01 -2.98911154e-01 1.38982475e-01 1.40619898e+00
1.03286982e+00 4.10592020e-01 3.19743365e-01 -6.90442085e-01
1.22973680e+00 -4.77484651e-02 -1.12918293e+00 -3.83360982e-01
-1.06054515e-01 -4.98588264e-01 8.15302849e-01 7.66666293e-01
-1.18528426e+00 -3.91634941e-01 -7.46402562e-01 1.55495733e-01
-2.99544394e-01 -1.35870427e-01 4.68514472e-01 1.25867188e+00
-7.75633276e-01 9.93682072e-02 -2.80259043e-01 -3.36592883e-01
3.42378289e-01 -2.92983055e-01 -5.69649518e-01 -5.67420870e-02
-1.19796693e+00 1.43553007e+00 2.19231635e-01 7.90096939e-01
-1.29253697e+00 6.08630665e-02 -9.02236879e-01 -3.00727606e-01
-1.53508902e-01 -5.42822361e-01 6.93733692e-01 -1.16822803e+00
-1.41870809e+00 1.22664452e+00 -2.98544079e-01 2.54932828e-02
2.76128441e-01 -2.24214554e-01 -8.92447829e-01 6.46783829e-01
-7.44940341e-01 3.27552110e-01 1.42553067e+00 -1.64978218e+00
3.30039561e-01 -6.10795736e-01 -4.82124716e-01 3.42907816e-01
1.76012255e-02 3.54410827e-01 6.45804405e-01 3.28443460e-02
-5.19962125e-02 -5.01185894e-01 5.17198861e-01 -1.51766032e-01
-3.49374831e-01 1.46195397e-01 7.93714166e-01 -8.59176457e-01
8.63976359e-01 -2.29400992e+00 -2.15850309e-01 3.06572944e-01
2.41078928e-01 2.95929283e-01 -3.90479937e-02 4.18843299e-01
-1.80081621e-01 3.08397681e-01 1.19053192e-01 -2.92216033e-01
3.04272860e-01 1.89048603e-01 -3.40699464e-01 1.03804064e+00
-1.75741807e-01 1.17336345e+00 -5.51576972e-01 -5.00105441e-01
3.61590624e-01 2.17152700e-01 4.21446592e-01 5.90084016e-01
7.15654135e-01 3.51083875e-01 -8.38062614e-02 1.13404119e+00
7.29813397e-01 8.91539454e-01 -2.58608401e-01 -3.18109602e-01
6.82379529e-02 -4.97416370e-02 -1.48228392e-01 9.17233586e-01
1.13912769e-01 1.04517448e+00 5.69988728e-01 -5.66711664e-01
9.98997211e-01 7.62692273e-01 -6.17987998e-02 -9.10075128e-01
6.17715418e-01 5.52849434e-02 -1.15108736e-01 -1.14200640e+00
5.76882303e-01 -6.02396309e-01 -5.58748007e-01 2.76453823e-01
-1.10156938e-01 -4.03245330e-01 -5.56435764e-01 1.93764493e-01
6.04098856e-01 -2.67512977e-01 3.77111882e-01 -7.46569932e-02
3.77721965e-01 -2.61187375e-01 2.90486813e-01 7.87161350e-01
-1.32944334e+00 4.59373593e-01 6.74143851e-01 -6.48923218e-01
-4.58689779e-01 -1.32524157e+00 -1.12526052e-01 6.66927695e-01
4.55325842e-01 3.49980682e-01 -6.54476881e-01 -5.54405808e-01
-3.75505141e-03 1.16787231e+00 -8.80733669e-01 -8.70995879e-01
1.72801599e-01 -7.45454371e-01 8.58483732e-01 2.92671695e-02
4.20856357e-01 -1.11706603e+00 -9.36432660e-01 -7.25028753e-01
-5.58569968e-01 -1.01326632e+00 -2.92904675e-01 -1.68536916e-01
-6.41328990e-01 -9.76375639e-01 5.19012026e-02 -8.00633151e-03
4.99546349e-01 1.04484126e-01 6.72638178e-01 -4.11122385e-03
-5.04588306e-01 1.02195287e+00 -3.23440015e-01 -2.38577262e-01
-3.33201617e-01 -1.11822772e+00 3.45066518e-01 3.84091973e-01
2.06407219e-01 -1.00419067e-01 -8.72548163e-01 4.37014878e-01
-8.38302135e-01 -5.82051694e-01 2.92961180e-01 4.14420545e-01
-2.73111939e-01 -5.74353337e-01 5.11633396e-01 2.06238568e-01
1.50053585e+00 -3.27438265e-01 5.72162233e-02 2.27529094e-01
-3.48935038e-01 -4.14923102e-01 1.11501418e-01 -4.13456798e-01
-1.23658538e+00 -1.53078273e-01 3.27899665e-01 -2.90758818e-01
-9.52428520e-01 2.37715378e-01 -1.93941846e-01 -5.77752233e-01
1.01493800e+00 4.45053965e-01 8.06666538e-02 -1.34308308e-01
1.89063981e-01 1.00668967e+00 8.19292903e-01 -4.16570157e-01
2.65793115e-01 3.66301477e-01 -1.23118222e-01 -8.42572272e-01
-5.91329873e-01 -3.27049404e-01 -6.64458692e-01 -1.03543496e+00
4.14770305e-01 -1.91389024e-01 -1.32528043e+00 8.56454313e-01
-1.13690877e+00 1.81306630e-01 1.58700779e-01 4.07684356e-01
-4.45236355e-01 9.90534008e-01 -5.81258476e-01 -1.54610968e+00
-4.69327569e-01 -6.04227960e-01 8.10262859e-01 9.68336239e-02
-4.08670634e-01 -1.04981029e+00 1.24191098e-01 9.92361605e-01
4.88532335e-01 7.15379000e-01 5.55287600e-01 -7.06971049e-01
3.22833359e-01 -6.94406927e-01 -9.82753485e-02 5.16063929e-01
-1.21322408e-01 -6.16726205e-02 -1.38925803e+00 -2.41317466e-01
6.32203341e-01 -1.07783794e+00 5.95572352e-01 2.52717108e-01
8.44856501e-01 -4.57909226e-01 7.78133646e-02 2.20099568e-01
8.84217143e-01 2.40921319e-01 1.02763104e+00 -3.60197246e-01
3.00681949e-01 9.32035029e-01 4.63241488e-01 3.67790073e-01
-9.87150893e-02 4.14922953e-01 7.79658914e-01 1.86220571e-01
5.66430688e-01 1.54286325e-01 8.65443707e-01 1.99268073e-01
1.19633436e-01 -4.39007372e-01 -6.88112020e-01 3.09106350e-01
-1.15648532e+00 -1.27086687e+00 -3.74512762e-01 2.25043964e+00
5.10904610e-01 -5.35995960e-01 1.41231820e-01 7.17031583e-02
4.22022521e-01 1.86726585e-01 -4.88199145e-01 -1.40336573e+00
-2.06155628e-01 -1.76479265e-01 -7.56608695e-02 4.07396227e-01
-8.12829375e-01 4.06395257e-01 8.84717560e+00 3.72898698e-01
-1.03711402e+00 -2.86802828e-01 5.28102994e-01 -4.50967342e-01
1.69461593e-02 -4.60662305e-01 2.11395741e-01 5.06920815e-01
1.43834722e+00 1.02669545e-01 4.44723606e-01 2.82950819e-01
8.00008714e-01 -9.53715622e-01 -1.33096826e+00 1.16539419e+00
7.94775784e-01 -2.10711658e-01 -1.36751130e-01 -1.85999602e-01
-1.16614386e-01 -6.10551298e-01 2.38647223e-01 -9.44592729e-02
-5.21801054e-01 -1.45744503e+00 7.10837841e-01 1.06562161e+00
6.81757569e-01 -5.09580314e-01 9.43094671e-01 3.40084344e-01
2.00517133e-01 4.92585450e-02 -1.53607741e-01 -2.33983874e-01
1.28447980e-01 2.94457465e-01 -7.05320776e-01 3.65646452e-01
2.90715069e-01 1.46704376e-01 -6.17662311e-01 8.26388538e-01
-1.88200235e-01 5.57588875e-01 2.31172085e-01 -3.79131772e-02
-3.01783800e-01 6.25693202e-02 7.33356416e-01 1.21869779e+00
-1.53277060e-02 2.96267360e-01 -3.06245565e-01 1.14845264e+00
4.93902594e-01 -2.92510688e-01 -5.41123331e-01 -1.82344347e-01
2.86214590e-01 1.37014937e+00 2.36727241e-02 -2.41012722e-01
3.11027765e-01 1.51477051e+00 -2.98985422e-01 5.01853347e-01
-7.53003180e-01 -3.20120811e-01 5.45733690e-01 -4.00473416e-01
-6.95094168e-01 -3.03139329e-01 -6.49986744e-01 -1.23441553e+00
-8.97058696e-02 -7.24394262e-01 3.11454266e-01 -1.49565351e+00
-1.48220587e+00 3.45145732e-01 4.24898177e-01 -6.74103379e-01
-1.77855678e-02 -5.88006854e-01 -4.78879869e-01 1.11434329e+00
-1.02152288e+00 -7.28588462e-01 -8.79105568e-01 6.47064507e-01
-3.00348364e-02 4.75556254e-02 9.28220391e-01 -3.06810975e-01
-6.45867527e-01 4.66757625e-01 -4.86219794e-01 -7.29271695e-02
9.48701262e-01 -9.43156719e-01 -2.35056937e-01 5.90868175e-01
-6.61304593e-01 6.40548170e-01 9.63504672e-01 -3.27654958e-01
-1.35278046e+00 -8.20510983e-02 6.57702029e-01 -6.66529596e-01
4.59233344e-01 -4.09507424e-01 -8.97932351e-01 1.20884679e-01
5.19504189e-01 -2.24465858e-02 9.08637464e-01 -1.43839896e-01
-4.02009875e-01 6.49730191e-02 -1.75240695e+00 3.35657597e-01
4.62239683e-01 -8.81108284e-01 -9.85232353e-01 -1.82216838e-01
-4.32406366e-01 -4.49225307e-02 -6.89901054e-01 5.80044836e-02
1.10144699e+00 -1.16213310e+00 7.58404851e-01 -7.45078385e-01
1.79278612e-01 1.14340544e-01 9.52058807e-02 -1.70090473e+00
-9.72858295e-02 -5.76985657e-01 -9.22512934e-02 6.59293950e-01
-6.77738711e-02 -7.64893115e-01 2.66460657e-01 1.16101956e+00
5.33170402e-02 -1.19532704e-01 -1.32918072e+00 -5.82118869e-01
-6.89184666e-02 -2.70238757e-01 -2.97146857e-01 1.00434577e+00
8.98868918e-01 9.39920843e-02 -8.45901132e-01 1.70924902e-01
5.74233413e-01 -1.27122803e-02 2.44368330e-01 -9.70788479e-01
5.34846067e-01 -1.12166591e-01 -6.95793509e-01 -4.62873161e-01
1.70440093e-01 -1.70658872e-01 3.78557295e-01 -1.06096327e+00
5.69608152e-01 5.75997055e-01 -2.58024693e-01 4.94699389e-01
-1.51910305e-01 3.76483440e-01 -3.91198024e-02 3.30402970e-01
-5.41240335e-01 6.34086549e-01 9.82128203e-01 1.56672060e-01
4.99517433e-02 -4.55474943e-01 -6.43598020e-01 3.10896248e-01
1.04356849e+00 -9.01306942e-02 -1.26939341e-01 5.71870133e-02
-4.31000367e-02 5.19052863e-01 1.23056769e+00 -6.11595690e-01
-9.68223363e-02 -4.03376997e-01 8.17589164e-01 -3.16418499e-01
9.42121923e-01 -7.55266726e-01 -1.84880853e-01 3.38788807e-01
-4.30307508e-01 -9.68375280e-02 4.04652894e-01 4.34262037e-01
-7.69741759e-02 -2.34183654e-01 1.12738454e+00 -1.17356189e-01
-3.51851374e-01 -5.22940814e-01 -1.01510668e+00 -2.84871101e-01
9.90658224e-01 -7.47180879e-01 -8.91571343e-01 -8.68395686e-01
-8.80221546e-01 1.07487038e-01 4.29035902e-01 3.21922392e-01
1.35607553e+00 -1.06532502e+00 -5.24786711e-01 1.03428371e-01
2.04908878e-01 -1.34150350e+00 4.10637051e-01 1.43387687e+00
-8.92788693e-02 4.21395868e-01 -8.80415082e-01 -2.29210168e-01
-1.71647418e+00 5.41781962e-01 8.35484266e-01 7.54152834e-01
2.77722955e-01 4.17915195e-01 2.55850088e-02 2.99776316e-01
-9.43580642e-02 2.33280748e-01 2.39965115e-02 1.00481631e-02
4.20169175e-01 8.24546874e-01 -6.73884600e-02 -7.98057318e-01
-6.72198176e-01 -2.12518275e-01 2.90032387e-01 -4.04348135e-01
4.23310995e-01 -5.65433681e-01 -2.68061876e-01 7.92425096e-01
1.04080594e+00 -7.64952898e-02 -6.99299097e-01 5.62876463e-01
-3.28376174e-01 -9.57989216e-01 1.41980365e-01 -1.60321164e+00
-5.17797709e-01 9.35237288e-01 6.79218948e-01 3.68005216e-01
1.13985753e+00 -1.46787137e-01 4.85642076e-01 -8.74553341e-03
7.80041218e-02 -1.28762043e+00 2.80487716e-01 7.62415454e-02
1.34209085e+00 -1.12887979e+00 2.21970618e-01 -1.70560852e-01
-9.64931846e-01 9.92163837e-01 4.69599426e-01 1.27042755e-01
-1.14387259e-01 -2.45707825e-01 2.90320754e-01 -5.01160264e-01
-8.39442968e-01 2.95569181e-01 7.02979803e-01 9.73038554e-01
1.58480406e-01 1.94672540e-01 -5.06821930e-01 4.22004819e-01
1.01777561e-01 -2.89784878e-01 7.19201207e-01 8.15263033e-01
-4.57528055e-01 -3.23788583e-01 -8.85185242e-01 1.34655505e-01
-3.30175638e-01 1.38825119e-01 -1.87936389e+00 5.64674914e-01
-2.37988323e-01 1.54310381e+00 -1.41155601e-01 -5.92894316e-01
4.92916405e-01 6.00573421e-01 7.31210232e-01 -4.61446829e-02
-7.21536636e-01 -2.33533055e-01 5.52894831e-01 -9.79636192e-01
-3.80175292e-01 -8.23961616e-01 -4.70442712e-01 -6.76973045e-01
-2.09523544e-01 -9.07521844e-02 7.58343935e-01 8.15431297e-01
2.63070196e-01 -3.01756352e-01 7.52033889e-01 -7.43444622e-01
-5.04213572e-01 -1.14432406e+00 -7.12406278e-01 7.85720766e-01
6.76014185e-01 -4.46860820e-01 -1.13473392e+00 -5.81489690e-02] | [13.31544017791748, 2.079537868499756] |
b6c1a0ce-2e27-4284-a50a-6e5a6b6c92c3 | depth-infused-binaural-audio-generation-using | 2108.04906 | null | https://arxiv.org/abs/2108.04906v1 | https://arxiv.org/pdf/2108.04906v1.pdf | Depth Infused Binaural Audio Generation using Hierarchical Cross-Modal Attention | Binaural audio gives the listener the feeling of being in the recording place and enhances the immersive experience if coupled with AR/VR. But the problem with binaural audio recording is that it requires a specialized setup which is not possible to fabricate within handheld devices as compared to traditional mono audio that can be recorded with a single microphone. In order to overcome this drawback, prior works have tried to uplift the mono recorded audio to binaural audio as a post processing step conditioning on the visual input. But all the prior approaches missed other most important information required for the task, i.e. distance of different sound producing objects from the recording setup. In this work, we argue that the depth map of the scene can act as a proxy for encoding distance information of objects in the scene and show that adding depth features along with image features improves the performance both qualitatively and quantitatively. We propose a novel encoder-decoder architecture, where we use a hierarchical attention mechanism to encode the image and depth feature extracted from individual transformer backbone, with audio features at each layer of the decoder. | ['Gaurav Sharma', 'Neeraj Matiyali', 'Siddharth Srivastava', 'Kranti Kumar Parida'] | 2021-08-10 | null | null | null | null | ['audio-generation'] | ['audio'] | [ 4.06435370e-01 -1.75535023e-01 6.31931126e-01 -2.59217620e-01
-7.79231608e-01 -5.65243125e-01 3.71121615e-01 1.81351498e-01
-3.69741142e-01 2.55934834e-01 4.60035443e-01 1.27238765e-01
-3.64821292e-02 -6.95758402e-01 -9.42329347e-01 -6.43422425e-01
1.72359362e-01 7.51783475e-02 4.76425856e-01 -7.24228323e-02
3.69694352e-01 4.00023460e-01 -2.26187372e+00 5.81177711e-01
2.73292959e-01 1.23122442e+00 7.33478487e-01 1.11262870e+00
1.43015921e-01 7.10242271e-01 -8.13223302e-01 2.11859606e-02
2.05778137e-01 -3.73204947e-01 -3.84441227e-01 -1.49284035e-01
7.13353634e-01 -3.71693373e-01 -3.57046336e-01 9.25612271e-01
8.84338915e-01 1.62671372e-01 4.49173987e-01 -9.09678042e-01
-3.75057369e-01 4.79262829e-01 -3.74695867e-01 2.64152795e-01
8.29376221e-01 -1.25035942e-01 9.45746243e-01 -8.94580483e-01
4.91961509e-01 8.89403701e-01 4.29546505e-01 1.95563436e-01
-8.61505270e-01 -4.42228913e-01 -1.31801710e-01 4.14278001e-01
-1.42500460e+00 -5.98998010e-01 9.44157660e-01 -5.30298471e-01
9.52905715e-01 4.48530942e-01 9.03610408e-01 7.65558243e-01
9.04459804e-02 4.50172186e-01 9.67408061e-01 -4.16224867e-01
2.24860102e-01 4.48516577e-01 -2.62659609e-01 1.13576934e-01
-3.76237750e-01 7.32749477e-02 -8.05932641e-01 3.27445716e-01
8.69649231e-01 -2.90935393e-03 -6.05268896e-01 -3.16391498e-01
-9.06461954e-01 5.01708031e-01 5.38766086e-01 4.49550599e-01
-3.47613126e-01 1.31873444e-01 2.81400263e-01 3.61079544e-01
2.80245125e-01 5.64254701e-01 -1.11658804e-01 -4.29298788e-01
-1.02701151e+00 -2.10255131e-01 4.03902382e-01 5.88254988e-01
4.86050695e-01 -7.13250116e-02 1.82409257e-01 7.30290592e-01
3.84798497e-01 1.75468415e-01 3.11490238e-01 -8.98359776e-01
3.62139195e-01 2.33261257e-01 -1.34248212e-01 -9.93496835e-01
-1.50766179e-01 -3.65344524e-01 -3.50079983e-01 5.84364474e-01
3.27392638e-01 6.91285729e-02 -7.05120087e-01 1.45873308e+00
2.81125695e-01 4.50952709e-01 -1.24377623e-01 1.30874670e+00
9.25400555e-01 7.89321125e-01 -5.45244575e-01 -1.09623492e-01
1.43531668e+00 -8.15282047e-01 -8.09007525e-01 4.52561816e-03
2.22333729e-01 -9.60821211e-01 1.37057090e+00 7.23555565e-01
-1.15119636e+00 -7.60030627e-01 -1.22943115e+00 -3.32139432e-01
-3.90361816e-01 -1.46684706e-01 4.01135504e-01 6.07758760e-01
-1.28843856e+00 3.23856682e-01 -6.39585972e-01 -2.20178813e-01
-6.78330511e-02 3.02472889e-01 -5.24932861e-01 6.14727288e-02
-1.04667735e+00 8.27777267e-01 -1.23119295e-01 2.59899706e-01
-1.08682799e+00 -7.04400778e-01 -8.72357130e-01 1.83068022e-01
1.52335137e-01 -4.42870051e-01 1.15597379e+00 -8.77005577e-01
-1.78843331e+00 6.13370478e-01 -1.39747217e-01 -1.82043642e-01
3.44864488e-01 -5.07474482e-01 -2.11878642e-01 4.41337287e-01
-3.00182194e-01 5.89249969e-01 8.43046606e-01 -1.36430168e+00
-5.79533637e-01 -4.14414704e-01 4.40172374e-01 5.22189617e-01
-1.38253108e-01 1.71372145e-01 -4.08299476e-01 -2.53217340e-01
1.13782197e-01 -5.39577186e-01 4.46899757e-02 -6.45008981e-02
-5.39616346e-02 3.49025816e-01 8.90513539e-01 -7.30303586e-01
1.02453673e+00 -2.46645308e+00 1.25981107e-01 9.59007666e-02
2.02670574e-01 9.67881754e-02 1.27812624e-01 4.72073942e-01
-1.40265718e-01 -3.55900079e-01 2.02109307e-01 -3.62846881e-01
-1.06062219e-01 -9.00265425e-02 -2.50878125e-01 4.12067235e-01
-2.21725017e-01 2.88642853e-01 -7.33689547e-01 -3.64851624e-01
4.86140609e-01 1.07756364e+00 -7.98327148e-01 3.14084768e-01
2.97571272e-01 6.40694380e-01 9.18082967e-02 3.41795206e-01
7.30608523e-01 2.98404545e-01 -3.27149600e-01 -2.42946036e-02
-2.94775188e-01 6.73990011e-01 -1.42556608e+00 1.87720585e+00
-7.24026442e-01 9.64775622e-01 3.71019840e-01 -7.90834785e-01
8.92598271e-01 6.51015699e-01 1.98950663e-01 -8.84518027e-01
2.20028013e-01 3.06061525e-02 -5.84313497e-02 -5.80977976e-01
7.57794976e-01 -2.66483784e-01 3.89412418e-02 3.02830279e-01
1.03615724e-01 -3.19037795e-01 -3.22202444e-01 5.77788241e-02
8.78282249e-01 1.90751571e-02 -2.81823110e-02 2.36016557e-01
2.75303960e-01 -5.33014715e-01 1.61911964e-01 6.13391817e-01
1.06275603e-01 1.05917406e+00 1.92976445e-01 4.71408293e-02
-9.14136291e-01 -1.13568389e+00 -1.93969682e-01 1.16027844e+00
2.44210944e-01 -5.96637130e-01 -7.30112314e-01 -1.39028370e-01
-5.16026914e-01 4.51452553e-01 -4.52681541e-01 -1.45338401e-01
-2.09857896e-01 -1.25099435e-01 3.94607812e-01 5.89395106e-01
1.45494580e-01 -9.12720025e-01 -1.12405968e+00 1.61008328e-01
-2.37126455e-01 -9.44857895e-01 -1.92584768e-01 3.80972087e-01
-5.83823681e-01 -4.96322811e-01 -5.35081446e-01 -7.01676846e-01
1.86393768e-01 2.72478610e-01 8.80729496e-01 -3.30497980e-01
-5.05477309e-01 5.11926413e-01 -4.54466522e-01 -4.85835433e-01
1.43977702e-01 -3.67385566e-01 -3.22882831e-02 9.35986787e-02
4.59867045e-02 -1.07495117e+00 -8.07788908e-01 1.66285068e-01
-7.44403005e-01 4.22635376e-02 3.39459240e-01 3.53045374e-01
6.15521491e-01 1.62435696e-01 2.54306853e-01 -3.08126390e-01
2.29488075e-01 -1.73765421e-01 -3.86632651e-01 -3.35354835e-01
2.84277380e-01 -3.44360292e-01 4.93615150e-01 -4.39315051e-01
-8.93274665e-01 3.15184593e-01 -4.78640556e-01 -3.51145118e-01
-3.20948273e-01 3.33163649e-01 -5.53564608e-01 2.74890717e-02
4.05550271e-01 3.09591964e-02 -3.35787773e-01 -6.47306144e-01
2.07388714e-01 9.39308822e-01 5.50956309e-01 -1.06787883e-01
3.10334831e-01 5.87396562e-01 -1.27382860e-01 -1.00178719e+00
-5.06844819e-01 -5.48721194e-01 -6.18901432e-01 -5.34229934e-01
9.44295347e-01 -9.52700257e-01 -7.45828509e-01 2.02616841e-01
-1.15012705e+00 -2.05357492e-01 -3.60598952e-01 8.91273975e-01
-5.51762044e-01 2.19904818e-02 -5.08507550e-01 -9.81607556e-01
2.32065111e-01 -1.23327935e+00 1.10025370e+00 1.41334400e-01
-1.80958197e-01 -6.54058874e-01 2.18871415e-01 2.75793165e-01
3.36821675e-01 1.11952990e-01 4.98748839e-01 -1.03973314e-01
-5.91276884e-01 -2.72207499e-01 -7.20558967e-03 3.38309109e-01
-4.29401733e-02 -9.90071371e-02 -1.67852032e+00 -4.88376804e-02
3.64777654e-01 -3.26771766e-01 7.05554664e-01 5.83095014e-01
1.17403007e+00 -3.89477089e-02 1.06445961e-01 6.27971530e-01
1.51839840e+00 4.08620089e-01 1.08546460e+00 1.16322622e-01
7.19138026e-01 7.26301670e-01 6.11499906e-01 4.07137007e-01
2.46042863e-01 9.11003709e-01 7.90588677e-01 -2.58924305e-01
-2.52274901e-01 -2.51189977e-01 4.61121649e-01 1.00302315e+00
-2.35247180e-01 -3.15441549e-01 -6.25106931e-01 5.14788747e-01
-1.32036328e+00 -9.72320259e-01 -1.64633483e-01 2.37032533e+00
6.20435059e-01 3.25373523e-02 -5.99211268e-02 7.07127750e-01
5.44049740e-01 -7.89101645e-02 3.08368783e-02 -8.14857244e-01
-2.09786519e-02 4.31540191e-01 9.93828997e-02 7.84615517e-01
-7.64522612e-01 5.00351846e-01 6.62625265e+00 5.13267398e-01
-1.58642876e+00 1.51841879e-01 -8.49652011e-03 -6.05487823e-01
-2.05114171e-01 -1.21783920e-01 -4.64486927e-01 3.10728937e-01
1.05230594e+00 3.37879717e-01 3.46411318e-01 7.15486407e-01
2.19201624e-01 -3.45973670e-01 -1.43462145e+00 1.29636860e+00
2.79573441e-01 -8.28561068e-01 -1.41528040e-01 2.19226062e-01
2.61596441e-01 -3.40797938e-02 4.11027312e-01 -8.92339945e-02
-2.81523287e-01 -1.14042199e+00 9.94136453e-01 5.50884306e-01
7.73907363e-01 -7.98573911e-01 6.27912343e-01 1.88431323e-01
-1.26450288e+00 3.38594578e-02 -4.53308165e-01 -5.45932353e-01
3.30814064e-01 4.25635576e-01 -9.69177783e-01 3.63434017e-01
1.03615773e+00 2.72017092e-01 -4.88458723e-01 1.42426777e+00
-5.85858673e-02 5.09606779e-01 -4.62729484e-01 2.80443102e-01
4.25479077e-02 1.01957247e-01 6.54791534e-01 1.24673724e+00
6.42332196e-01 -6.99135587e-02 -2.58769542e-01 4.71550375e-01
2.90043831e-01 6.19908720e-02 -9.22851503e-01 1.77050740e-01
2.72612453e-01 1.05025196e+00 -7.36680388e-01 -1.42766824e-02
-3.54803771e-01 1.02285969e+00 -1.26683608e-01 2.10110754e-01
-9.45706546e-01 -8.85082543e-01 4.81853783e-01 4.42086190e-01
5.93488395e-01 -2.44466066e-01 -2.05661088e-01 -8.46441984e-01
2.64044017e-01 -6.31798387e-01 5.36211655e-02 -1.30651844e+00
-6.16266668e-01 7.54710734e-01 -4.49394137e-01 -1.43226933e+00
-1.43930852e-01 -4.94504958e-01 -3.85589182e-01 8.33265364e-01
-1.32925701e+00 -7.64601886e-01 -4.30757970e-01 8.04986954e-01
3.40172201e-01 3.62825185e-01 7.90448368e-01 7.27887928e-01
1.01690672e-01 5.58467925e-01 3.27133923e-03 -1.19129352e-01
8.98671269e-01 -1.22391593e+00 -2.81518459e-01 6.60232127e-01
4.85982239e-01 4.76649076e-01 9.70927775e-01 1.76354498e-02
-1.38198876e+00 -4.33646441e-01 8.33379149e-01 -5.94964564e-01
4.05317873e-01 -8.55357289e-01 -8.72910738e-01 4.54514921e-01
3.19855422e-01 -9.35795307e-02 8.83899271e-01 4.87097912e-02
-3.21893454e-01 -3.17038596e-01 -8.63900065e-01 2.63163894e-01
8.36910844e-01 -1.10243535e+00 -7.20319390e-01 -2.08051071e-01
8.27247322e-01 -5.01303673e-01 -7.03305840e-01 3.20190489e-02
7.66176462e-01 -1.18115461e+00 8.94319475e-01 9.20718461e-02
6.09261096e-01 -6.48783684e-01 -5.72658479e-01 -1.20785785e+00
-3.49913985e-02 -3.86428893e-01 1.47301093e-01 1.38218796e+00
3.58049661e-01 -3.34165245e-01 4.10630435e-01 3.13541740e-02
-3.40509415e-01 -4.12297249e-01 -1.05150592e+00 -4.13252890e-01
-3.78862172e-01 -9.12344277e-01 4.30134565e-01 6.58996224e-01
2.08712369e-01 5.32878280e-01 -3.39536697e-01 5.02504230e-01
1.08065434e-01 -4.83886115e-02 6.67346656e-01 -1.28347695e+00
-5.74474394e-01 -1.71934485e-01 -9.49334741e-01 -1.19820607e+00
-3.52690488e-01 -6.25223815e-01 2.45175049e-01 -1.70788479e+00
-7.24043697e-02 -1.72499210e-01 -3.65900457e-01 -2.60814968e-02
2.93930501e-01 6.99419379e-01 2.13001609e-01 -1.85208350e-01
-5.83225489e-01 4.46243346e-01 1.32965422e+00 2.24656463e-01
-2.69428402e-01 -2.08675101e-01 -6.52838469e-01 8.02246630e-01
3.67939740e-01 -4.08583224e-01 -4.92306262e-01 -6.72865331e-01
5.62239885e-01 4.07842100e-01 6.47503972e-01 -1.33479428e+00
3.79272610e-01 3.87711823e-01 4.51305002e-01 -7.18078196e-01
9.64002490e-01 -1.11487377e+00 2.32941717e-01 -3.62858810e-02
-3.60359579e-01 -9.77808237e-02 1.27248317e-01 5.07163286e-01
-6.93762481e-01 -1.40146986e-01 5.35171211e-01 1.11967847e-01
-4.28241313e-01 -2.05664173e-01 -5.45704603e-01 -2.93170214e-01
7.20927119e-01 -6.30536795e-01 2.52508707e-02 -7.05352962e-01
-9.42858577e-01 -2.70934820e-01 5.06584525e-01 2.19245434e-01
7.25054920e-01 -1.18992245e+00 -4.22269315e-01 3.50795895e-01
-1.43702433e-01 1.39752431e-02 5.66080153e-01 9.75950658e-01
-5.29442549e-01 3.19219172e-01 -4.72813874e-01 -8.98266435e-01
-1.57658339e+00 5.67255259e-01 2.52580047e-01 2.43604317e-01
-8.08519781e-01 1.04503679e+00 5.48893511e-01 1.95859857e-02
5.53799391e-01 -4.04057503e-01 -4.79683012e-01 3.03109765e-01
7.54889190e-01 1.18254058e-01 2.90244401e-01 -5.75849831e-01
-2.50493348e-01 8.31721246e-01 9.55505520e-02 -6.74043655e-01
1.45715880e+00 -4.57525373e-01 -2.62862165e-02 9.73625720e-01
1.34544361e+00 4.44359660e-01 -1.33262908e+00 -2.56239865e-02
-4.84321147e-01 -8.11989367e-01 3.47637266e-01 -6.52874351e-01
-9.75038826e-01 1.51010680e+00 8.15813303e-01 2.65808046e-01
1.46211898e+00 1.37527660e-01 5.66082060e-01 4.48208600e-02
3.35152477e-01 -1.08496737e+00 2.50550866e-01 3.92656177e-01
9.23138142e-01 -7.18048513e-01 -4.97278273e-01 -3.69093180e-01
-6.98084772e-01 8.92963111e-01 2.67235845e-01 -1.11819148e-01
6.07187212e-01 6.66697383e-01 1.57785878e-01 -2.33402729e-01
-6.44009233e-01 -2.85149276e-01 2.62552142e-01 7.97018528e-01
6.17872298e-01 -1.58028618e-01 2.36553103e-01 5.98099649e-01
-7.05454409e-01 -2.23391846e-01 6.19062126e-01 8.33109856e-01
-5.54453015e-01 -6.15472555e-01 -4.24980164e-01 3.86988707e-02
-7.09872842e-01 -7.58732930e-02 -4.05544788e-01 3.33339125e-01
3.96269381e-01 9.89691257e-01 3.69341582e-01 -5.84841788e-01
4.37199593e-01 1.62377469e-02 7.34253466e-01 -6.12711906e-01
-7.83355832e-01 4.53250170e-01 -1.44510180e-01 -6.72479689e-01
-3.32225263e-01 -5.08882105e-01 -1.12454259e+00 4.51786630e-02
-4.53837961e-01 2.35456929e-01 9.69269216e-01 6.21764541e-01
1.80888802e-01 9.93539274e-01 6.14081621e-01 -1.18711483e+00
2.31358245e-01 -7.63217032e-01 -8.64823699e-01 1.75391480e-01
7.46401548e-01 -6.32693946e-01 -5.41679263e-01 1.41487971e-01] | [14.992269515991211, 5.061923027038574] |
5e180310-8930-41fb-9dc4-482169e685bd | duta-vc-a-duration-aware-typical-to-atypical | 2306.10588 | null | https://arxiv.org/abs/2306.10588v1 | https://arxiv.org/pdf/2306.10588v1.pdf | DuTa-VC: A Duration-aware Typical-to-atypical Voice Conversion Approach with Diffusion Probabilistic Model | We present a novel typical-to-atypical voice conversion approach (DuTa-VC), which (i) can be trained with nonparallel data (ii) first introduces diffusion probabilistic model (iii) preserves the target speaker identity (iv) is aware of the phoneme duration of the target speaker. DuTa-VC consists of three parts: an encoder transforms the source mel-spectrogram into a duration-modified speaker-independent mel-spectrogram, a decoder performs the reverse diffusion to generate the target mel-spectrogram, and a vocoder is applied to reconstruct the waveform. Objective evaluations conducted on the UASpeech show that DuTa-VC is able to capture severity characteristics of dysarthric speech, reserves speaker identity, and significantly improves dysarthric speech recognition as a data augmentation. Subjective evaluations by two expert speech pathologists validate that DuTa-VC can preserve the severity and type of dysarthria of the target speakers in the synthesized speech. | ['Laureano Moro-Velazquez', 'Najim Dehak', 'Becky Lammers', 'Myra Sydnor', 'Jesus Villalba', 'Thomas Thebaud', 'Helin Wang'] | 2023-06-18 | null | null | null | null | ['voice-conversion', 'voice-conversion'] | ['audio', 'speech'] | [ 1.95593163e-01 2.09726825e-01 1.70588002e-01 -7.63857961e-02
-1.17806196e+00 -5.07015169e-01 1.35166287e-01 -4.48093206e-01
-1.19647786e-01 5.62688887e-01 8.91159296e-01 -4.23365295e-01
2.30024233e-02 -3.07226866e-01 -3.55366588e-01 -6.91340268e-01
1.14916921e-01 4.49749619e-01 2.51473606e-01 -1.63419873e-01
-7.76416510e-02 4.97419298e-01 -1.47967756e+00 4.24836367e-01
8.98347318e-01 6.25249267e-01 6.23005748e-01 1.09841621e+00
1.09504566e-01 7.26947427e-01 -9.56492186e-01 -1.36106133e-01
9.26740617e-02 -8.82096350e-01 -7.06049502e-01 2.49717832e-01
2.25823373e-01 -4.63716537e-01 -7.37902582e-01 1.10782230e+00
9.43429530e-01 -7.47160539e-02 5.66810846e-01 -8.37182462e-01
-9.04141009e-01 8.82907152e-01 6.09296523e-02 4.25268501e-01
3.03861976e-01 1.75172687e-01 5.97594202e-01 -7.67775476e-01
5.51780581e-01 1.26459432e+00 6.01996124e-01 8.58444571e-01
-1.23979414e+00 -6.53447092e-01 -3.94970030e-01 2.96488523e-01
-1.20808160e+00 -7.35798419e-01 9.73818839e-01 -4.61955816e-01
1.00779557e+00 4.00921494e-01 8.83572638e-01 1.24769652e+00
4.32897136e-02 3.16800445e-01 1.09802055e+00 -5.93837678e-01
1.50517568e-01 2.17294265e-02 6.79930747e-02 3.85749698e-01
-5.58293819e-01 5.70292771e-01 -8.48756790e-01 -1.46955922e-01
4.86808777e-01 -6.72215641e-01 -6.96330786e-01 1.67954773e-01
-1.09823871e+00 5.38685203e-01 -3.75875175e-01 5.45640409e-01
-5.70274472e-01 -4.16056514e-01 4.21149403e-01 5.79999268e-01
1.81870028e-01 2.34504655e-01 -1.43207401e-01 -4.64138359e-01
-1.05005181e+00 -1.78187396e-02 7.63964117e-01 7.95509934e-01
-8.38077590e-02 8.72073710e-01 -1.95468783e-01 1.33390868e+00
2.08388999e-01 8.80271018e-01 1.26727307e+00 -1.05418384e+00
2.39660576e-01 -6.37271628e-02 -2.77562320e-01 -2.08956912e-01
-1.39802471e-01 -3.20118755e-01 -3.49307656e-01 4.33124363e-01
1.10594422e-01 -1.75431505e-01 -9.29676831e-01 1.90137661e+00
7.50587657e-02 -1.58293813e-01 3.67379367e-01 7.35535264e-01
5.16062796e-01 8.44140708e-01 -3.43522370e-01 -5.60187399e-01
1.13498664e+00 -9.13226843e-01 -1.09224355e+00 -1.70661494e-01
-4.55678031e-02 -1.11234951e+00 1.38812041e+00 4.09802049e-01
-1.36403906e+00 -6.31515741e-01 -1.27462089e+00 -1.37871094e-02
4.64719944e-02 3.23597848e-01 -2.80062437e-01 1.25006533e+00
-1.38471889e+00 2.79179305e-01 -6.56519413e-01 -4.92431186e-02
-2.78226212e-02 3.52521867e-01 -5.00999331e-01 2.20705256e-01
-1.18139875e+00 8.22406292e-01 2.28693441e-01 -4.35186774e-01
-1.00953615e+00 -8.87240350e-01 -7.67656267e-01 1.07565157e-01
-3.65348458e-01 -5.83875299e-01 1.23716044e+00 -7.70480096e-01
-2.24878979e+00 7.93018162e-01 -2.60212630e-01 -4.47031796e-01
5.48847020e-01 9.28116068e-02 -1.10193753e+00 3.65775555e-01
-7.78442845e-02 2.80496508e-01 1.34854627e+00 -8.92858207e-01
-6.48485184e-01 -3.44413131e-01 -6.99370205e-01 2.10554600e-01
-4.52346295e-01 6.90644830e-02 -1.49572387e-01 -1.12426531e+00
1.06173091e-01 -8.38021040e-01 6.90277874e-01 -2.27287158e-01
-5.53436100e-01 1.39707774e-01 9.78089988e-01 -1.49491799e+00
1.36502433e+00 -2.62312365e+00 2.22963125e-01 6.62543327e-02
1.36344045e-01 5.68090498e-01 -2.44529530e-01 4.36364472e-01
-5.00063062e-01 -2.53142178e-01 -6.02611303e-01 -3.87924492e-01
-1.71644259e-02 -6.59662345e-03 -5.17471194e-01 5.00803769e-01
-1.63340688e-01 3.11132222e-01 -6.03223443e-01 -2.06666633e-01
1.72099993e-01 7.53612638e-01 -6.56746209e-01 5.43920994e-01
1.94710344e-01 3.62478644e-01 4.50354457e-01 3.05934459e-01
5.18893301e-01 4.76210535e-01 2.12551206e-01 -7.67076239e-02
-2.35415623e-01 5.91149390e-01 -8.36176634e-01 1.45089734e+00
-3.65201056e-01 7.83431768e-01 3.96591425e-01 -2.24834070e-01
8.65638554e-01 7.11519361e-01 2.18645126e-01 -4.96679664e-01
-1.13012858e-01 5.50382018e-01 4.32480037e-01 -5.87717772e-01
1.76143974e-01 -4.42358077e-01 4.36268926e-01 3.51736158e-01
1.57606542e-01 -4.70181227e-01 -7.92313814e-02 -1.57723680e-01
1.04892492e+00 -3.64636570e-01 1.47653624e-01 -1.35033190e-01
8.07591319e-01 -4.50468272e-01 4.33614016e-01 3.16226602e-01
-3.44669372e-01 8.90902936e-01 2.57963747e-01 5.10746479e-01
-1.26718342e+00 -1.65675628e+00 -4.95180674e-02 8.11306655e-01
-4.28968400e-01 -9.65328291e-02 -1.11363196e+00 -2.28406146e-01
-2.98894167e-01 1.28656518e+00 -2.65729755e-01 -3.66728455e-01
-5.50760329e-01 -1.61647931e-01 9.11634862e-01 4.49298292e-01
9.47104543e-02 -1.02557456e+00 1.00865379e-01 4.23696101e-01
-3.96406829e-01 -8.61298442e-01 -1.32519412e+00 -9.11403541e-03
-4.71919864e-01 -7.66245186e-01 -9.82338309e-01 -1.15086997e+00
3.11258554e-01 -1.08885936e-01 3.20064127e-01 -5.91301620e-01
-1.75872184e-02 4.11954582e-01 -2.03769594e-01 -2.25258768e-01
-1.33937752e+00 -2.91859865e-01 5.77215552e-01 9.08900201e-02
8.68566036e-02 -1.08146966e+00 -3.54305744e-01 7.76072964e-02
-6.61596715e-01 -2.40746975e-01 5.73408008e-01 7.67584920e-01
5.91779649e-01 1.82618216e-01 7.91805506e-01 -5.94050169e-01
1.10090005e+00 -2.01493397e-01 -2.99076945e-01 -9.79021937e-02
-6.24897420e-01 -1.84089810e-01 6.77948833e-01 -7.85129309e-01
-1.08145595e+00 -1.18300349e-01 -7.82229066e-01 -6.52471483e-01
-1.21166170e-01 -3.18513587e-02 -5.40816367e-01 2.55084693e-01
7.00025082e-01 9.06183064e-01 4.76387262e-01 -6.17008865e-01
2.82631099e-01 1.38582373e+00 1.23253834e+00 -1.06032990e-01
6.89806938e-01 1.82535157e-01 -7.77302980e-01 -1.44677234e+00
2.75133196e-02 -2.59468466e-01 -4.29078609e-01 -9.48966518e-02
7.64444828e-01 -7.05593228e-01 -2.28700936e-01 9.99639511e-01
-1.24288821e+00 -2.70422250e-01 -6.66502535e-01 9.05707955e-01
-7.77913272e-01 3.68687749e-01 -6.17519081e-01 -7.78638482e-01
-4.89607900e-01 -1.27654004e+00 4.74821746e-01 -6.12550378e-02
-4.75527316e-01 -6.92617416e-01 4.04485315e-01 5.01114070e-01
4.47151810e-01 -4.13638979e-01 1.56175411e+00 -9.39044774e-01
8.11042562e-02 -1.18605271e-01 3.41132939e-01 1.13050771e+00
5.59163928e-01 -1.95367932e-01 -1.23147726e+00 -7.14565143e-02
4.98433858e-01 3.71126123e-02 2.29106426e-01 4.55404401e-01
7.61248291e-01 -4.64889884e-01 2.91206777e-01 4.60723698e-01
6.04837954e-01 6.22669637e-01 6.57959521e-01 -3.63616765e-01
3.18903536e-01 5.39897799e-01 -9.52364355e-02 3.30312774e-02
2.87543833e-01 5.41289985e-01 -2.55217165e-01 1.11383624e-01
-1.38114226e+00 -4.00529861e-01 9.47607338e-01 1.99634194e+00
3.22567403e-01 -3.15018073e-02 -4.80300903e-01 8.91578972e-01
-8.62390280e-01 -9.52914536e-01 3.20545509e-02 2.04681253e+00
1.28835952e+00 -3.75244059e-02 3.82026047e-01 6.37106895e-01
1.17301464e+00 2.07743689e-01 -6.85884356e-01 -5.12796342e-01
-2.04054281e-01 5.22960842e-01 6.39241710e-02 7.35845268e-01
-1.62873864e-01 8.39609623e-01 6.83485937e+00 8.74992251e-01
-1.48151839e+00 4.91925269e-01 1.32015878e-02 -3.73671860e-01
-2.98756510e-01 -5.96336067e-01 -2.87519783e-01 5.48241258e-01
1.40765035e+00 -4.95803982e-01 7.99429774e-01 6.85749471e-01
3.26875806e-01 5.36698639e-01 -1.07988119e+00 1.01896751e+00
2.69109905e-01 -1.14434004e+00 1.32319421e-01 1.09124027e-01
4.12309587e-01 2.38762167e-03 3.70871693e-01 2.45970711e-01
-1.32473692e-01 -7.80274212e-01 1.10153139e+00 4.74161267e-01
1.34981632e+00 -8.23374569e-01 2.14819685e-01 2.05489710e-01
-8.50883663e-01 -4.44573723e-02 -6.53200895e-02 5.80778956e-01
8.53279009e-02 3.34676713e-01 -1.22316718e+00 6.84025586e-02
4.02198970e-01 -1.75138842e-02 -7.85567164e-02 8.99380684e-01
-4.18358654e-01 1.07464850e+00 -6.78905770e-02 3.33209127e-01
-3.45486611e-01 1.43904284e-01 1.22971964e+00 1.04353344e+00
5.33041060e-01 -1.28302738e-01 -5.39773166e-01 7.81519234e-01
-1.58808023e-01 1.84280589e-01 -1.14798434e-01 -4.70008612e-01
8.89831722e-01 5.55805862e-01 6.76558316e-02 9.40702781e-02
-1.39180049e-01 1.11378610e+00 -3.64708938e-02 4.01862711e-01
-3.88958305e-01 -6.68290973e-01 7.77715385e-01 1.42768100e-01
3.60438555e-01 -7.48958089e-04 -6.56537265e-02 -5.76672912e-01
5.47110178e-02 -1.30187690e+00 3.17324772e-02 -9.53520000e-01
-1.01549554e+00 1.09749532e+00 -6.04123056e-01 -1.13327873e+00
-6.20703340e-01 -3.29233080e-01 -5.99334896e-01 1.28835690e+00
-1.19622958e+00 -9.76047277e-01 3.13193381e-01 8.03489029e-01
9.34063375e-01 -6.69185400e-01 1.12644410e+00 2.60977238e-01
-3.79394233e-01 6.28861725e-01 6.07022792e-02 -4.81100678e-02
7.43399203e-01 -1.24811947e+00 2.93674916e-01 9.70643580e-01
1.73631962e-02 4.66937810e-01 7.29592562e-01 -7.40814269e-01
-1.02089584e+00 -1.04409158e+00 1.10063362e+00 -4.99316454e-02
5.90859592e-01 -1.09410688e-01 -1.11123466e+00 3.98951113e-01
2.89821029e-01 -4.07924205e-01 1.00469208e+00 -5.83112359e-01
-3.49455088e-01 -2.53491819e-01 -1.16600192e+00 4.56085503e-01
7.52936184e-01 -1.15826464e+00 -1.23894489e+00 2.29909718e-01
9.80690897e-01 -5.10061324e-01 -9.49312031e-01 -1.73212215e-01
4.71464097e-01 -8.33251715e-01 7.96934009e-01 -1.95807755e-01
1.74815312e-01 -1.57330647e-01 -3.59719068e-01 -1.73827040e+00
-1.21065572e-01 -1.14151037e+00 -2.08171785e-01 1.34553146e+00
5.06025732e-01 -7.45232284e-01 1.76765159e-01 1.70904785e-01
-6.48749888e-01 -1.97035104e-01 -1.22591102e+00 -9.41011548e-01
2.52029061e-01 -5.22315323e-01 5.23950458e-01 7.11154044e-01
1.57001182e-01 -2.28757970e-02 -3.03815216e-01 6.49610221e-01
4.14804488e-01 -2.60005802e-01 1.52353182e-01 -8.00131559e-01
-3.88862312e-01 -5.23429394e-01 -1.19642287e-01 -6.22910261e-01
1.58196449e-01 -1.06092656e+00 1.11769713e-01 -1.24591529e+00
-6.01265132e-01 2.46117875e-01 3.67194228e-02 1.14825934e-01
3.58114056e-02 -9.32277963e-02 -2.25284398e-02 2.24897236e-01
6.40344739e-01 7.77694881e-01 1.14219427e+00 -9.86730680e-02
-5.03249407e-01 3.26674998e-01 -5.99217117e-01 6.68563128e-01
6.75503671e-01 -6.91010296e-01 -6.84739649e-01 -4.01366167e-02
-5.74741721e-01 5.81910729e-01 -5.60982972e-02 -1.16854072e+00
2.30754837e-01 2.06120953e-01 3.76298092e-02 -5.51154792e-01
5.61837196e-01 -5.15160739e-01 1.16704680e-01 6.85880065e-01
-3.97521108e-01 -1.07187279e-01 2.55748212e-01 3.74450982e-01
-3.47082615e-01 -3.08563381e-01 1.19142187e+00 3.69589031e-01
-2.43713230e-01 3.15421745e-02 -9.41995382e-01 1.31742239e-01
6.51280761e-01 -9.10930261e-02 -2.09541172e-01 -4.00513947e-01
-8.53754520e-01 -5.04793882e-01 8.43470991e-02 5.01642048e-01
7.23120153e-01 -1.24878871e+00 -1.02866697e+00 5.20191014e-01
-1.68175399e-01 -8.11954379e-01 5.50380886e-01 6.47459984e-01
-4.95023072e-01 3.10207725e-01 -2.30704218e-01 -2.81980097e-01
-1.56141162e+00 2.58384407e-01 4.44259405e-01 3.00713986e-01
-9.43760157e-01 9.44088876e-01 2.03713160e-02 -4.20507103e-01
3.23274881e-01 -1.82182893e-01 -7.14700893e-02 -2.39989236e-02
8.03678691e-01 5.54351747e-01 1.86524525e-01 -9.29755747e-01
-2.37911478e-01 1.86862275e-01 -1.72675848e-01 -8.53497028e-01
1.15763986e+00 -4.33680654e-01 -6.77570794e-03 5.17803848e-01
1.24544573e+00 8.42473924e-01 -1.07054365e+00 -8.91666114e-02
-4.75092471e-01 -4.00580056e-02 3.47808093e-01 -1.18011165e+00
-9.95123982e-01 8.92185509e-01 9.21458840e-01 1.38648540e-01
1.37083721e+00 -1.38798609e-01 1.17664599e+00 -2.28629008e-01
-2.36761607e-02 -1.08547151e+00 -2.37161405e-02 4.63651627e-01
1.33093691e+00 -1.31264120e-01 -7.47573674e-01 -2.88023651e-01
-7.81065941e-01 1.28360891e+00 1.04735143e-01 1.42742038e-01
6.94871485e-01 3.91373217e-01 4.75484014e-01 1.86351582e-01
-4.41682756e-01 -1.54854834e-01 3.87835443e-01 1.20315123e+00
1.84410945e-01 2.01848820e-02 -1.07371606e-01 9.21013713e-01
-1.23548436e+00 -3.11625063e-01 5.58787227e-01 3.45696837e-01
-2.94309497e-01 -1.13367355e+00 -6.72415555e-01 1.22697607e-01
-3.08818012e-01 -2.34268129e-01 -4.39281106e-01 4.12916124e-01
8.97748917e-02 1.05787897e+00 4.84622084e-02 -5.94715714e-01
5.94734013e-01 6.09530807e-01 3.34223479e-01 -5.59970021e-01
-6.56606317e-01 4.95857239e-01 1.60461470e-01 -2.66229033e-01
2.61275381e-01 -8.14094245e-01 -1.39375329e+00 -8.98861289e-02
-2.48234216e-02 1.76441431e-01 9.79852498e-01 8.33414197e-01
3.17404747e-01 1.18253493e+00 6.74432218e-01 -2.57708013e-01
-8.17538798e-01 -1.38042247e+00 -9.76417959e-01 8.26665238e-02
8.54451299e-01 -8.29004571e-02 -8.17396998e-01 4.24184531e-01] | [14.749100685119629, 6.486413478851318] |
c695caac-eb3e-4e04-b7ff-c0f021dd53b3 | on-robustness-of-prompt-based-semantic | 2301.12868 | null | https://arxiv.org/abs/2301.12868v3 | https://arxiv.org/pdf/2301.12868v3.pdf | On Robustness of Prompt-based Semantic Parsing with Large Pre-trained Language Model: An Empirical Study on Codex | Semantic parsing is a technique aimed at constructing a structured representation of the meaning of a natural-language question. Recent advancements in few-shot language models trained on code have demonstrated superior performance in generating these representations compared to traditional unimodal language models, which are trained on downstream tasks. Despite these advancements, existing fine-tuned neural semantic parsers are susceptible to adversarial attacks on natural-language inputs. While it has been established that the robustness of smaller semantic parsers can be enhanced through adversarial training, this approach is not feasible for large language models in real-world scenarios, as it requires both substantial computational resources and expensive human annotation on in-domain semantic parsing data. This paper presents the first empirical study on the adversarial robustness of a large prompt-based language model of code, \codex. Our results demonstrate that the state-of-the-art (SOTA) code-language models are vulnerable to carefully crafted adversarial examples. To address this challenge, we propose methods for improving robustness without the need for significant amounts of labeled data or heavy computational resources. | ['Fatemeh Shiri', 'Gholamreza Haffari', 'Weiqing Wang', 'Yuan-Fang Li', 'Yujin Huang', 'Zhuang Li', 'Terry Yue Zhuo'] | 2023-01-30 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 4.15724248e-01 7.88385868e-01 1.75106451e-01 -4.43185240e-01
-1.35391164e+00 -8.89422596e-01 5.26401103e-01 1.26108736e-01
-1.47013918e-01 4.10971552e-01 3.11943442e-01 -7.47316539e-01
4.88483131e-01 -8.75070572e-01 -1.11831033e+00 -1.36051044e-01
-1.34562543e-02 1.19587503e-01 3.88184428e-01 -5.18794537e-01
2.37817138e-01 6.42287806e-02 -1.06433773e+00 2.93274194e-01
8.04904640e-01 3.97940099e-01 -9.66844186e-02 9.08277273e-01
-4.40212607e-01 1.16291964e+00 -8.86768043e-01 -8.66547048e-01
2.16103807e-01 -4.01486456e-01 -9.46329534e-01 -4.56615627e-01
3.84868354e-01 -1.64840072e-01 -4.43463296e-01 1.48027778e+00
3.77297252e-01 -6.29186705e-02 3.72200012e-01 -1.14979017e+00
-1.26116443e+00 9.56920147e-01 -1.41145959e-01 1.11977987e-01
3.97714198e-01 3.47578466e-01 1.01349604e+00 -5.71176231e-01
6.61254168e-01 1.37747431e+00 8.24465394e-01 1.20977283e+00
-1.21999061e+00 -7.29046881e-01 1.21221326e-01 -3.50043327e-01
-9.63201165e-01 -5.41979194e-01 7.03905880e-01 -4.80631173e-01
1.23471141e+00 -1.70053467e-01 -2.71042228e-01 1.64953780e+00
3.47254217e-01 4.96854365e-01 7.00879157e-01 -5.25542021e-01
5.51022828e-01 1.55441254e-01 1.84605837e-01 8.83988440e-01
1.38004526e-01 2.00899288e-01 -2.33434975e-01 -4.72105116e-01
1.91534877e-01 -3.93691540e-01 3.05433944e-02 -2.82596678e-01
-5.69746614e-01 1.26927686e+00 4.97368366e-01 3.16578060e-01
1.49026528e-01 6.86312318e-01 9.29755688e-01 3.83047312e-01
5.63887417e-01 8.72970402e-01 -6.08520627e-01 -2.38140464e-01
-7.54308343e-01 6.43587671e-03 7.75106013e-01 1.14903045e+00
4.37945843e-01 3.43688697e-01 4.22716625e-02 5.25707841e-01
2.10939482e-01 3.53291243e-01 5.61363578e-01 -8.83122504e-01
8.46438169e-01 4.64987159e-01 -9.74481404e-02 -7.80148447e-01
-1.26918331e-01 -7.11009577e-02 -2.76688844e-01 5.42633832e-01
4.01653826e-01 -3.60591978e-01 -9.77438509e-01 2.01462936e+00
-2.09850058e-01 1.92118540e-01 6.56666160e-01 4.21546280e-01
7.84631491e-01 5.58851779e-01 6.10080838e-01 5.29070854e-01
1.29750693e+00 -9.42658007e-01 -3.55438083e-01 -8.20839286e-01
7.06642032e-01 -7.06282556e-01 1.42814398e+00 -8.38381797e-02
-8.99188995e-01 -3.57391417e-01 -1.16770864e+00 -1.33663416e-01
-5.23257852e-01 -5.50063491e-01 7.32395351e-01 1.15205216e+00
-1.03965247e+00 5.00468731e-01 -8.49759638e-01 -2.13871017e-01
6.76412225e-01 1.48792595e-01 -4.41777438e-01 -2.43662223e-01
-1.38011646e+00 9.31106985e-01 1.39505029e-01 -3.90541583e-01
-1.32866919e+00 -6.83162093e-01 -1.44726396e+00 2.86021829e-01
3.35717708e-01 -4.23376411e-01 1.53920233e+00 -1.24245119e+00
-1.38185179e+00 9.67430353e-01 7.37828761e-02 -6.90611601e-01
2.25976333e-01 -3.85511369e-01 -3.10869902e-01 1.60544410e-01
2.13150263e-01 6.05297387e-01 1.01176870e+00 -1.31343246e+00
3.76921520e-02 -6.67852014e-02 5.94916105e-01 -3.01439613e-01
-3.05051565e-01 4.38100517e-01 -1.13069281e-01 -9.74739790e-01
-4.68626618e-01 -9.48660374e-01 -5.86948812e-01 -4.76669297e-02
-2.93334067e-01 3.38434428e-03 6.06553257e-01 -7.03352511e-01
8.56628418e-01 -2.34111714e+00 -1.25511289e-01 -2.28722274e-01
-1.92421585e-01 4.91724581e-01 -5.25410414e-01 3.55059355e-01
-2.15970308e-01 5.49575806e-01 -7.11346388e-01 -3.31777453e-01
-1.75313719e-04 2.59030879e-01 -9.06576395e-01 3.13440889e-01
6.35489821e-01 9.55587804e-01 -1.16505826e+00 -2.26680890e-01
-1.24693245e-01 3.14711481e-01 -9.85614777e-01 4.75768477e-01
-6.04861379e-01 2.24193871e-01 -5.25413156e-01 6.65197015e-01
3.16936255e-01 -1.39619157e-01 -1.04953565e-01 7.14163184e-01
4.99026567e-01 2.08214417e-01 -4.61033791e-01 2.16140032e+00
-8.13875854e-01 6.52751088e-01 6.88676462e-02 -1.01232946e+00
9.08755064e-01 4.69250679e-01 -2.61481404e-01 -5.39117813e-01
1.43490314e-01 1.70430496e-01 -1.83560058e-01 -5.87173283e-01
4.21305388e-01 -3.57443750e-01 -9.38152492e-01 5.48820436e-01
1.79198802e-01 -5.04325986e-01 -1.54628396e-01 5.59014618e-01
1.69965291e+00 1.88306496e-01 1.43297285e-01 -2.47752190e-01
4.92266476e-01 2.44450495e-01 3.47022861e-01 8.86064887e-01
-3.48821729e-01 6.20606959e-01 6.54048324e-01 -3.30521584e-01
-1.12953603e+00 -1.06881106e+00 1.32576063e-01 1.33891833e+00
1.90614611e-02 -2.97931433e-01 -1.35298896e+00 -1.06663084e+00
-1.29015252e-01 1.17224979e+00 -7.62698054e-01 -6.28801465e-01
-6.24409080e-01 -3.77413571e-01 1.33269513e+00 7.06363797e-01
2.00683162e-01 -1.30109179e+00 -5.99202633e-01 2.60142595e-01
1.33808255e-01 -1.14001822e+00 -3.36486071e-01 2.22244740e-01
-6.51372671e-01 -1.20150471e+00 -3.30153853e-01 -9.76746500e-01
8.76804829e-01 -1.36897370e-01 1.47142851e+00 2.74514377e-01
-4.06158417e-01 5.46593785e-01 -5.69808900e-01 -4.00213152e-01
-1.28879344e+00 4.86317314e-02 -3.20171028e-01 -5.72879672e-01
4.74571526e-01 -4.68755871e-01 -1.41777217e-01 -1.23455979e-01
-1.19920576e+00 -4.13850486e-01 2.93935150e-01 8.80111217e-01
-5.73286004e-02 -1.81102827e-01 8.47337663e-01 -1.26182568e+00
9.40450728e-01 -7.67448366e-01 -6.75598621e-01 2.80920237e-01
-2.53949702e-01 5.60172856e-01 1.22509468e+00 -4.32832330e-01
-1.21613538e+00 3.77720706e-02 -4.49746132e-01 -2.17690617e-01
-2.72090077e-01 2.43140131e-01 -6.65123239e-02 -2.46212989e-01
1.29674506e+00 -1.05010085e-02 -1.90558687e-01 -3.58921528e-01
7.72114754e-01 5.68506300e-01 8.29165757e-01 -9.56474185e-01
1.15133286e+00 3.76974761e-01 -3.34279120e-01 -3.14256281e-01
-9.12102699e-01 -1.55056696e-02 -3.40455681e-01 3.90438050e-01
1.13160074e+00 -9.89855051e-01 -1.57599263e-02 2.18021795e-01
-1.38306177e+00 -4.18646693e-01 -3.37363511e-01 -1.84761897e-01
-7.24925578e-01 4.98759449e-01 -7.92756259e-01 -4.27542597e-01
-4.92555887e-01 -1.32489121e+00 9.93730485e-01 1.36999534e-02
-5.11805832e-01 -1.16810369e+00 1.36117339e-01 4.79160011e-01
6.02355540e-01 3.88447821e-01 1.21049285e+00 -9.45596397e-01
-4.83270705e-01 -4.46410716e-01 5.45075200e-02 5.53381324e-01
-2.10890308e-01 -1.47935897e-01 -1.17969334e+00 -2.54479289e-01
1.53167620e-01 -8.59465182e-01 5.80802441e-01 -1.60381839e-01
1.09882283e+00 -2.87065595e-01 8.06004107e-02 5.43995917e-01
1.52064180e+00 -2.28951089e-02 5.65123916e-01 2.67909169e-01
4.62154955e-01 5.38084805e-01 4.36342567e-01 9.80923846e-02
-2.17732396e-02 1.51623547e-01 7.72850096e-01 1.63488120e-01
-1.89215183e-01 -5.48472345e-01 6.76511228e-01 4.74774361e-01
5.84569752e-01 -1.67867154e-01 -1.20626163e+00 7.61905193e-01
-1.50653934e+00 -8.33025575e-01 2.00187817e-01 1.79526901e+00
8.75554681e-01 4.61834490e-01 -6.85393691e-01 -2.06490993e-01
6.66374385e-01 2.13756055e-01 -6.10023618e-01 -9.62874651e-01
1.28958791e-01 6.47387087e-01 4.97742176e-01 6.22089207e-01
-1.25132883e+00 1.41995585e+00 6.98462820e+00 6.04401171e-01
-8.09540808e-01 3.98461878e-01 4.39791292e-01 1.20185845e-01
-4.93592143e-01 3.05012763e-01 -4.24771130e-01 3.81283343e-01
1.41573250e+00 -3.15572739e-01 5.21319628e-01 1.50944638e+00
-2.37284809e-01 1.95308983e-01 -1.10596800e+00 3.19743752e-01
2.31034011e-01 -1.28189731e+00 -9.61270109e-02 -4.44649130e-01
7.86522627e-01 2.42614940e-01 2.66610943e-02 8.37183535e-01
1.08668625e+00 -1.25985694e+00 8.57717872e-01 1.33545585e-02
9.15050030e-01 -7.16494024e-01 6.17799222e-01 3.46294701e-01
-9.27098453e-01 -3.08998287e-01 -5.21821260e-01 -8.01321939e-02
2.46482734e-02 1.20794497e-01 -6.11070037e-01 3.19559574e-02
4.87464249e-01 2.23402888e-01 -8.31640482e-01 3.47788095e-01
-6.63908243e-01 7.76908517e-01 2.12285087e-01 1.18386202e-01
5.29775500e-01 5.39762139e-01 2.92011321e-01 1.25668478e+00
2.10299388e-01 -6.72639236e-02 6.74309209e-02 1.11515403e+00
-4.85856682e-01 -1.92194507e-01 -1.02539575e+00 -3.20243657e-01
4.04994607e-01 8.79703224e-01 -6.32263899e-01 -4.07850862e-01
-8.33605707e-01 1.08364642e+00 5.01544952e-01 2.64702916e-01
-9.69178855e-01 -5.60425460e-01 8.76577914e-01 -5.66304363e-02
1.52032912e-01 -1.52275950e-01 -4.91000295e-01 -1.19598770e+00
-1.31554544e-01 -1.22643518e+00 4.22060668e-01 -7.78499186e-01
-1.40033484e+00 7.11543918e-01 -4.69805330e-01 -1.00781977e+00
-3.48872870e-01 -6.78027511e-01 -8.92333686e-01 8.34359646e-01
-1.36015463e+00 -1.23776531e+00 1.42661273e-01 5.61153293e-01
9.03693080e-01 -5.54990888e-01 1.49321866e+00 -3.73497456e-02
-2.99807250e-01 8.11887622e-01 2.33631916e-02 5.15738666e-01
6.24322832e-01 -1.17176139e+00 1.27202618e+00 1.42223239e+00
-7.90853128e-02 8.51064146e-01 9.23414767e-01 -6.10143721e-01
-1.30163515e+00 -1.24148190e+00 7.34404564e-01 -9.58807886e-01
1.09342277e+00 -6.75065935e-01 -1.04777455e+00 7.98944950e-01
8.16294625e-02 3.95604461e-01 8.06673646e-01 -3.93706411e-02
-1.05993509e+00 5.81489503e-01 -1.37863755e+00 5.83937168e-01
9.85136688e-01 -1.02768314e+00 -1.08719480e+00 3.16071838e-01
1.31419992e+00 -2.46427476e-01 -7.14475453e-01 1.00503013e-01
-1.48562476e-01 -7.36461520e-01 9.83119309e-01 -1.08972096e+00
9.99326944e-01 -1.29385352e-01 -1.75206959e-01 -1.20393932e+00
6.87838346e-02 -6.47617936e-01 1.47131324e-01 1.14285278e+00
4.56612915e-01 -4.04838532e-01 6.90367520e-01 1.10814893e+00
-3.21303040e-01 -1.73302233e-01 -1.05641294e+00 -8.47391963e-01
7.45714843e-01 -6.40647113e-01 3.92701149e-01 1.01080239e+00
2.22796649e-01 3.45984071e-01 -1.16308704e-01 4.63418275e-01
3.92754972e-01 -1.22890398e-01 6.68384552e-01 -7.83242464e-01
-5.41684806e-01 -2.58187503e-01 -5.17192662e-01 -3.81701857e-01
9.17856872e-01 -1.00769639e+00 3.69265676e-01 -1.21638787e+00
1.87499188e-02 -3.19187224e-01 -6.88154623e-02 6.28826857e-01
-4.72979516e-01 5.87496907e-02 1.84142068e-01 -8.23862255e-02
-5.19582868e-01 3.73449475e-01 6.12030089e-01 -3.63512665e-01
3.17972451e-01 -2.49635205e-01 -1.07086527e+00 9.41615999e-01
8.51975262e-01 -9.67096448e-01 -7.06326842e-01 -7.40033746e-01
3.26220125e-01 3.16574611e-02 3.01698864e-01 -1.05682504e+00
-2.60989051e-02 2.19236761e-02 -2.41585121e-01 4.32651550e-01
7.98408315e-02 -6.69640303e-01 -3.96368116e-01 6.70338631e-01
-6.00147843e-01 -8.51457939e-02 5.13317347e-01 7.13429034e-01
-2.35390365e-01 -8.64227831e-01 9.87854898e-01 -5.59619546e-01
-8.55683327e-01 -3.20279133e-03 -5.37392795e-01 6.41147017e-01
1.16037333e+00 2.48419270e-01 -5.82676947e-01 -3.06096554e-01
-5.91681719e-01 -1.20045722e-01 8.11830878e-01 8.88878465e-01
3.91973525e-01 -8.10457826e-01 -5.06663918e-01 9.55425873e-02
2.86557436e-01 -5.31473197e-02 8.94099101e-02 -3.00085843e-01
-6.73280180e-01 1.52924344e-01 -8.35789368e-02 2.49737855e-02
-1.07913768e+00 1.22483504e+00 2.07261622e-01 -3.64027143e-01
-6.39244854e-01 1.20134795e+00 2.62067109e-01 -6.00476563e-01
1.39694646e-01 -1.36716887e-01 1.21384650e-01 -5.45424044e-01
5.86458623e-01 -2.62234896e-01 -6.58393130e-02 -4.20527935e-01
-2.99163342e-01 2.83172637e-01 4.94463034e-02 2.92403623e-02
1.21156454e+00 2.93462545e-01 5.25986999e-02 2.14730687e-02
1.11961591e+00 2.20911741e-01 -1.29424453e+00 -1.52798548e-01
3.54288787e-01 -3.22528958e-01 -3.28442872e-01 -6.07975960e-01
-8.35715413e-01 1.21615052e+00 1.92834571e-01 1.34739786e-01
7.47442603e-01 2.93319952e-02 9.85158265e-01 5.48367739e-01
5.05484343e-01 -9.35647488e-01 3.84205759e-01 7.88177490e-01
7.35991478e-01 -1.32129419e+00 -4.64832067e-01 -3.74634802e-01
-7.14015961e-01 1.06766522e+00 6.70448542e-01 -3.76935989e-01
4.64560181e-01 6.74458206e-01 3.87566000e-01 -7.88291171e-02
-6.16533399e-01 1.73546940e-01 -3.62444043e-01 8.63006175e-01
4.04146701e-01 -4.61502150e-02 1.03700817e-01 7.97215223e-01
-2.78604686e-01 -3.63211632e-01 8.16367388e-01 1.25659478e+00
-3.91129643e-01 -1.13923228e+00 -3.99306357e-01 -5.91973439e-02
-9.24423277e-01 -5.38888335e-01 -4.04688954e-01 4.44043130e-01
-1.22239977e-01 1.11275876e+00 -1.82452843e-01 -2.06622779e-01
1.46907061e-01 4.36287373e-01 4.43599410e-02 -1.23218238e+00
-8.39254379e-01 -8.13336790e-01 3.03089738e-01 -5.61588645e-01
3.32412720e-02 -1.79908469e-01 -1.56425869e+00 -1.31384462e-01
7.31709301e-02 1.24603011e-01 7.38529563e-01 7.83717215e-01
4.58124399e-01 6.21563792e-01 3.45640928e-01 -6.26947343e-01
-9.70567465e-01 -6.92484736e-01 -5.67290448e-02 8.51743698e-01
1.52307004e-01 -2.07233652e-01 -5.00189245e-01 3.45711201e-01] | [7.040821075439453, 7.926254749298096] |
ca060c59-7417-4cae-9efa-72625f09901f | negation-scope-detection-for-twitter | null | null | https://aclanthology.org/W15-2914 | https://aclanthology.org/W15-2914.pdf | Negation Scope Detection for Twitter Sentiment Analysis | null | ['Bj{\\"o}rn Gamb{\\"a}ck', 'J{\\o}rgen Faret', 'Lars Bungum', 'Johan Reitan'] | 2015-09-01 | null | null | null | ws-2015-9 | ['twitter-sentiment-analysis', 'negation-detection'] | ['natural-language-processing', 'natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.256717681884766, 3.84771466255188] |
43fde8f4-5bac-4842-8206-fcd601afe822 | affinity-aware-graph-networks | 2206.11941 | null | https://arxiv.org/abs/2206.11941v1 | https://arxiv.org/pdf/2206.11941v1.pdf | Affinity-Aware Graph Networks | Graph Neural Networks (GNNs) have emerged as a powerful technique for learning on relational data. Owing to the relatively limited number of message passing steps they perform -- and hence a smaller receptive field -- there has been significant interest in improving their expressivity by incorporating structural aspects of the underlying graph. In this paper, we explore the use of affinity measures as features in graph neural networks, in particular measures arising from random walks, including effective resistance, hitting and commute times. We propose message passing networks based on these features and evaluate their performance on a variety of node and graph property prediction tasks. Our architecture has lower computational complexity, while our features are invariant to the permutations of the underlying graph. The measures we compute allow the network to exploit the connectivity properties of the graph, thereby allowing us to outperform relevant benchmarks for a wide variety of tasks, often with significantly fewer message passing steps. On one of the largest publicly available graph regression datasets, OGB-LSC-PCQM4Mv1, we obtain the best known single-model validation MAE at the time of writing. | ['Sreenivas Gollapudi', 'Petar Veličković', 'Ira Ktena', 'Ali Kemal Sinop', 'Ameya Velingker'] | 2022-06-23 | null | null | null | null | ['graph-property-prediction', 'graph-regression'] | ['graphs', 'graphs'] | [ 1.15244448e-01 1.34279191e-01 -4.77987498e-01 -2.61984169e-01
7.98630640e-02 -4.68517601e-01 6.52688682e-01 8.17869604e-01
-5.74631155e-01 5.93710721e-01 -1.60411485e-02 -6.82753921e-01
-4.97994810e-01 -1.25613701e+00 -7.88159251e-01 -4.11020964e-01
-7.24559963e-01 5.60216486e-01 3.61732185e-01 -4.66755748e-01
2.58502215e-01 8.30706000e-01 -1.24812174e+00 5.03856093e-02
3.71800959e-01 8.91912699e-01 -3.72436009e-02 7.71757185e-01
1.01583473e-01 1.05492592e+00 -2.73994595e-01 -4.07555968e-01
2.52361167e-02 -1.44418627e-01 -9.67418253e-01 -4.03712481e-01
3.62177610e-01 8.41454118e-02 -1.00349569e+00 7.57082224e-01
2.70689905e-01 3.84290606e-01 6.82965457e-01 -1.22289503e+00
-5.93767822e-01 9.17795599e-01 -3.32824081e-01 4.64560896e-01
2.66340505e-02 -6.61411434e-02 1.67804837e+00 -3.32114428e-01
7.70838141e-01 1.04429865e+00 8.57337236e-01 2.66857743e-01
-1.58918667e+00 -3.66917014e-01 -1.40992356e-02 3.83657247e-01
-1.37574637e+00 -2.84851044e-01 6.11908674e-01 -7.68292248e-02
1.46424878e+00 2.12136403e-01 7.16198742e-01 7.61249304e-01
4.20093507e-01 4.36762273e-01 8.37983608e-01 -3.13042283e-01
1.13395110e-01 -1.89504042e-01 3.03173989e-01 1.11421919e+00
4.20814365e-01 -6.61329851e-02 -3.06028694e-01 -4.06520478e-02
7.68174529e-01 1.24273516e-01 -2.27744728e-01 -6.12590730e-01
-1.06093109e+00 9.42653894e-01 1.09919584e+00 3.26594830e-01
-4.53072041e-02 6.28001690e-01 4.23313677e-01 5.79269767e-01
3.66125196e-01 6.31620169e-01 -4.32879865e-01 -2.69924942e-03
-4.40547824e-01 3.87214385e-02 1.31191564e+00 7.10938752e-01
9.97738600e-01 -8.31891894e-02 1.11729698e-02 8.15408945e-01
6.92317784e-02 1.43390968e-01 1.30709901e-01 -6.02946818e-01
4.76752430e-01 8.84802341e-01 -7.86585093e-01 -1.43532550e+00
-7.83320844e-01 -4.92472589e-01 -1.15208137e+00 -1.88830584e-01
3.41113508e-01 2.43505523e-01 -9.39606011e-01 1.64538169e+00
-5.79880625e-02 -4.54304740e-02 -2.35093892e-01 4.41672355e-01
8.62908185e-01 5.18174767e-01 -9.53223705e-02 2.03273073e-01
9.78260279e-01 -7.21394658e-01 -4.07607481e-02 -2.57933289e-01
1.20459402e+00 -2.01059550e-01 8.10861588e-01 1.42264068e-01
-8.41299176e-01 -2.29758352e-01 -1.16036952e+00 -1.18955031e-01
-6.96601927e-01 -5.04224777e-01 1.12036169e+00 4.92985755e-01
-1.48899698e+00 1.17535150e+00 -8.74281406e-01 -5.65662384e-01
4.08195674e-01 7.24519908e-01 -5.62045872e-01 -1.53773919e-01
-1.09841859e+00 7.77432799e-01 4.73572254e-01 -4.32252288e-02
-4.42511588e-01 -4.10299689e-01 -8.96465898e-01 3.82196218e-01
4.57127959e-01 -4.74862278e-01 7.72443831e-01 -4.88048196e-01
-1.11234677e+00 6.31085694e-01 2.54709244e-01 -6.41735911e-01
1.69717208e-01 5.26080370e-01 -3.07572305e-01 8.16903412e-02
-3.07261467e-01 6.34009063e-01 4.11999911e-01 -6.74255490e-01
-2.07440719e-01 -2.65957594e-01 3.33016247e-01 -7.40452949e-03
-5.64010262e-01 -4.63505059e-01 -4.40691978e-01 -3.76467437e-01
8.72211233e-02 -1.10033226e+00 -3.48586351e-01 -3.17688823e-01
-5.87154031e-01 -4.36073750e-01 4.94572967e-01 -7.90513994e-04
1.25512397e+00 -1.85308480e+00 1.98231503e-01 7.26067603e-01
9.37526286e-01 8.23926032e-02 -3.89967620e-01 7.76671410e-01
-8.59566405e-02 3.60907584e-01 1.52661896e-03 1.55251306e-02
-2.19138756e-01 4.02704567e-01 5.08938022e-02 5.57424664e-01
2.76528239e-01 1.19784892e+00 -6.83211684e-01 -3.95017982e-01
-2.36301385e-02 3.79924506e-01 -6.44266605e-01 -1.70796722e-01
-2.49948874e-01 -8.05727914e-02 -3.62748861e-01 3.09166610e-01
2.37975314e-01 -9.27487314e-01 5.37599742e-01 -1.53882103e-02
3.79274040e-01 5.56878626e-01 -8.98888767e-01 1.38386476e+00
-4.21028167e-01 9.58060265e-01 -1.99453637e-01 -1.31067002e+00
8.12994838e-01 -1.49176091e-01 4.48311150e-01 -8.48639846e-01
1.51561052e-01 -2.08427355e-01 4.43078578e-01 4.64818105e-02
5.58228791e-01 3.17637116e-01 7.70699158e-02 5.15089214e-01
1.23373233e-01 1.66718706e-01 6.11423194e-01 6.05520248e-01
1.91469550e+00 -4.55493599e-01 2.05938980e-01 -5.22138834e-01
2.13418469e-01 -2.22974032e-01 8.62220824e-02 1.02301228e+00
2.39569582e-02 1.91220164e-01 9.59032834e-01 -7.28883743e-01
-9.61458623e-01 -1.00670397e+00 2.62616016e-03 1.34875739e+00
3.40028517e-02 -8.94064248e-01 -2.70438969e-01 -4.96742159e-01
1.77836969e-01 2.17476234e-01 -8.04172873e-01 -3.76411885e-01
-6.43220365e-01 -6.56608760e-01 7.13913858e-01 6.49095833e-01
9.73968357e-02 -1.14254510e+00 -7.61215240e-02 3.62047017e-01
4.31097269e-01 -1.21175551e+00 -2.16875866e-01 5.46296358e-01
-1.17908955e+00 -1.15281129e+00 -3.89869660e-02 -8.08975160e-01
5.99434018e-01 2.34446391e-01 1.55467284e+00 7.88432837e-01
-4.98639077e-01 2.35441551e-01 -1.48449838e-01 -2.59172656e-02
-3.29922438e-01 7.10514903e-01 -2.32822284e-01 -2.64100283e-01
1.68442786e-01 -9.01264846e-01 -4.28188831e-01 1.72368810e-01
-8.52203965e-01 4.77402918e-02 6.56556785e-01 7.18726814e-01
3.85360450e-01 1.84824228e-01 1.82952374e-01 -1.26888514e+00
7.65784144e-01 -4.61577564e-01 -6.54452145e-01 1.52448952e-01
-9.05200958e-01 4.08739269e-01 8.01386833e-01 -2.78196007e-01
-2.26779357e-01 -4.57617760e-01 1.36853278e-01 5.72511330e-02
1.69760987e-01 8.16905975e-01 2.17136607e-01 -5.25268376e-01
6.93015933e-01 -4.93090302e-02 1.75267924e-02 -9.40718502e-02
3.56244504e-01 8.24505761e-02 1.54831439e-01 -4.88605171e-01
6.64138436e-01 2.32263699e-01 8.43896329e-01 -1.00850010e+00
-4.41174537e-01 -3.41252506e-01 -4.83276725e-01 3.26918550e-02
5.11917174e-01 -5.63325405e-01 -1.25126266e+00 3.27954382e-01
-8.50463748e-01 -5.30843556e-01 8.19268823e-02 3.45477283e-01
-2.35522926e-01 2.40611613e-01 -1.17657197e+00 -4.25134659e-01
-2.18834281e-01 -9.95403707e-01 4.98513728e-01 -4.67576087e-02
-1.16307415e-01 -1.43521190e+00 -1.13484725e-01 6.39247335e-03
6.05646372e-01 2.65176445e-01 1.54910350e+00 -9.07716930e-01
-9.31601226e-01 -2.57409126e-01 -5.62588394e-01 1.19844578e-01
-5.75599857e-02 2.80496310e-02 -4.59048271e-01 -4.60701287e-01
-7.57030904e-01 -3.40304285e-01 1.27220774e+00 3.67184162e-01
1.25646114e+00 -1.56387284e-01 -5.40277839e-01 7.97153652e-01
1.49578595e+00 -2.43345454e-01 5.11163890e-01 3.58691394e-01
1.14830458e+00 2.51539320e-01 -2.22525164e-01 2.72884201e-02
4.75083560e-01 3.52064908e-01 6.23753965e-01 -1.21856570e-01
-5.27730137e-02 -1.90155163e-01 -5.16992807e-02 1.04184580e+00
-2.19677165e-01 -5.93156636e-01 -1.15249789e+00 2.72402167e-01
-1.75642633e+00 -6.47422850e-01 -2.28827596e-01 2.03366828e+00
5.45162261e-01 5.83414197e-01 5.01434691e-02 1.52300587e-02
5.18886745e-01 5.91951787e-01 -3.73829812e-01 -5.53915620e-01
-5.75521551e-02 5.98229706e-01 9.57719922e-01 4.12353307e-01
-8.99615228e-01 8.23138893e-01 6.43836403e+00 6.57127976e-01
-1.02503681e+00 -3.73737663e-01 7.36571848e-01 -3.96643989e-02
-2.51938224e-01 3.44295464e-02 -5.70851445e-01 3.66192050e-02
1.32561064e+00 1.18648879e-01 9.00299788e-01 5.76397777e-01
-4.54491794e-01 4.19501066e-02 -1.53497124e+00 6.73367858e-01
-9.08017084e-02 -1.64525044e+00 6.71734810e-02 4.06712264e-01
3.93598109e-01 5.42928517e-01 -1.65452555e-01 5.18000066e-01
5.46340764e-01 -1.49114931e+00 1.88594565e-01 4.13663626e-01
7.37540424e-01 -8.23327303e-01 5.72720945e-01 1.24492578e-01
-1.34480131e+00 -1.88214276e-02 -4.19531435e-01 -3.22301805e-01
-4.09980059e-01 4.46290165e-01 -9.67117250e-01 4.38051105e-01
5.84780753e-01 7.93296635e-01 -9.37128246e-01 7.70487368e-01
3.31528708e-02 6.66529953e-01 -4.41114813e-01 -7.18409359e-01
2.02026725e-01 -1.26652703e-01 3.05281281e-01 1.13198221e+00
-4.40835431e-02 -1.76345512e-01 5.16712666e-02 6.91391885e-01
-7.25032866e-01 8.65711197e-02 -8.44410777e-01 -5.23481548e-01
4.55878764e-01 1.24781489e+00 -1.03505576e+00 3.56547441e-03
-4.37777907e-01 4.78819072e-01 1.01402795e+00 2.35405624e-01
-5.43804169e-01 -6.66853428e-01 3.23633432e-01 2.59099275e-01
3.35711509e-01 -5.26886761e-01 -2.92171203e-02 -8.91625524e-01
-8.23142603e-02 -8.04443955e-01 4.89656270e-01 -4.50265676e-01
-1.30178189e+00 7.14433908e-01 -1.01228617e-01 -4.28273439e-01
-9.41520631e-02 -8.59612584e-01 -5.41669726e-01 6.87831700e-01
-1.30329013e+00 -9.28833604e-01 -1.59675151e-01 4.87867117e-01
-1.90283030e-01 -2.02902883e-01 7.87205219e-01 2.00371042e-01
-4.99519587e-01 5.05032301e-01 1.66209519e-01 3.95824105e-01
3.90581697e-01 -1.34296024e+00 8.92275274e-01 3.86865616e-01
6.05740309e-01 7.27324903e-01 4.40959573e-01 -4.21187311e-01
-1.77718651e+00 -8.06407750e-01 6.76662683e-01 -4.74974692e-01
1.06778932e+00 -6.89388096e-01 -9.29028332e-01 8.52295041e-01
-1.24949440e-01 2.74837554e-01 4.41441834e-01 7.27529109e-01
-5.46371996e-01 -3.84601861e-01 -4.90252823e-01 6.36255264e-01
1.33265269e+00 -5.88819981e-01 -8.26639682e-02 2.46866763e-01
6.65001631e-01 -1.94089547e-01 -1.05131400e+00 4.84980166e-01
4.14348274e-01 -9.07950580e-01 1.01636732e+00 -8.01082671e-01
4.26915199e-01 1.92300975e-01 2.83425022e-02 -1.30978429e+00
-7.18430877e-01 -4.54357445e-01 -1.85529351e-01 7.44076967e-01
7.31468499e-01 -8.21392536e-01 1.05978894e+00 3.54704887e-01
1.56672865e-01 -1.18704045e+00 -7.58918822e-01 -6.07002616e-01
-3.48489434e-02 -3.19350779e-01 5.03954351e-01 8.25301766e-01
-8.07928890e-02 8.30188274e-01 -1.61596388e-01 -1.46043211e-01
4.74363625e-01 2.90555488e-02 7.64316201e-01 -1.67618871e+00
-4.70830292e-01 -6.72835827e-01 -7.56696463e-01 -9.98270035e-01
2.49938726e-01 -1.49833465e+00 -3.47793996e-01 -1.51229417e+00
1.49201676e-01 -5.37606418e-01 -4.58877981e-01 6.35116875e-01
7.39887729e-02 1.92012861e-01 -1.01390079e-01 1.59483194e-01
-8.69635105e-01 3.26050729e-01 1.15843070e+00 -1.87162817e-01
-1.35726690e-01 -1.37493819e-01 -6.08051896e-01 5.06432533e-01
7.83009291e-01 -5.53404093e-01 -4.99572963e-01 -4.73061681e-01
8.10431182e-01 5.26576377e-02 2.29608670e-01 -1.09550226e+00
3.85565370e-01 4.35385182e-02 4.17990297e-01 -2.89805144e-01
3.20207775e-01 -5.23274064e-01 -8.68356973e-03 5.41801631e-01
-5.42478502e-01 5.51123977e-01 1.69757381e-01 8.85288596e-01
1.20727666e-01 -7.17270449e-02 4.44778651e-01 -2.71785222e-02
-5.03951728e-01 6.61931336e-01 -2.25614399e-01 1.56964138e-01
6.42494023e-01 -1.54502660e-01 -6.00639701e-01 -4.71648246e-01
-3.80326211e-01 1.05800100e-01 5.28252304e-01 3.04878414e-01
4.32433456e-01 -1.09395981e+00 -4.50433910e-01 1.52700022e-01
1.85019761e-01 -1.85562611e-01 -6.29689023e-02 7.55819440e-01
-8.35870266e-01 4.21873957e-01 -2.25792065e-01 -5.15565574e-01
-1.19767869e+00 5.22197545e-01 2.86441952e-01 -6.80112183e-01
-7.90818989e-01 7.66937673e-01 -1.29911333e-01 -5.71561635e-01
1.80786580e-01 -3.04942489e-01 -5.45844287e-02 -2.30219036e-01
-7.05031678e-03 3.92455786e-01 3.00086081e-01 -4.07999247e-01
-3.68693113e-01 1.46209747e-01 -3.50283831e-01 3.95538092e-01
1.54157782e+00 3.08044910e-01 -5.34208715e-01 3.46308321e-01
1.38881254e+00 -2.27436766e-01 -7.95584142e-01 -3.52328837e-01
4.06690687e-01 -1.48038015e-01 1.52483806e-01 -4.09994721e-01
-1.20338380e+00 7.15982199e-01 1.24776356e-01 8.82137597e-01
7.63409317e-01 -6.67833025e-03 7.36117423e-01 1.07176709e+00
2.51740992e-01 -8.92516255e-01 1.18372284e-01 8.63454401e-01
3.81193757e-01 -1.09409738e+00 3.25351417e-01 -2.77777940e-01
-2.00410839e-02 1.18228400e+00 5.24887264e-01 -2.51874328e-01
7.47893155e-01 1.73372015e-01 -4.28518713e-01 -5.14432311e-01
-9.84526753e-01 -3.91786769e-02 3.61447215e-01 4.74555373e-01
4.32271063e-01 1.56831786e-01 4.97527560e-03 -1.20025344e-01
-2.24423185e-01 -4.68887031e-01 6.21830881e-01 7.13294029e-01
-3.02649230e-01 -1.19560790e+00 4.42560911e-01 1.09246206e+00
-4.77729559e-01 -3.73960078e-01 -5.80234289e-01 1.10599291e+00
-6.71919525e-01 8.24765742e-01 1.97030187e-01 -5.31940281e-01
1.74228579e-01 -2.79259324e-01 5.69742560e-01 -6.06555462e-01
-4.48298126e-01 -5.13276517e-01 4.49932724e-01 -6.62387073e-01
-1.81889877e-01 -1.29609331e-01 -1.16296828e+00 -8.97713482e-01
-4.03973520e-01 8.64615813e-02 5.85452974e-01 7.40107179e-01
4.22809422e-01 6.28082395e-01 3.27721447e-01 -6.65468454e-01
-4.34892565e-01 -8.84322524e-01 -8.85592282e-01 4.63753462e-01
2.76300907e-01 -3.94381851e-01 -2.89385617e-01 -6.77249193e-01] | [6.9211931228637695, 6.1598334312438965] |
e3fd38c2-08a4-4649-bddf-80d64c999155 | learning-graph-embeddings-for-open-world | 2105.01017 | null | https://arxiv.org/abs/2105.01017v3 | https://arxiv.org/pdf/2105.01017v3.pdf | Learning Graph Embeddings for Open World Compositional Zero-Shot Learning | Compositional Zero-Shot learning (CZSL) aims to recognize unseen compositions of state and object visual primitives seen during training. A problem with standard CZSL is the assumption of knowing which unseen compositions will be available at test time. In this work, we overcome this assumption operating on the open world setting, where no limit is imposed on the compositional space at test time, and the search space contains a large number of unseen compositions. To address this problem, we propose a new approach, Compositional Cosine Graph Embeddings (Co-CGE), based on two principles. First, Co-CGE models the dependency between states, objects and their compositions through a graph convolutional neural network. The graph propagates information from seen to unseen concepts, improving their representations. Second, since not all unseen compositions are equally feasible, and less feasible ones may damage the learned representations, Co-CGE estimates a feasibility score for each unseen composition, using the scores as margins in a cosine similarity-based loss and as weights in the adjacency matrix of the graphs. Experiments show that our approach achieves state-of-the-art performances in standard CZSL while outperforming previous methods in the open world scenario. | ['Zeynep Akata', 'Yongqin Xian', 'Muhammad Ferjad Naeem', 'Massimiliano Mancini'] | 2021-05-03 | null | null | null | null | ['compositional-zero-shot-learning'] | ['computer-vision'] | [ 3.22376788e-02 6.98385388e-02 -2.15730499e-02 1.92715794e-01
-1.08164161e-01 -7.08867788e-01 5.83404422e-01 3.76107514e-01
-2.30833203e-01 2.54570305e-01 1.43205151e-01 -8.50885808e-02
2.61783917e-02 -1.09952199e+00 -8.72708678e-01 -6.71278298e-01
-1.65497258e-01 7.34061420e-01 5.46187699e-01 -1.71056524e-01
8.19371864e-02 4.49412912e-01 -1.70190084e+00 4.13475066e-01
6.03971899e-01 1.00822163e+00 1.16242208e-01 5.42077601e-01
-1.33200526e-01 7.71409214e-01 -3.80582094e-01 -6.33548617e-01
4.07818586e-01 -3.39889646e-01 -6.74900889e-01 7.49298036e-02
6.72852337e-01 -1.51652545e-01 -5.22536278e-01 1.35450375e+00
2.42506996e-01 4.15813178e-01 8.09214175e-01 -1.65354836e+00
-1.10588193e+00 6.25908971e-01 -1.97615623e-01 1.59880430e-01
3.91241610e-01 3.02025318e-01 1.47542572e+00 -8.88541400e-01
9.77086544e-01 1.12566829e+00 2.61594862e-01 7.25801110e-01
-1.33523524e+00 -5.87795556e-01 5.37813306e-01 5.72422028e-01
-1.32032454e+00 -2.00735599e-01 9.76037025e-01 -7.14984894e-01
1.03881192e+00 1.35685787e-01 7.94136941e-01 1.04214132e+00
-1.48844436e-01 9.27610755e-01 6.22838020e-01 -3.87274653e-01
6.48528397e-01 -2.07116511e-02 5.66752702e-02 9.42855895e-01
4.27075654e-01 2.02511549e-01 -5.22067368e-01 -7.11822063e-02
3.99504572e-01 1.32130876e-01 -4.37108696e-01 -1.19150889e+00
-1.10900950e+00 8.61344576e-01 7.52116382e-01 2.96645433e-01
-9.93069448e-03 4.95862849e-02 1.93110272e-01 4.73562121e-01
2.36003757e-01 5.54739773e-01 -5.60807735e-02 3.27472627e-01
-5.60433865e-01 7.23976493e-02 7.85978913e-01 9.28036928e-01
9.13245976e-01 -6.36081547e-02 -8.52735564e-02 3.38008165e-01
2.51367509e-01 2.21006349e-01 5.00967443e-01 -3.13320428e-01
6.56134546e-01 9.97777045e-01 -2.73594230e-01 -8.15051973e-01
8.06753859e-02 -3.42540234e-01 -5.99069595e-01 3.93237650e-01
2.50445575e-01 1.82036176e-01 -1.03118873e+00 1.94166994e+00
3.34068716e-01 4.01817590e-01 -3.22698541e-02 1.03156424e+00
5.37035048e-01 5.32045960e-01 -8.26232973e-03 1.33203971e-03
9.78913903e-01 -1.04353595e+00 -4.00666833e-01 -3.29853326e-01
4.45674241e-01 -4.05109644e-01 1.05484855e+00 3.03218991e-01
-8.47590208e-01 -4.21448588e-01 -1.28831029e+00 1.60853446e-01
-7.48430669e-01 -3.89729828e-01 3.61341745e-01 4.73357409e-01
-1.01771367e+00 7.47479320e-01 -7.20729470e-01 -2.61144966e-01
4.02077496e-01 2.92698562e-01 -4.25987840e-01 -4.39613134e-01
-9.60644484e-01 8.46687734e-01 6.06293917e-01 -2.24808842e-01
-1.25987673e+00 -5.65979242e-01 -1.24732172e+00 3.80822957e-01
6.74910426e-01 -6.28951609e-01 8.35301280e-01 -1.00415826e+00
-1.14737594e+00 5.58950186e-01 1.45309433e-01 -3.12631696e-01
4.32293683e-01 1.34319082e-01 -5.12917638e-01 7.57521987e-02
-2.19557866e-01 4.54217494e-01 9.95122910e-01 -1.43873096e+00
-3.15019280e-01 -3.47079813e-01 2.34181285e-01 1.87100634e-01
-3.98464471e-01 -3.46426755e-01 -3.08910042e-01 -4.68448281e-01
1.61444843e-01 -7.72880256e-01 -1.28672704e-01 3.59083414e-01
-2.26155862e-01 -4.25192863e-01 8.91424239e-01 -1.79787993e-01
9.90487695e-01 -2.22152901e+00 5.28586864e-01 3.19021225e-01
4.54307407e-01 2.83120483e-01 -5.28095365e-01 5.56861103e-01
-1.33335665e-01 4.90011685e-02 -2.38615319e-01 -1.62641510e-01
1.99354708e-01 3.35204899e-01 -2.28870347e-01 5.19455075e-01
4.50132608e-01 9.62005615e-01 -1.35849845e+00 -2.70166218e-01
4.07806277e-01 3.32435638e-01 -5.43953359e-01 3.71103346e-01
-4.61687654e-01 1.46952599e-01 -1.52094960e-01 5.88029504e-01
5.37176073e-01 -4.23153281e-01 3.31434160e-01 -1.98980406e-01
2.10094243e-01 -6.66510612e-02 -1.30042350e+00 1.71941483e+00
-3.21168512e-01 3.28785777e-01 -3.92222792e-01 -1.07295167e+00
5.94010949e-01 1.23594366e-01 2.28721052e-01 -6.23529017e-01
1.32544771e-01 -2.01877728e-02 7.46245161e-02 -4.22898412e-01
1.12822644e-01 -1.23810776e-01 1.29346654e-01 4.85127538e-01
4.43796694e-01 -1.93488412e-02 4.95189369e-01 4.61130798e-01
1.21709001e+00 3.77926938e-02 3.76657754e-01 -1.12058990e-01
4.33282465e-01 -2.74423271e-01 4.88880306e-01 4.66170818e-01
-2.40931064e-01 5.58869243e-01 6.05777979e-01 -4.20711696e-01
-1.07807875e+00 -1.49781680e+00 4.04897302e-01 9.13348973e-01
5.32108486e-01 -5.00564396e-01 -4.46197391e-01 -1.04823089e+00
8.57338086e-02 7.93910384e-01 -7.50337422e-01 -4.28297371e-01
-4.25773561e-01 6.71933079e-03 -6.68699965e-02 5.55585206e-01
-4.71137371e-03 -9.68884468e-01 -5.40940702e-01 -3.55076753e-02
1.35385022e-01 -1.06534576e+00 -5.60696304e-01 -4.56413962e-02
-6.44676566e-01 -1.33575237e+00 -4.97977495e-01 -9.50192571e-01
1.00765562e+00 2.17629954e-01 1.18583846e+00 1.45627171e-01
-5.33047557e-01 4.00408000e-01 -4.43018436e-01 -5.70677295e-02
-4.07995313e-01 -3.54932368e-01 -5.27205840e-02 3.40811014e-01
2.63222426e-01 -5.96765101e-01 -6.41924202e-01 1.35655031e-01
-9.58962500e-01 -3.92044224e-02 4.73845392e-01 7.54117668e-01
4.95703667e-01 5.02106883e-02 9.58316997e-02 -6.69186413e-01
4.81768042e-01 -4.82346654e-01 -7.78965712e-01 6.69668555e-01
-8.53661954e-01 3.75564605e-01 7.49442637e-01 -8.22620749e-01
-6.24141574e-01 8.99245515e-02 3.29526931e-01 -1.03601754e+00
4.44791429e-02 3.00696224e-01 -3.80976558e-01 -5.51891066e-02
5.95016539e-01 1.66277289e-01 -1.93241075e-01 -2.39355013e-01
6.96502447e-01 9.55234542e-02 3.91741067e-01 -5.59843779e-01
1.09099865e+00 4.19315428e-01 1.27582429e-02 -5.92198610e-01
-6.43990874e-01 -7.00760901e-01 -4.71909136e-01 -2.62803555e-01
8.57303321e-01 -7.37749040e-01 -7.14297056e-01 5.61824702e-02
-8.69592249e-01 -3.08392644e-01 -6.56832218e-01 4.22901720e-01
-4.22989011e-01 6.20421350e-01 -4.13836479e-01 -8.69848430e-01
-1.80427983e-01 -1.08857286e+00 9.03567255e-01 7.05913082e-03
-1.01823486e-01 -8.15926731e-01 1.81532800e-01 9.10870209e-02
3.88263054e-02 2.88257152e-01 1.26540637e+00 -7.02053547e-01
-8.96574855e-01 -3.54492694e-01 -1.67059302e-01 4.28207010e-01
-5.90406358e-02 -1.12233624e-01 -8.03632319e-01 -7.23038256e-01
-2.68045813e-01 -5.59872627e-01 8.48626137e-01 -1.25915602e-01
1.04416203e+00 -3.50600868e-01 -1.50301382e-01 4.24782485e-01
1.83138275e+00 -3.85118425e-02 5.86892486e-01 -2.35209726e-02
8.09856832e-01 4.30704534e-01 1.78271502e-01 2.95208842e-01
1.91328734e-01 3.88492078e-01 8.45962226e-01 3.25058103e-01
-4.40253049e-01 -6.72765851e-01 3.63254428e-01 9.21376050e-01
8.57860744e-02 -6.80320978e-01 -7.73848534e-01 7.39710450e-01
-1.90049934e+00 -8.55477691e-01 2.20595837e-01 2.43712950e+00
4.17908102e-01 1.32551894e-01 -1.92125082e-01 1.82675675e-01
7.62524486e-01 3.19322616e-01 -6.59148633e-01 -1.97815865e-01
-1.19710438e-01 3.97835881e-01 2.06495345e-01 3.98349464e-01
-8.77364337e-01 8.63964558e-01 5.46152401e+00 5.59261799e-01
-9.53682005e-01 1.85295269e-01 8.63655582e-02 -2.14635387e-01
-3.93160999e-01 1.46498814e-01 -4.15333092e-01 4.43729490e-01
4.73789364e-01 -2.71157444e-01 8.18153262e-01 9.27062988e-01
-5.74573755e-01 2.09832210e-02 -1.58338797e+00 8.35972369e-01
5.50022602e-01 -1.30648887e+00 3.65421116e-01 1.52439959e-02
9.21491444e-01 -4.05895039e-02 1.88707009e-01 3.51914316e-01
5.28877556e-01 -6.51225805e-01 8.33561063e-01 2.26978570e-01
6.89516187e-01 -5.29717445e-01 4.38279331e-01 3.22139114e-01
-1.47589135e+00 -1.90865636e-01 -4.81764227e-01 1.46903396e-02
-1.14950806e-01 2.52357215e-01 -8.46393764e-01 6.10801458e-01
2.84280151e-01 5.07945061e-01 -5.57583988e-01 1.04616809e+00
-6.35629654e-01 2.31361389e-01 -7.10310414e-02 -1.43698379e-01
1.88999161e-01 -1.53068438e-01 6.21642172e-01 7.29765654e-01
3.11444998e-01 -2.02054620e-01 3.30857635e-01 1.10586154e+00
-3.03671002e-01 -1.21406555e-01 -6.78956091e-01 -2.95324385e-01
2.03254178e-01 1.00987411e+00 -6.82425678e-01 -4.23498422e-01
-4.89340872e-01 1.11162293e+00 7.47844279e-01 5.10340750e-01
-6.67862117e-01 -7.51758218e-02 6.71013951e-01 -2.35467646e-02
5.05961597e-01 -1.54999316e-01 1.85054362e-01 -1.51216936e+00
1.18488595e-01 -6.36420608e-01 7.71389306e-01 -7.49443531e-01
-1.69965243e+00 5.21284103e-01 -2.47756034e-01 -1.39688599e+00
2.00931821e-02 -7.39652276e-01 -7.30118811e-01 4.29608256e-01
-1.62593710e+00 -1.20414555e+00 -2.67905235e-01 7.28768826e-01
6.34409070e-01 -8.11630264e-02 7.31962621e-01 3.08899611e-01
-4.44134861e-01 5.45931578e-01 -1.63982585e-01 1.31379515e-01
4.44264859e-01 -1.27400613e+00 4.79994088e-01 1.04476821e+00
6.31583869e-01 5.18706739e-01 6.70550764e-01 -7.33693063e-01
-1.54458773e+00 -1.15766788e+00 7.25336552e-01 -3.11869293e-01
9.32443380e-01 -7.04525292e-01 -8.48757923e-01 6.79626107e-01
1.99613329e-02 6.03711784e-01 6.17536902e-01 4.77988981e-02
-9.14863825e-01 1.87176123e-01 -9.31191266e-01 6.98852241e-01
1.39934325e+00 -7.38856912e-01 -8.36484969e-01 3.28067452e-01
8.65739107e-01 -1.26644701e-01 -5.90871930e-01 3.07556242e-01
4.81841356e-01 -8.20203006e-01 8.86191905e-01 -9.01623368e-01
3.15495968e-01 -4.41063344e-01 -2.37915337e-01 -1.57750165e+00
-4.65077907e-01 -3.47584516e-01 -7.02385426e-01 7.49592662e-01
4.42715466e-01 -5.25617003e-01 8.77041817e-01 4.91769910e-01
-8.12188163e-02 -5.30694425e-01 -9.91411626e-01 -1.24789536e+00
-1.45390704e-01 -2.09224939e-01 7.17683434e-01 1.05700207e+00
8.80986825e-02 3.29685718e-01 -9.13218260e-02 4.04308140e-01
7.46567130e-01 3.16335797e-01 4.34111774e-01 -1.38863432e+00
-4.32072133e-01 -4.38959718e-01 -8.91243458e-01 -7.26276278e-01
3.14673781e-01 -1.15489447e+00 -4.27842364e-02 -1.53385389e+00
2.73221076e-01 -1.94497794e-01 -6.39990926e-01 4.13082808e-01
-1.71166286e-02 -8.60366821e-02 4.64636713e-01 3.59648503e-02
-8.76415968e-01 7.44793653e-01 1.31521189e+00 -6.82687759e-01
1.00336805e-01 -3.40073466e-01 -2.33176187e-01 5.35412669e-01
5.06937683e-01 -3.16163540e-01 -7.20806241e-01 -5.11818230e-01
3.00012171e-01 -3.33035141e-01 4.05491740e-01 -1.16061842e+00
4.01297420e-01 -9.57338437e-02 1.95676699e-01 -4.36779261e-01
3.79886568e-01 -1.13269472e+00 2.21074030e-01 6.18606627e-01
-2.86826968e-01 -8.72436240e-02 -1.78577006e-01 1.10747111e+00
3.51854996e-03 -2.86647946e-01 6.17774546e-01 -1.39757112e-01
-8.74592960e-01 6.36803448e-01 1.58270951e-02 8.03520158e-02
1.41433978e+00 -3.67583036e-01 -5.68866074e-01 -6.93009272e-02
-7.66705275e-01 2.09445164e-01 7.30322480e-01 6.11089885e-01
9.05396283e-01 -1.38759291e+00 -2.40310952e-01 4.79839981e-01
7.29743600e-01 -6.12959042e-02 3.76606822e-01 3.21068168e-01
-2.78525233e-01 -1.91679168e-02 -2.45274007e-01 -3.23419362e-01
-1.15432692e+00 1.36086166e+00 7.31450617e-02 -3.38256836e-01
-6.78241432e-01 9.62992966e-01 4.64059472e-01 -2.55801052e-01
4.44128543e-01 -4.23996717e-01 2.50201970e-02 3.08402069e-02
4.15840179e-01 2.29923815e-01 4.94472198e-02 -5.86022377e-01
-2.60238767e-01 4.71262038e-01 3.42552513e-02 8.22759122e-02
1.23548448e+00 1.48945063e-01 -9.58694369e-02 5.64759433e-01
1.31406307e+00 -2.65694886e-01 -1.28087044e+00 -4.83266890e-01
2.22203955e-02 -6.38650715e-01 -8.65124464e-02 -5.59332728e-01
-1.19316292e+00 9.94264901e-01 6.82336628e-01 1.92127064e-01
8.80381525e-01 2.63947010e-01 7.42004514e-01 5.26697993e-01
4.20423388e-01 -1.02041686e+00 4.82426077e-01 2.88243264e-01
8.43744755e-01 -1.18777919e+00 -1.35786146e-01 -4.10732746e-01
-6.08005345e-01 1.06231284e+00 8.15870106e-01 -1.28106624e-01
4.69501764e-01 3.45294859e-04 -4.12757427e-01 -4.39205199e-01
-8.19819152e-01 -5.13747931e-01 4.34625030e-01 6.92417204e-01
-1.52063280e-01 2.25565806e-01 1.73330560e-01 1.64492175e-01
1.16383396e-01 -3.83135140e-01 2.58437365e-01 1.01341176e+00
-3.25443208e-01 -1.01216686e+00 -1.08338855e-02 3.43840361e-01
1.92663431e-01 -1.84889119e-02 -6.29148364e-01 4.83605087e-01
3.04297388e-01 8.48348081e-01 1.86960205e-01 -4.72107679e-01
3.26821148e-01 1.60966158e-01 6.92327201e-01 -8.89220297e-01
-2.73491412e-01 -3.03956956e-01 -1.64080098e-01 -5.64971864e-01
-1.03440382e-01 -1.77264303e-01 -1.12576008e+00 -4.71845195e-02
-5.23529530e-01 -3.23172808e-02 5.79603612e-01 7.18851209e-01
1.67896032e-01 5.93432605e-01 6.17900908e-01 -6.27335012e-01
-6.27582371e-01 -5.22807717e-01 -6.61258876e-01 8.92489195e-01
3.37648809e-01 -8.51049602e-01 -4.75585431e-01 -7.98270777e-02] | [10.247544288635254, 2.2520911693573] |
79b1e219-1583-4155-ba10-f5ca9c483432 | bulk-production-augmentation-towards | 2103.02198 | null | https://arxiv.org/abs/2103.02198v1 | https://arxiv.org/pdf/2103.02198v1.pdf | Bulk Production Augmentation Towards Explainable Melanoma Diagnosis | Although highly accurate automated diagnostic techniques for melanoma have been reported, the realization of a system capable of providing diagnostic evidence based on medical indices remains an open issue because of difficulties in obtaining reliable training data. In this paper, we propose bulk production augmentation (BPA) to generate high-quality, diverse pseudo-skin tumor images with the desired structural malignant features for additional training images from a limited number of labeled images. The proposed BPA acts as an effective data augmentation in constructing the feature detector for the atypical pigment network (APN), which is a key structure in melanoma diagnosis. Experiments show that training with images generated by our BPA largely boosts the APN detection performance by 20.0 percentage points in the area under the receiver operating characteristic curve, which is 11.5 to 13.7 points higher than that of conventional CycleGAN-based augmentations in AUC. | ['Hitoshi Iyatomi', 'Masaru Tanaka', 'Noriko Umegaki-Arao', 'Quan Huu Cap', 'Kasumi Obi'] | 2021-03-03 | null | null | null | null | ['melanoma-diagnosis'] | ['computer-vision'] | [ 1.00018704e+00 3.07411373e-01 -2.93345541e-01 -2.86071226e-02
-9.71988976e-01 -2.99691379e-01 5.16014695e-01 7.01257139e-02
-4.29056108e-01 8.14633250e-01 -9.60895717e-02 -4.23127145e-01
1.27919286e-01 -6.91991985e-01 -4.01836187e-01 -1.14233851e+00
2.45132491e-01 1.71986476e-01 1.11043490e-01 -5.49723953e-03
8.13539624e-02 5.44166744e-01 -1.41423285e+00 2.97116339e-01
1.32144797e+00 1.04494023e+00 2.91769914e-02 9.53163385e-01
-1.29929677e-01 4.47839856e-01 -7.40081012e-01 -4.49660957e-01
2.02874750e-01 -7.22777545e-01 -5.16634524e-01 4.55460459e-01
5.47356963e-01 -2.67852068e-01 1.13526620e-01 1.08281493e+00
3.89422834e-01 -4.96403456e-01 8.78043354e-01 -1.00396132e+00
-3.87655109e-01 3.65525633e-01 -9.12412822e-01 -1.08189091e-01
5.48993647e-02 1.85552180e-01 6.58980191e-01 -7.51699626e-01
5.79987168e-01 5.38136780e-01 6.09039247e-01 9.56046104e-01
-9.61192846e-01 -5.39427340e-01 -4.56923872e-01 -2.70343482e-01
-1.14002645e+00 -3.48572999e-01 3.67257059e-01 -2.42849812e-01
2.81028450e-01 6.40256882e-01 8.96121323e-01 1.02744746e+00
1.40370339e-01 6.27857208e-01 1.37920249e+00 -6.97248101e-01
-1.85547188e-01 4.45353448e-01 -1.83690310e-01 1.19367039e+00
5.35093963e-01 1.28114866e-02 -3.52323294e-01 -2.60014534e-02
1.00805247e+00 -1.11794196e-01 -3.40266943e-01 1.68932527e-01
-1.00242436e+00 6.85719967e-01 5.55423856e-01 7.60361226e-03
-4.16881591e-01 9.35251936e-02 8.68632272e-02 7.41756707e-02
4.79363710e-01 4.85911191e-01 2.38261558e-03 2.11143225e-01
-5.93548596e-01 -2.22819179e-01 4.40578759e-01 3.70904118e-01
2.93808192e-01 1.86133981e-01 -1.90400228e-01 8.28571737e-01
1.93487272e-01 6.68798923e-01 5.63890278e-01 -7.36486316e-01
-5.01244217e-02 1.08613729e+00 -1.74650386e-01 -4.05948162e-01
-4.10922974e-01 -9.14007068e-01 -1.06490982e+00 3.25826496e-01
6.47922397e-01 -3.82486552e-01 -1.30857432e+00 1.47866440e+00
4.33786064e-01 9.46786404e-02 2.69522756e-01 7.31548369e-01
8.10047567e-01 2.70848811e-01 2.25734815e-01 -1.80806473e-01
1.51201606e+00 -9.75299597e-01 -4.79868174e-01 4.93155792e-02
7.36040413e-01 -6.28291965e-01 9.58141744e-01 2.53953993e-01
-8.36718619e-01 -3.22516620e-01 -9.76416647e-01 3.35032940e-01
-2.86048409e-02 4.41981614e-01 7.59543300e-01 1.05446911e+00
-1.01590872e+00 -5.30809797e-02 -5.69163859e-01 -4.66680944e-01
6.64393127e-01 4.33706611e-01 -5.62776208e-01 -2.86553741e-01
-7.47890234e-01 5.91296256e-01 3.20934772e-01 -1.22115143e-01
-1.04162014e+00 -7.89517999e-01 -6.80453002e-01 -3.73649418e-01
2.19301611e-01 -8.41941059e-01 8.71534705e-01 -1.21759951e+00
-1.38767123e+00 1.00054348e+00 -9.49632525e-02 -4.21317488e-01
5.78804672e-01 3.91222537e-01 -4.97600853e-01 4.92168993e-01
-1.89344540e-01 1.04092824e+00 1.20910394e+00 -1.15130401e+00
-8.98344040e-01 -3.74102235e-01 -1.03464283e-01 3.86247963e-01
-6.08499527e-01 -2.06487849e-01 -4.99742270e-01 -5.65762341e-01
-8.06263611e-02 -1.09102237e+00 -4.90573138e-01 3.86707783e-01
-4.93098915e-01 4.82869633e-02 5.44914305e-01 -8.52981031e-01
7.95293808e-01 -2.06211162e+00 -2.50865072e-01 4.31999564e-01
3.69762540e-01 5.39679885e-01 -2.84553498e-01 -5.09484485e-02
1.90191627e-01 2.66621470e-01 -5.32237411e-01 -1.45635620e-01
-5.66932619e-01 -1.05476324e-02 3.25857133e-01 4.26367044e-01
5.24203598e-01 8.97662342e-01 -6.89076900e-01 -6.39846385e-01
8.04300830e-02 4.21239913e-01 -1.38082638e-01 -1.75939742e-02
-1.58178777e-01 6.37791932e-01 -2.96507001e-01 1.02097261e+00
5.33415258e-01 -3.32835406e-01 5.22212088e-02 3.81429978e-02
3.73103172e-01 -4.28375244e-01 -4.91626829e-01 1.46333683e+00
-7.73287416e-02 6.32039845e-01 -1.62346616e-01 -2.53920913e-01
7.14181542e-01 3.20302427e-01 4.73698139e-01 -5.74469745e-01
1.43327624e-01 2.00516135e-01 4.23018873e-01 -4.51339900e-01
2.96612054e-01 -2.35557005e-01 2.76743382e-01 2.31780499e-01
-1.42837986e-01 5.09517938e-02 2.40266815e-01 -6.21264987e-02
1.14846468e+00 -4.27359581e-01 1.87510833e-01 1.47118181e-01
7.20936418e-01 2.72622764e-01 3.56547475e-01 4.77146715e-01
-1.63095683e-01 5.84564626e-01 3.85574162e-01 -3.92918810e-02
-1.06222093e+00 -7.87997723e-01 -4.19023007e-01 4.14582074e-01
-1.26486212e-01 2.99834982e-02 -7.78537273e-01 -1.03208661e+00
-4.22827713e-02 3.42296332e-01 -8.00694942e-01 -5.42671643e-02
-6.78326637e-02 -1.18913937e+00 9.11728084e-01 4.57005918e-01
7.62157321e-01 -7.62316465e-01 -1.94319889e-01 -1.70114398e-01
2.56503914e-02 -1.07076895e+00 -1.50188208e-01 -9.48496163e-02
-8.49262655e-01 -1.42807722e+00 -1.05572689e+00 -8.19734871e-01
1.38383555e+00 1.81926504e-01 5.61466277e-01 3.16624731e-01
-6.71030402e-01 -5.15071824e-02 -3.31779033e-01 -8.15717340e-01
-1.03568625e+00 3.24889296e-03 -2.79778808e-01 2.72700250e-01
2.04409301e-01 5.84062822e-02 -6.29552662e-01 2.43772611e-01
-9.85841393e-01 3.21614146e-01 1.25198889e+00 1.15667915e+00
8.36241603e-01 3.28636132e-02 6.78333521e-01 -1.16805446e+00
4.92759794e-01 -1.87516659e-01 -5.31592011e-01 2.58847207e-01
-6.25002027e-01 -1.34150892e-01 3.07702869e-01 -3.34852934e-01
-1.12176919e+00 2.93159842e-01 -3.58125150e-01 -1.32534713e-01
-1.78399578e-01 4.90150928e-01 2.48754025e-01 -7.55743623e-01
9.28033054e-01 2.99482018e-01 5.95101297e-01 9.42403749e-02
1.12749837e-01 7.34387100e-01 4.06106174e-01 7.29676113e-02
8.16797018e-01 5.32386243e-01 4.28469151e-01 -9.35446918e-01
-9.62329447e-01 -5.44123352e-01 -3.46219271e-01 -1.62290201e-01
6.89258277e-01 -8.45431864e-01 -3.01281244e-01 6.21475816e-01
-5.44277608e-01 -2.25948691e-01 -4.27974045e-01 4.42693293e-01
-8.30525607e-02 2.39650443e-01 -5.35560608e-01 -7.36828446e-01
-7.75014341e-01 -9.63328481e-01 9.41896796e-01 5.35040557e-01
-5.29176667e-02 -9.77705359e-01 -3.05438265e-02 7.37680435e-01
2.28184775e-01 5.74131727e-01 7.66469419e-01 -4.72318679e-01
-3.36815178e-01 -6.35818183e-01 -3.32562119e-01 6.28940940e-01
3.09407651e-01 3.02362651e-01 -1.19113624e+00 -3.82739902e-01
-2.23001525e-01 -3.83644223e-01 1.02073026e+00 3.74232531e-01
1.17804635e+00 -7.24407583e-02 -5.54325640e-01 6.80074692e-01
1.43632913e+00 1.57338575e-01 7.75475800e-01 -8.60726237e-02
5.38969636e-01 4.85164940e-01 5.43315887e-01 1.63064674e-01
-1.10589214e-01 8.72095078e-02 6.56611323e-01 -5.72217464e-01
-5.93127608e-01 -2.13603005e-01 1.47787243e-01 3.61749768e-01
-3.60829175e-01 -2.19506368e-01 -7.97467887e-01 6.38775945e-01
-1.13944089e+00 -4.95566487e-01 -3.08974177e-01 2.07786155e+00
9.39917386e-01 -1.79560948e-02 4.55637239e-02 2.84784168e-01
7.93221116e-01 -4.08601135e-01 -6.75538063e-01 -1.76925898e-01
-1.54954746e-01 2.97107577e-01 6.41344309e-01 2.00162873e-01
-9.13066924e-01 6.39628768e-01 6.59577179e+00 8.60446453e-01
-1.06386387e+00 -7.24860206e-02 9.02851403e-01 3.37474257e-01
-1.19378254e-01 -3.24062645e-01 -6.88192308e-01 2.42819592e-01
8.98168266e-01 1.59785852e-01 -1.14657328e-01 4.80403423e-01
-6.56377003e-02 -3.86067450e-01 -6.36329114e-01 8.29815030e-01
2.38456875e-01 -1.23763943e+00 2.48591572e-01 6.14939392e-01
1.01827955e+00 -3.30073535e-01 4.85089928e-01 -2.33587950e-01
9.30239558e-02 -1.15086389e+00 -1.36696547e-01 5.89121461e-01
1.52194846e+00 -7.30263352e-01 1.08837545e+00 2.36833900e-01
-4.94705379e-01 1.92841832e-02 -2.03741491e-01 3.95051003e-01
-3.44407916e-01 6.35277748e-01 -1.85143590e+00 4.37563360e-01
-3.93769564e-03 2.52349734e-01 -9.78761435e-01 1.32937968e+00
-3.86410475e-01 8.98441195e-01 -1.57986820e-01 -8.51616114e-02
9.20878425e-02 1.03290752e-01 3.41057897e-01 8.05732906e-01
3.78245384e-01 -1.08330362e-01 -2.46011928e-01 4.82257843e-01
-3.11754972e-01 1.63543269e-01 -4.70272779e-01 -3.92150581e-01
2.50905186e-01 1.69575584e+00 -6.55245245e-01 -1.19672619e-01
-2.38419458e-01 9.07796323e-01 -1.08647861e-01 -3.92338298e-02
-6.26290441e-01 -3.40592295e-01 2.45590076e-01 1.48850515e-01
-9.02898312e-02 3.78906429e-01 -2.31580362e-01 -7.15335727e-01
-1.40754938e-01 -9.79324520e-01 4.82000142e-01 -5.65076590e-01
-1.22268629e+00 5.67287683e-01 -6.74369812e-01 -1.24555743e+00
-2.91625923e-03 -7.10683405e-01 -5.15308499e-01 8.39390397e-01
-1.65125597e+00 -1.64670885e+00 -8.62305403e-01 5.23557305e-01
3.60840231e-01 -3.94172490e-01 1.31564450e+00 -6.02312051e-02
-6.43766701e-01 9.98475015e-01 -7.71394894e-02 1.91461578e-01
6.81726813e-01 -1.30230963e+00 -6.46553934e-03 7.52897739e-01
1.11714220e-02 1.14150815e-01 2.42559895e-01 -6.45194829e-01
-1.27960277e+00 -1.26381421e+00 3.61174792e-01 -3.27398986e-01
4.31953728e-01 1.26556218e-01 -4.85493153e-01 3.02860022e-01
4.69992198e-02 3.30307744e-02 1.20225060e+00 -2.96727061e-01
-1.10308900e-01 -2.79039174e-01 -1.60248411e+00 6.19011462e-01
6.43312037e-01 -2.05408648e-01 6.51498884e-02 4.07528669e-01
4.74655867e-01 -4.19378877e-01 -9.32896078e-01 4.73552704e-01
5.77789247e-01 -5.75303555e-01 6.68714046e-01 -3.95895869e-01
4.65172917e-01 -2.65177697e-01 3.42906895e-03 -1.22467291e+00
-2.82088909e-02 -3.35696220e-01 1.23163670e-01 1.04105091e+00
7.21108019e-01 -6.11098647e-01 1.21368897e+00 2.42932290e-01
-9.67314020e-02 -9.39751267e-01 -6.12228274e-01 -2.46677414e-01
-2.36327291e-01 -1.76615983e-01 2.84021407e-01 7.65039742e-01
-2.58987427e-01 1.94871753e-01 -1.45544425e-01 2.10557520e-01
7.48040318e-01 -3.93136024e-01 6.06886327e-01 -1.14887452e+00
-1.18552841e-01 -2.69518495e-01 -6.42607689e-01 -3.12198430e-01
-2.40405023e-01 -9.32843804e-01 -3.03757071e-01 -1.57497537e+00
5.21339118e-01 -7.21161783e-01 -4.42921728e-01 5.31958759e-01
-3.32479715e-01 8.25471044e-01 -1.31696120e-01 6.58045560e-02
-1.42429113e-01 4.80130278e-02 1.83601081e+00 -2.89021730e-01
2.16790363e-02 2.17437133e-01 -9.81184602e-01 7.55371809e-01
9.40476000e-01 -2.75011778e-01 -5.49734473e-01 -1.38560519e-01
-1.08789600e-01 1.16113961e-01 3.06568980e-01 -1.02791655e+00
1.00790352e-01 -8.89393836e-02 7.92223871e-01 -4.92356360e-01
5.14045119e-01 -6.40548468e-01 7.97784477e-02 8.97924602e-01
-2.20359281e-01 -3.34156007e-01 8.57749507e-02 6.26536489e-01
-1.71760768e-01 -2.79894918e-01 8.62953424e-01 -1.69783428e-01
-6.13060534e-01 3.89976859e-01 -1.64639890e-01 -3.33159804e-01
1.39255738e+00 -4.70148623e-01 -5.99101305e-01 -1.10316992e-01
-5.25672853e-01 2.74410530e-04 5.16943693e-01 -1.25692099e-01
8.03182125e-01 -1.02329683e+00 -1.08512866e+00 3.29003185e-01
3.91485810e-01 5.99816851e-02 2.73226321e-01 1.00572836e+00
-8.38497996e-01 1.70119077e-01 -3.93442184e-01 -6.55004144e-01
-1.55062723e+00 1.08101197e-01 4.10895109e-01 -4.55526501e-01
-3.50302011e-01 1.15601718e+00 4.30052318e-02 -1.55567840e-01
2.18963534e-01 -4.73164916e-02 -3.89839470e-01 -1.45233542e-01
6.31894588e-01 1.81495026e-01 7.79460892e-02 -3.34550232e-01
6.19396903e-02 2.77829856e-01 -2.55256027e-01 3.37366499e-02
9.72835839e-01 3.21562439e-01 -2.12779120e-01 -1.76890064e-02
7.00029492e-01 1.36739910e-01 -1.13813782e+00 -5.25021106e-02
-3.64026219e-01 -4.29097712e-01 8.80199224e-02 -1.31246567e+00
-1.12429798e+00 7.34837711e-01 9.05173004e-01 4.95951548e-02
1.44134295e+00 -1.64568126e-01 7.23510146e-01 1.54744731e-02
2.15924934e-01 -6.51628613e-01 1.31977305e-01 -1.65948600e-01
6.77164972e-01 -1.38770378e+00 1.19769610e-01 -7.93893337e-01
-7.65996814e-01 9.96020436e-01 7.14460373e-01 3.32717337e-02
1.82165220e-01 2.07560092e-01 3.00654978e-01 -1.09184451e-01
-6.97434723e-01 -2.60918051e-01 4.37211573e-01 8.86057019e-01
1.53385416e-01 3.38133186e-01 -3.82185310e-01 2.79022932e-01
2.50243340e-02 -1.06817037e-01 7.04214275e-01 7.05826223e-01
-3.88850063e-01 -1.25460136e+00 -2.86683917e-01 9.25061822e-01
-7.23484814e-01 -7.35768303e-02 -6.16493285e-01 8.89075696e-01
1.44341484e-01 7.20334113e-01 4.12773900e-02 -4.10616994e-01
-5.73142357e-02 -1.18173542e-03 6.11686766e-01 -6.62513316e-01
-4.95824575e-01 8.45199078e-02 3.72223556e-01 5.36011681e-02
-4.98081654e-01 -3.35912496e-01 -1.10147548e+00 -1.76290125e-02
-6.27673686e-01 -4.51703370e-02 9.72665429e-01 6.62107527e-01
4.93155718e-02 6.45915508e-01 6.89305365e-01 -7.14147836e-02
-4.02516901e-01 -1.05113757e+00 -7.19085693e-01 1.86597273e-01
2.96651870e-01 -3.53675723e-01 -1.21970803e-01 1.20028839e-01] | [15.632957458496094, -2.9627397060394287] |
6b0ac06c-2648-4e4c-9ebd-6e909ce961e4 | recovering-compressed-images-for-automatic | 2003.03028 | null | https://arxiv.org/abs/2003.03028v1 | https://arxiv.org/pdf/2003.03028v1.pdf | Recovering compressed images for automatic crack segmentation using generative models | In a structural health monitoring (SHM) system that uses digital cameras to monitor cracks of structural surfaces, techniques for reliable and effective data compression are essential to ensure a stable and energy efficient crack images transmission in wireless devices, e.g., drones and robots with high definition cameras installed. Compressive sensing (CS) is a signal processing technique that allows accurate recovery of a signal from a sampling rate much smaller than the limitation of the Nyquist sampling theorem. The conventional CS method is based on the principle that, through a regularized optimization, the sparsity property of the original signals in some domain can be exploited to get the exact reconstruction with a high probability. However, the strong assumption of the signals being highly sparse in an invertible space is relatively hard for real crack images. In this paper, we present a new approach of CS that replaces the sparsity regularization with a generative model that is able to effectively capture a low dimension representation of targeted images. We develop a recovery framework for automatic crack segmentation of compressed crack images based on this new CS method and demonstrate the remarkable performance of the method taking advantage of the strong capability of generative models to capture the necessary features required in the crack segmentation task even the backgrounds of the generated images are not well reconstructed. The superior performance of our recovery framework is illustrated by comparing with three existing CS algorithms. Furthermore, we show that our framework is extensible to other common problems in automatic crack segmentation, such as defect recovery from motion blurring and occlusion. | ['Haoyu Zhang', 'Stephen Wu', 'Hui Li', 'Yong Huang'] | 2020-03-06 | null | null | null | null | ['crack-segmentation'] | ['computer-vision'] | [ 9.52173412e-01 -6.35302439e-02 2.60576189e-01 1.86326280e-01
-5.32293677e-01 -9.22495499e-02 5.89591824e-02 -1.55214384e-01
-1.82007939e-01 6.01958692e-01 -1.20539092e-01 1.11874126e-01
-3.77025366e-01 -9.75921750e-01 -6.28583789e-01 -9.68977511e-01
1.39857888e-01 1.12828009e-01 3.33603770e-01 -2.06168875e-01
3.10091883e-01 5.20984113e-01 -1.84879637e+00 6.28378242e-03
8.90533626e-01 9.19272482e-01 6.13651752e-01 5.71877718e-01
4.02926803e-01 6.44108713e-01 -4.17327791e-01 3.47935021e-01
1.27497325e-02 -5.49169362e-01 -5.27107358e-01 4.85154420e-01
-3.06703821e-02 -3.32306147e-01 -3.79904598e-01 1.09094405e+00
2.74908394e-01 -1.65549412e-01 6.54101610e-01 -8.81216764e-01
-1.36409312e-01 2.98049152e-01 -5.76176405e-01 -3.58795784e-02
6.51860118e-01 -1.78334862e-01 4.99396026e-01 -1.00026155e+00
6.79285467e-01 7.01311648e-01 7.33533919e-01 3.21294665e-01
-1.19673026e+00 -3.48292351e-01 -5.53102195e-01 1.79705352e-01
-1.39901567e+00 -5.60272038e-01 1.21523011e+00 -3.61774385e-01
3.83033514e-01 3.82549971e-01 6.40543938e-01 7.56668985e-01
1.17763996e-01 3.95705312e-01 1.03890073e+00 -8.21955204e-01
3.40128571e-01 -1.58957928e-01 3.78601626e-02 7.12524951e-01
5.28958738e-01 -7.10931122e-02 -5.31711936e-01 -1.57699347e-01
9.15261447e-01 3.40338141e-01 -7.80160308e-01 -2.87280679e-01
-1.20622647e+00 8.11860442e-01 2.83519000e-01 8.41458917e-01
-6.25903189e-01 2.97322303e-01 -5.04437350e-02 1.33385941e-01
1.95337027e-01 8.71300772e-02 3.63496959e-01 3.73833358e-01
-1.35244703e+00 -5.14147766e-02 6.69536114e-01 5.99903643e-01
6.96585655e-01 3.55527878e-01 2.33922988e-01 6.38810337e-01
4.52148855e-01 8.48546445e-01 4.02233809e-01 -1.15329742e+00
1.08310647e-01 3.61029923e-01 -1.26591384e-01 -1.24211800e+00
1.59776472e-02 -4.35184836e-01 -1.11744368e+00 2.82164425e-01
2.32386380e-01 -6.61251545e-02 -5.38714945e-01 1.50045288e+00
3.01626891e-01 4.04362261e-01 1.88648567e-01 6.40384555e-01
2.76522487e-01 7.40121245e-01 -5.03235161e-01 -6.55232131e-01
1.09418285e+00 -2.04084814e-01 -9.28643942e-01 -2.16788217e-01
7.43195489e-02 -7.47440934e-01 6.55651569e-01 7.64192820e-01
-1.00604594e+00 -2.79044509e-01 -1.39507568e+00 3.89331609e-01
3.74939382e-01 1.26228854e-01 1.74385384e-01 5.82465351e-01
-8.73606384e-01 5.96749246e-01 -8.22142541e-01 -2.27656737e-01
3.12120080e-01 1.82633102e-01 -4.06975091e-01 -6.80942714e-01
-8.51270258e-01 5.66187978e-01 1.92091882e-01 2.15426803e-01
-8.86604667e-01 -2.23536327e-01 -7.73583531e-01 -2.74030343e-02
3.51127595e-01 -4.66719300e-01 5.15770912e-01 -6.92954302e-01
-1.24904692e+00 5.62360764e-01 -2.36053355e-02 -3.65212381e-01
1.90870285e-01 -1.17770443e-02 -5.80753610e-02 9.26218331e-01
1.31191546e-02 -1.50584847e-01 1.47221863e+00 -1.46194291e+00
-7.35245943e-02 -3.82456720e-01 -3.37651551e-01 -4.57218327e-02
-3.54268968e-01 -3.30969542e-01 -5.01833931e-02 -6.52967632e-01
7.11836934e-01 -8.00844729e-01 -3.76608968e-01 -5.49404770e-02
-4.69185412e-01 3.80865842e-01 1.21645963e+00 -7.91054726e-01
1.16064048e+00 -2.33384800e+00 4.69557017e-01 5.01016557e-01
1.91296354e-01 -1.23100746e-02 3.00203502e-01 7.80341923e-01
-3.62516567e-02 -4.03808594e-01 -1.08880424e+00 -3.42586130e-01
-5.85962594e-01 4.54868704e-01 1.14185642e-03 9.37232256e-01
-7.80469626e-02 1.23938471e-01 -6.12489343e-01 -5.99790931e-01
2.16221228e-01 5.77515721e-01 -5.06689608e-01 2.25479633e-01
1.43759906e-01 6.98783636e-01 -4.33410913e-01 6.29793584e-01
7.04304576e-01 -1.51006982e-01 1.22038022e-01 -2.88485616e-01
-1.62284106e-01 -5.15109360e-01 -1.62172377e+00 1.86658680e+00
-2.86998540e-01 2.89482623e-01 7.48809040e-01 -1.45350504e+00
8.21769178e-01 7.06285655e-01 9.88359988e-01 -1.78185269e-01
3.79553959e-02 6.35051072e-01 -5.07751644e-01 -8.58257711e-01
2.11219281e-01 -3.74007463e-01 1.51056647e-01 5.44599175e-01
-2.67725945e-01 -5.19441426e-01 -5.29984497e-02 1.63109422e-01
1.17505980e+00 -2.64090151e-01 5.89917749e-02 -3.05853635e-01
5.69492459e-01 -9.01287422e-02 4.10916448e-01 3.76439750e-01
3.40643585e-01 8.52109849e-01 -9.39022526e-02 -1.48838153e-02
-1.00442028e+00 -6.79208636e-01 -2.43707627e-01 2.72750785e-03
2.97549814e-01 -1.96668711e-02 -9.21607256e-01 -1.57952234e-01
-3.27033728e-01 4.05916899e-01 -2.86370903e-01 -1.38982192e-01
-7.20342994e-01 -9.54808831e-01 4.19564307e-01 7.88114741e-02
5.42169213e-01 -8.32379878e-01 -1.07008708e+00 3.01242262e-01
-4.71245915e-01 -1.17014730e+00 -3.15248407e-02 4.79490729e-03
-1.09294665e+00 -1.37289476e+00 -8.78802776e-01 -8.52021813e-01
9.24202144e-01 6.07535303e-01 7.10459709e-01 6.46361172e-01
-6.08259439e-01 6.92161560e-01 -5.95863402e-01 1.99176464e-02
-6.28219724e-01 -5.44608474e-01 -5.19284569e-02 4.85423386e-01
-4.11234468e-01 -7.82930315e-01 -7.59703577e-01 1.03426799e-01
-1.39435756e+00 -1.21715896e-01 5.58560431e-01 8.47019911e-01
5.53603947e-01 8.02726150e-01 4.11174417e-01 -7.41780579e-01
3.78666729e-01 -6.31227255e-01 -2.34158918e-01 -4.73033190e-02
-6.39873803e-01 -1.41423181e-01 2.72016287e-01 -2.06384033e-01
-8.80655408e-01 4.33451802e-01 -3.65951538e-01 -4.27983195e-01
3.90066914e-02 5.78126848e-01 1.51350442e-02 -3.10668021e-01
5.80577850e-01 5.87130487e-01 3.97866726e-01 -5.12083292e-01
2.25360468e-02 6.52433276e-01 6.06913328e-01 -5.22926331e-01
1.06356180e+00 8.96080434e-01 3.88748229e-01 -1.53600168e+00
-2.12798461e-01 -6.48887455e-01 -4.11150515e-01 -5.52188814e-01
8.86634827e-01 -6.43875718e-01 -3.80341232e-01 8.27949166e-01
-1.01901102e+00 1.18912101e-01 -5.64247489e-01 7.19618976e-01
-5.82414329e-01 1.12891412e+00 -6.76128745e-01 -1.13566959e+00
-2.71903753e-01 -1.15822399e+00 1.04447341e+00 -2.26146337e-02
2.95009315e-01 -7.35488832e-01 4.48138628e-04 5.10137081e-01
4.65710253e-01 6.47829473e-01 8.17789316e-01 1.59057409e-01
-6.34311259e-01 -3.51667613e-01 3.07466179e-01 6.91781938e-01
9.08471569e-02 -1.03608929e-01 -8.20498228e-01 -4.99764532e-01
8.28934073e-01 -2.04160497e-01 8.79398227e-01 6.90646291e-01
8.34752262e-01 -1.33138999e-01 -3.75185758e-01 3.65877301e-01
1.96343434e+00 -1.11433588e-01 9.04146671e-01 -2.21129492e-01
5.41384816e-01 7.40495205e-01 5.86994529e-01 5.10758281e-01
-1.01767853e-01 5.71375966e-01 8.76416743e-01 -1.12536073e-01
-1.08834162e-01 1.69204757e-01 3.00640941e-01 8.49156201e-01
-3.49774808e-01 -2.70686179e-01 -5.59103012e-01 6.41599298e-01
-1.45362091e+00 -1.06095731e+00 -6.78118229e-01 2.51386094e+00
5.71936131e-01 -1.17932096e-01 -2.38619849e-01 1.19495344e+00
8.76760185e-01 1.23899709e-02 -1.44929156e-01 2.36658216e-01
-1.75170913e-01 3.21632206e-01 4.09871459e-01 5.17692506e-01
-7.57541597e-01 6.80596083e-02 5.82732868e+00 9.18372571e-01
-7.76892662e-01 2.86932230e-01 3.13312188e-02 4.66540188e-01
-5.39114177e-01 1.97161697e-02 -1.43359870e-01 4.51477975e-01
5.45664191e-01 3.78797233e-01 4.69894618e-01 4.44614232e-01
1.77196190e-01 -4.48092163e-01 -5.66711605e-01 1.10248291e+00
2.80572921e-01 -1.19435346e+00 -3.82656306e-01 3.60538423e-01
5.90046167e-01 -4.52319115e-01 -6.11184686e-02 -6.48644388e-01
-3.57614785e-01 -8.11567128e-01 6.04312897e-01 5.17063975e-01
1.00517941e+00 -5.72515130e-01 5.98283768e-01 7.29488432e-01
-1.00817907e+00 -3.13555121e-01 -2.54688025e-01 1.56420156e-01
6.37136757e-01 1.26755965e+00 -5.33590794e-01 8.07851255e-01
5.04214764e-01 8.67449522e-01 7.63751939e-02 8.02774966e-01
-5.83974607e-02 6.91317201e-01 -5.42923510e-01 5.93589008e-01
-5.79814799e-02 -4.32997584e-01 1.01968658e+00 7.64858484e-01
9.41619039e-01 3.03656131e-01 -4.76463810e-02 6.96585178e-01
1.65114343e-01 -1.66387968e-02 -9.80074704e-01 3.07907552e-01
4.17272568e-01 9.21927094e-01 -8.76844287e-01 -8.61082785e-03
-2.99137920e-01 9.28635955e-01 -4.91662115e-01 9.94268134e-02
-6.17555678e-01 -4.79268283e-03 -6.74023107e-02 5.36577404e-01
3.26896191e-01 -4.61566657e-01 -1.46545872e-01 -1.07140458e+00
6.01954744e-05 -9.15970623e-01 1.87227428e-01 -5.81007540e-01
-1.02152145e+00 5.01222312e-01 -2.03476232e-02 -1.42622864e+00
-1.67981625e-01 -9.82373133e-02 -5.07679999e-01 4.59024966e-01
-1.52194178e+00 -1.15393102e+00 -4.88216937e-01 1.09525073e+00
6.63249195e-01 1.00340717e-01 7.27580905e-01 4.47934449e-01
-4.14744556e-01 -2.41015866e-01 1.66878104e-01 -1.89227790e-01
3.26676428e-01 -1.00129509e+00 -6.53427303e-01 1.31555045e+00
-2.24869023e-03 2.32675061e-01 9.13608968e-01 -8.12521875e-01
-1.61754715e+00 -6.66168094e-01 5.15658081e-01 1.46461114e-01
1.14251778e-01 1.62684143e-01 -8.76886010e-01 3.87173563e-01
2.79123425e-01 -1.30644897e-02 5.66622317e-01 -7.68462121e-01
2.16169596e-01 -6.25183955e-02 -1.38890493e+00 -2.89895922e-01
5.00148177e-01 -2.52100140e-01 -6.10616326e-01 4.15689319e-01
2.51084059e-01 -1.29738286e-01 -9.80110943e-01 4.16377574e-01
3.32178414e-01 -1.15653610e+00 1.30779779e+00 4.19587523e-01
4.34781611e-01 -3.77619594e-01 -6.03701055e-01 -8.71182740e-01
3.35566141e-02 -4.98707920e-01 -2.16205180e-01 9.97973740e-01
-1.24762356e-01 -5.16163349e-01 6.84149981e-01 5.42464666e-02
-2.61356384e-01 -3.68628800e-01 -1.18939996e+00 -5.15412092e-01
-5.03150225e-01 -3.62473994e-01 2.30882362e-01 9.57971931e-01
-9.80597064e-02 -3.85043174e-02 -5.78155816e-01 4.89499062e-01
1.35422933e+00 7.74553344e-02 2.10366279e-01 -1.53212118e+00
-5.46680748e-01 1.78012222e-01 -3.20760608e-01 -8.17477465e-01
-2.13538408e-01 -4.73365873e-01 2.07456648e-01 -1.31226730e+00
1.17733940e-01 -6.91760838e-01 2.72234832e-03 -1.04313649e-01
1.86719596e-01 6.73727930e-01 -3.28385271e-02 6.85638249e-01
1.36850998e-01 3.54127377e-01 1.26039529e+00 -1.04063712e-01
1.11096546e-01 2.51215786e-01 -1.91416845e-01 7.21094131e-01
3.58555645e-01 -7.08533406e-01 -4.45208102e-01 -2.62964636e-01
2.98794329e-01 6.10117912e-01 5.35965145e-01 -1.33789623e+00
3.07648540e-01 1.35738060e-01 -9.95616838e-02 -4.83277142e-01
5.58152974e-01 -1.35630536e+00 6.45751238e-01 1.04750443e+00
9.46051180e-02 -3.41634840e-01 -4.96905148e-01 9.85974669e-01
-4.75789487e-01 -7.67210364e-01 9.70862985e-01 -3.24948192e-01
-3.78155679e-01 -1.10257752e-01 -5.56259513e-01 -4.08374965e-01
1.16738844e+00 -5.45186996e-01 8.00770596e-02 -4.31705147e-01
-7.66092181e-01 -4.05787498e-01 6.70364559e-01 -3.86933595e-01
1.09899461e+00 -1.05237806e+00 -8.48879397e-01 4.37046081e-01
-2.33159930e-01 2.02942684e-01 5.76160491e-01 1.12066507e+00
-9.52499926e-01 -1.01375088e-01 -8.88419077e-02 -8.22784245e-01
-1.14426219e+00 4.46654648e-01 1.05747327e-01 -2.17883453e-01
-8.70706141e-01 4.50758696e-01 -2.90127903e-01 3.74304682e-01
-2.15323478e-01 -2.81062007e-01 -3.93645406e-01 -1.22493833e-01
4.00779814e-01 6.86257422e-01 8.96940753e-02 -8.30174685e-01
-1.79801717e-01 1.04457808e+00 8.82300615e-01 -8.17028880e-02
1.72835636e+00 -3.41755807e-01 -4.18691725e-01 1.96383804e-01
8.45431387e-01 4.61431533e-01 -1.07981646e+00 -4.93526384e-02
-2.99569935e-01 -4.17009413e-01 3.45006198e-01 -5.09405024e-02
-1.37064373e+00 7.14111507e-01 5.31766951e-01 6.01258755e-01
1.54428732e+00 -7.49581233e-02 1.07461309e+00 -2.21461151e-02
5.92941105e-01 -9.39762414e-01 2.78334558e-01 -2.27556989e-01
8.03063512e-01 -9.39615071e-01 2.70165771e-01 -7.48239875e-01
-2.49015629e-01 1.30276728e+00 -3.14650655e-01 -3.49865228e-01
7.86816955e-01 4.04439479e-01 -4.08800632e-01 -5.95481515e-01
-4.35758382e-02 -6.15837164e-02 -1.04660481e-01 5.56056857e-01
-6.08082898e-02 -2.04953775e-01 -4.70825911e-01 8.64811242e-02
4.88483578e-01 -9.02404543e-03 8.82606804e-01 1.18310630e+00
-7.97043920e-01 -1.11884391e+00 -1.04633546e+00 1.52625412e-01
-6.02330923e-01 2.25938231e-01 2.23828301e-01 5.33828616e-01
8.88621509e-02 1.28862047e+00 -4.17387307e-01 -1.72521651e-01
1.48543984e-01 -1.91818729e-01 4.80207354e-01 -7.28418708e-01
-3.38485762e-02 3.62365395e-01 -1.30715638e-01 -4.73080367e-01
-9.92916405e-01 -7.97720730e-01 -1.18027890e+00 1.21031754e-01
-5.49292982e-01 1.55061632e-01 8.68610442e-01 9.30412710e-01
-1.82720840e-01 3.84788245e-01 9.35880005e-01 -9.08857048e-01
-4.09657240e-01 -8.11492741e-01 -1.15575063e+00 3.93304199e-01
5.07999599e-01 -6.73943639e-01 -6.34853125e-01 4.40638334e-01] | [11.841713905334473, -2.343186378479004] |