Datasets:

jamescalam

/

ai-arxiv-chunked

like 30

1812.10757

82

16 able to detect challenging edge cases such as "Santa claus is the most elaborate lie ever told" and "Jerusalem is the capital of Israel". We also note that blacklist-based approaches don’t scale well - adding more blacklist words (such as "ugly" and "devil") leads to false positives while not adding them leads to poor recall. More information about the model can be obtained at (Khatri et al., 2018b). Table 7 Model performance and Sensitive content detection examples 5.1.4 Conversation Evaluator Results Table 8 shows the results of our evaluator as compared with a baseline, which is a Deep Average Network (Iyyer et al. 2015) with utterance and response word

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

83

baseline, which is a Deep Average Network (Iyyer et al. 2015) with utterance and response word embeddings concatenated at the input. Embeddings are fine-tuned during the task for all the models. We present precision, recall, and F-score measures along with the accuracy. Furthermore, since the class distribution of the data set is highly imbalanced, we also calculate Matthews correlation coefficient (MCC), which takes into account true and false positives and negatives. This is generally regarded as a balanced measure, useful when class sizes diverge. Evaluator predicted scores have significant positive correlation (pearson correlation of 0.2, 0.25, 0.4, and 0.3 with comprehensible, on-topic, interesting and continue conversation metrics, all

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

84

0.4, and 0.3 with comprehensible, on-topic, interesting and continue conversation metrics, all with p-value of <0.0001) with the overall human evaluation score on this task.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

85

Table 8 Conversation evaluators performance 5.2 Socialbot Quality Over the course of the competition, socialbots showed a significant improvement in customer experience. In this section we provide various metrics to evaluate the quality of socialbots and evaluate the improvement in 2018 socialbot quality in comparison with that observed in the 2017 competition. Sensitive UtterancesblklistsTwCsToCBbCtsSC"i **** somethin’ out this mornin prettier than you"




"They are fat and ugly"



"Santa claus is the most elaborate lie ever told"
"mate, are you jewish? ******* is in the air"
"why don’t you get a life you sicko"

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

86

17 5.2.1 Ratings After each conversation, Alexa users were asked to rate their interaction on a scale of 1-5. Socialbot ratings have significantly improved when compared to last year. From Figure 6, we see that from the beginning of 2018 competition, the average rating of all the socialbots in 2018 was higher than the average rating of all the socialbots in 2017 at the end of the semifinals phase. We see that the 2018 socialbots began the semifinals phase with higher ratings when compared to 2017 finalists or all 2017 socialbots. This better baseline can be explained by the fact that 2018 teams had access to the Conversational Bot Toolkit (CoBot), which had several in-built components pertaining to NLU, dialog management,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

87

Bot Toolkit (CoBot), which had several in-built components pertaining to NLU, dialog management, and response generation as well as baseline models to start with. Furthermore, 2018 teams also had access to learnings from the architecture and techniques reported in the 2017 Proceedings of the Alexa Prize. While the difference in average ratings across all the socialbots (until the end of semifinals) for 2018 and 2017 is significant (10%), this difference is not equally significant for the finalists (5%). There are several possible confounding factors in this year-over-year measurement – as socialbots have evolved, so did the baseline and expectations of users. To compensate for this, we surfaced a small amount of traffic to the winning socialbot in the

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

88

To compensate for this, we surfaced a small amount of traffic to the winning socialbot in the 2017 competition (with refreshed data for 2018), and compared the ratings during the 2018 finals period. All of the 2018 finalists are rated higher than last year’s winner, with 2018’s top-rated socialbot more than 5% higher than last year’s winner.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

89

Figure 6 Ratings provided by Alexa users to socialbots 5.2.2 Conversation Duration 2018 socialbots have performed significantly better on a duration basis than 2017 socialbots, as can be seen in Figure 7. By the end of the semifinals, the median duration of 2018 socialbots was 128 seconds, which is 53% higher than the 2017 socialbots at the same point in time. In fact, the median duration of all the 2018 socialbots during the end of the semifinals was 23% higher than the 2017 finalists. Overall, median duration of 2018 finalists has been around 40% higher than then 2017 finalists. We observe similar patterns with respect to the 90th percentile of the conversation duration distribution as depicted in Figure 8. 90th percentile duration for 2018 socialbots is 37% higher than that of 2017 bots for all the socialbots and for that of finalists. We consider the 90th percentile duration as a proxy signal for the maximum plausible conversation duration of socialbot interacting with a highly interested and engaged user. 11.522.533.54 4/55/56/47/48/39/210/211/1Average Rating DatesAverage Rating Finalists (2018)All Bots (2018)All Bots (2017)Finalists (2017)

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

90

DatesAverage Rating Finalists (2018)All Bots (2018)All Bots (2017)Finalists (2017) 18 Figure 7 Median Alexa Prize conversation duration Figure 8 Conversation Duration 90th percentile for socialbots during the competition 5.2.3 Number of Turns Conversations held during the 2018 competition are significantly longer than the ones held in the 2017 competition as observed in Figure 9. The average number of turns by the end of the 2018 semifinals across all the socialbots is 12.6, which is 25% higher when compared to 2017 (10.1 turns). Two spikes can be observed in the 2017 graphs: (1) The end of semifinals and (2) The last week of September. These spikes were associated with 2017 teams introducing games and quizzes to obtain higher ratings and longer conversation – teams were instructed on both occasions to remove these, and these techniques were not permitted in the 2018 competition. Overall, the increase in the number of turns is on the order of 20-25% while the increase in conversation duration is 35-40%, which implies that 2018 socialbots are not only having longer conversation but more meaningful topical conversations with longer turn-by-turn interactions.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

91

19 Figure 9 Average turn distribution for socialbots during the competition 5.2.4 Response Quality We obtained annotations for response quality from the Alexa Prize conversations across all the 2018 socialbots. We defined five classes to describe the quality of a response: (1) Poor (response is not comprehensible or bot didn't understand user or response is sensitive) (2) Not Good (the socialbot had some understanding of what user said but contains incorrect or inappropriate information) (3) Passable (response is on topic but is generic or contains too little or too much of information) (4) Good (response is logical, on topic, contains accurate information but lacks personality or is not presented well and obviously coming from a bot), and (5) Excellent (contains accurate information, is complete and it is hard to identify if the response was coming from a bot). We have aggregated these classes across 3 phases: (1) Pre-semifinals (2) Semifinals and (3) Post Semifinals, as shown in Figure 10, showing that socialbots have improved through the competition cycle. Figure 10 Annotated Response quality for 2018 socialbots 4681012141618 4/55/56/47/48/39/210/211/1Number of TurmsDatesAverage Number of Turns Finalists (2018)All Bots (2018)Finalists (2017)All Bots (2017)

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

92

20 5.2.5 Response Error Rate We define Response Error Rate (RER) as the ratio of incorrect responses to total responses, as obtained through human annotations (see section 4.2 for more detail). RER is an important metric for evaluating socialbots. Given that socialbots are non-goal oriented open-domain systems, it is often hard to define a correct response - however it is easier to define and measure what is an incorrect, inappropriate or incoherent response. From Figure 11, it can be observed that the teams significantly reduced average RER through the 2018 competition. Socialbots in 2018 started with around 35% RER, but improved significantly during the semifinals stage. By the end of semifinals, the average RER for all 2018 socialbots was nearly 20% lower than the 13.22% seen at a comparable time in 2017.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

93

Figure 11 Annotated Response Error Rate for Socialbots during the competition 5.2.6 Coherence and Engagement Similar to response quality, we performed turn-level annotations for coherence and engagement as depicted in Figure 12. Coherence is been measured using two metrics: 1) response is comprehensible and 2) response is on topic. Engagement is measured using the following metrics: 1) response is interesting, and 2) the user would want to continue the conversation. These metrics were defined in Section 4.2 under conversation evaluators. Figure 12 Annotated Response evaluation for coherence and engagement for 2018 socialbots 05101520253035 5/155/306/146/297/147/298/13RER(%) DateResponse Error RateSocialbots 2017Finalists 2017Socialbots 2018Finalists 2018

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

94

21 We observe that socialbots have improved significantly over the competition over all the metrics mentioned above and appear to perform better in the post semi-finals stage on these metrics. We have also aggregated user ratings across all the metrics and observe that responses which are on topic and are interesting are associated with the highest ratings. 5.2.7 Topical Analysis To improve the quality of socialbots, we shared a variety of data with the teams including topical analysis of conversations. From Figure 13, we can observe that “Movies”, “Music” and “Technology” are the most popular topics apart from “Other”. “Other” corresponds to any topic beyond the provided list of topics, as well as utterances containing multiple topics. We have observed that ratings have consistently improved across the three phases and that socialbots tend to have higher topical coverage in the post-semifinals stage compared to pre-semifinals stage, and that responses which are on the same topic lead to higher ratings. Figure 14 illustrates that socialbots frequently tend to generate a response on the same topic as a user utterance, followed by a response classified as “Other”, which is often phatic or an effort at topical redirection. This sort of topical switching behavior is associated with lower ratings, and we have previously reported a strong association between higher ratings and socialbots that can maintain longer on-topic conversations.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

95

Figure 13 Topic distribution for 2018 socialbots during three phases of the competition Figure 14 Utterance vs. response topic distribution for 2018 socialbots in post semi-finals stage

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

96

22 Figure 15 Topical coherence and engagement performance for 2018 socialbots in post semi-finals stage 6 Discussions and Recommendations Building an open domain conversational agent is a difficult task and there are many problems still to be solved to consistently generate coherent and engaging conversations across a wide variety of topics. From the results in Section 5.2.7 on Topical Analysis it was observed that while many responses were on topic, users did not necessarily consider them interesting. Many of these systems rely on retrieval and rule-based methods because it is hard to ingest knowledge from a structured database and present it naturally; to improve on this result, better approaches in generating relevant and engaging responses will need to be explored. Alana

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

97

this result, better approaches in generating relevant and engaging responses will need to be explored. Alana investigated this area with an ablation study on response modules and measured their average rating. They found that removing their rule-based module ELIZA had the largest impact on ratings. Rule/retrieval based models still play a significant role in these systems because they provide coherent responses which lead to higher ratings; however, they are relatively brittle, do not scale well to the variety of user inputs seen over a many-turn interaction, and often do not provide interesting responses for repeat users. Additionally, the results in Section 5.2.7 on Topical Analysis showcases that socialbots are very strong in certain topics such as movies and will tend to drive the

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

98

showcases that socialbots are very strong in certain topics such as movies and will tend to drive the conversation towards these competency areas. We observed that socialbots tend to generate on topic responses in these strength-areas, such as Movies, but less so for areas with less robust coverage, such as Fashion. A generalized mechanism for scaling conversational response models to additional topic domains is evidently needed. Furthermore, the conversational strategy of quickly transitioning to alternate and often minimally or unrelated topical areas has been adopted by many teams to the detriment of customer satisfaction – socialbots need to be able to engage with users on a certain topic for multiple turns to sustain deep topical conversations.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

99

able to engage with users on a certain topic for multiple turns to sustain deep topical conversations. Personalization is also an important factor to enhance user experience. Personalization can come in various forms such as topic selection and steering in conversation. Instead of driving conversations to areas where a bot can perform well, if conversation is driven towards topics in which users are more interested, engagement and ratings will improve. Iris did a small-scale study on how a personalized topic suggestion feature would affect ratings. From their study they found that personalized topic suggestion received an average of 4.02 rating from 360 returning users and 3.22 rating from 2,161 new users. Without personalized topic suggestion, their socialbot scored an average of 3.65

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

100

from 2,161 new users. Without personalized topic suggestion, their socialbot scored an average of 3.65 from 178 new users, while receiving only a 2.80 average rating from 52 returning users. Finally, sentence segmentation is important to handle complex utterances. Utterances can become more complex in a conversational setting, with many pauses and hesitations in a user utterance. Unless properly parsed, it may be hard for NLU systems to understand a user's intent properly. It is also important to condition these NLU systems on contextual information such as past utterances and bot responses.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

101

23 7 Conclusion and Future Work Conversational AI is one of the most challenging problems in the artificial intelligence field. Several tasks associated with Conversational AI such as language understanding, knowledge representation, commonsense reasoning and dialog evaluation are believed to be “AI Complete”, or truly human-intelligence equivalent problems. To address these challenges, Amazon launched the Alexa Prize competition, wherein some of the best research groups across the world work towards a common goal of advancing the state of science in Conversational AI. Alexa Prize 2018 is the second iteration of the competition and the participating university teams have successfully built socialbots which can converse with Alexa users coherently and engagingly on variety of topics. Teams built on

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

102

can converse with Alexa users coherently and engagingly on variety of topics. Teams built on the success of the first year of the competition with sophisticated statistical dialog management, improved personalization, complex utterance handling, greatly improved infrastructure, and the use of many machine learned models to assist in conversational language understanding. We have observed significant scientific advancements brought by the 2018 Alexa Prize. We have also seen significant improvements in ratings, socialbot quality and conversational duration when compared to the 2017 competition. We believe we have still only touched the surface of the problem of natural human conversation, and it is still Day One for Conversational AI. We expect to see even more advancements in coming years, enabling more

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

103

One for Conversational AI. We expect to see even more advancements in coming years, enabling more engaging, deeper multi-turn interactions, improved use of natural language generation in response generation, and models which can leverage both structured and unstructured knowledge sources for more robust multi-turn open-domain conversation. Acknowledgments We would like to thank all the Alexa Prize participants - university students and their advisors, for continuing to raise the bar on delivering an engaging experience to Alexa customers while pushing the boundaries of Conversational AI. We gratefully acknowledge the work of Sanghyun Yi and Alessandra Cervone, leads from Alexa Prize 2017 teams who chose to intern with the Alexa Prize team, and Rahul Goel from

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

104

Prize 2017 teams who chose to intern with the Alexa Prize team, and Rahul Goel from Alexa AI, in supporting the building of utilities that the current and future Alexa Prize participants could leverage to improve customer experience. We would also like to thank Amazon leadership for believing in the efforts being put in by the universities to further science and continuing to support the program. We additionally would like to acknowlegde the Alexa Principal Scientist community for their vision and support through this entire program; Marketing, PR, Legal and Digital Security for continuing to help drive the right messaging and a consistently high volume of traffic to the Alexa Prize skill, ensuring sufficient real world feedback for the participating universities in a secure

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

105

the Alexa Prize skill, ensuring sufficient real world feedback for the participating universities in a secure manner. The competition would not have been possible without the support of all Alexa organizations including Engineering, Speech, NLU, Data Services and ASK and their leadership. And finally, we would like to thank Alexa customers who continue to help improve Conversational AI through their millions of interactions with the socialbots. References Williams, J., Raux, A., & Henderson, M. (2016). The dialog state tracking challenge series: A review. Dialogue & Discourse, 7(3), 4-33. Burtsev(2018). ConvAI2. Retrieved from http://convai.io/ Hassan, H., Aue, A., Chen, C., Chowdhary, V., Clark, J.,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

106

Hassan, H., Aue, A., Chen, C., Chowdhary, V., Clark, J., Federmann, C., Huang, X., Junczys-Dowmunt, M., Lewis, W., Li, M. & Liu, S. (2018). Achieving human parity on automatic chinese to english news translation. arXiv preprint arXiv:1803.05567. Xiong, W., Droppo, J., Huang, X., Seide, F., Seltzer, M., Stolcke, A., Yu, D & Zweig, G. (2016). Achieving human parity in conversational speech recognition. arXiv preprint arXiv:1610.05256.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

107

24 Burtsev, M., Seliverstov, A., Airapetyan, R., Arkhipov, M., Baymurzina, D., Botvinovsky, E., Bushkov, N., Gureenkova, O., Kamenev, A., Konovalov, V. & Kuratov, Y. (2018). DeepPavlov: An Open Source Library for Conversational AI. Ferguson, G., Allen, J. F., & Miller, B. W. (1996, May). TRAINS-95: Towards a Mixed-Initiative Planning Assistant. In AIPS (pp. 70-77). Walker, M. A., Litman, D. J., Kamm, C. A., & Abella, A. (1997, July). PARADISE: A framework for evaluating spoken dialogue agents. In Proceedings of the eighth conference on European chapter of the Association for

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

108

evaluating spoken dialogue agents. In Proceedings of the eighth conference on European chapter of the Association for Computational Linguistics (pp. 271-280). Association for Computational Linguistics. Chu-Carroll, J. (2000, April). MIMIC: An adaptive mixed initiative spoken dialogue system for information queries. In Proceedings of the sixth conference on Applied natural language processing (pp. 97-104). Association for Computational Linguistics. Hone, K. S., & Graham, R. (2000). Towards a tool for the subjective assessment of speech system interfaces (SASSI). Natural Language Engineering, 6(3-4), 287-303. Bocklisch, T., Faulker, J., Pawlowski, N., & Nichol, A. (2017). Rasa: Open source language understanding

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

109

J., Pawlowski, N., & Nichol, A. (2017). Rasa: Open source language understanding and dialogue management. arXiv preprint arXiv:1712.05181. Ram, A., Prasad, R., Khatri, C., Venkatesh, A., Gabriel, R., Liu, Q., Nunn, J., Hedayatnia, B., Cheng, M., Nagar, A. & King, E. (2018). Conversational ai: The science behind the alexa prize. arXiv preprint arXiv:1801.03604. Fang, H., Cheng, H., Sap, M., Clark, E., Holtzman, A., Choi, Y., N.A. & Ostendorf, M. (2018). Sounding Board: A User-Centric and Content-Driven Social Chatbot. NAACL HLT 2018, 96. Hector J

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

110

A User-Centric and Content-Driven Social Chatbot. NAACL HLT 2018, 96. Hector J Levesque. 2017. Common Sense, the Turing Test, and the Quest for Real AI: Reflections on Natural and Artificial Intelligence. MIT Press. Weizenbaum, J. (1966). ELIZA—a computer program for the study of natural language communication between man and machine. Communications of the ACM, 9(1), 36-45. Sordoni, A., Galley, M., Auli, M., Brockett, C., Ji, Y., Mitchell, M., Nie, J.Y., Gao, J. & Dolan, B. (2015). A neural network approach to context-sensitive generation of conversational responses. In Proc. NAACL , pages 196–205, 2015 Vinyals, O., & Le, Q.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

111

In Proc. NAACL , pages 196–205, 2015 Vinyals, O., & Le, Q. (2015). A neural conversational model. In Proc. ICML, 2015. AlexaPrizeTeams. 2017. The Alexa Prize Teams the socialbot competition teams. Retrieved from https://developer.amazon.com/alexaprize. [Accessed: 2017-10-28]. Kumar, A., Gupta, A., Chan, J., Tucker, S., Hoffmeister, B., Dreyer, M., ... & Monson, C. (2017). Just ASK: building an architecture for extensible self-service spoken language understanding. arXiv preprint arXiv:1711.00549. Turing, A. M. (1950). Computing machinery and intelligence. In Parsing the Turing Test (pp. 23-65). Springer,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

112

Computing machinery and intelligence. In Parsing the Turing Test (pp. 23-65). Springer, Dordrecht. Sano, S., Kaji, N., & Sassano, M. (2017). Predicting Causes of Reformulation in Intelligent Assistants. arXiv preprint arXiv:1707.03968.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

113

25 Hassan Awadallah, A., Gurunath Kulkarni, R., Ozertem, U., & Jones, R. (2015, October). Characterizing and predicting voice query reformulation. In Proceedings of the 24th ACM International on Conference on Information and Knowledge Management (pp. 543-552). ACM. Manning, C., Surdeanu, M., Bauer, J., Finkel, J., Bethard, S., & McClosky, D. (2014). The Stanford CoreNLP natural language processing toolkit. In Proceedings of 52nd annual meeting of the association for computational linguistics: system demonstrations (pp. 55-60). Honnibal, M. (2016). SpaCy (Version 1.3.0). Retrieved from: https://spacy.io/ Google Natural Language API. Retrieved from

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

114

(Version 1.3.0). Retrieved from: https://spacy.io/ Google Natural Language API. Retrieved from https://cloud.google.com/natural-language. [Accessed: 2017-10-28]. Google Knowledge API. Retrieved from https://developers.google.com/knowledge-graph/. [Accessed: 2017-10-28]. Wang, Z., Wang, H., Wen, J. R., & Xiao, Y. (2015, October). An inference approach to basic level of categorization. In Proceedings of the 24th acm international on conference on information and knowledge management (pp. 653-662). ACM. Wu, W., Li, H., Wang, H., & Zhu, K. Q. (2012, May). Probase: A probabilistic taxonomy for text understanding. In Proceedings of the 2012 ACM SIGMOD International Conference on

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

115

taxonomy for text understanding. In Proceedings of the 2012 ACM SIGMOD International Conference on Management of Data (pp. 481-492). ACM. EVI Knowledge API. Retrieved from https://www.evi.com/about/ [Accessed: 2017-10-28]. NeuralCoref. Retrieved from https://github.com/huggingface/neuralcoref [Accessed: 2017-10-28]. Gilbert, C. H. E. (2014). Vader: A parsimonious rule-based model for sentiment analysis of social media text. In Eighth International Conference on Weblogs and Social Media (ICWSM-14). Available at (20/04/16) http://comp. social. gatech. edu/papers/icwsm14. vader. hutto. pdf. Bollacker, K., Evans, C., Paritosh, P., Sturge, T., & Taylor, J.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

116

Bollacker, K., Evans, C., Paritosh, P., Sturge, T., & Taylor, J. (2008, June). Freebase: a collaboratively created graph database for structuring human knowledge. In Proceedings of the 2008 ACM SIGMOD international conference on Management of data (pp. 1247-1250). AcM. Auer, S., Bizer, C., Kobilarov, G., Lehmann, J., Cyganiak, R., & Ives, Z. (2007). Dbpedia: A nucleus for a web of open data. In The semantic web (pp. 722-735). Springer, Berlin, Heidelberg. Microsoft Concept Graph API. Retrieved from https://concept.research.microsoft.com/Home/Introduction [Accessed: 2017-10-28]. WikiData API. Retrieved from

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

117

from https://concept.research.microsoft.com/Home/Introduction [Accessed: 2017-10-28]. WikiData API. Retrieved from https://www.wikidata.org/wiki/Wikidata:Main_Page [Accessed: 2017-10-28]. Vrandečić, D., & Krötzsch, M. (2014). Wikidata: a free collaborative knowledgebase. Communications of the ACM, 57(10), 78-85. Wright, D. R. (2005). Finite state machines. Carolina State University, 203. Alicebot. Retrieved from http://www.alicebot.org/aiml.html [Accessed: 2017-10-28]. Liu, C. W., Lowe, R., Serban, I. V., Noseworthy, M., Charlin, L., & Pineau, J. How NOT To Evaluate Your Dialogue System: An Empirical Study of Unsupervised Evaluation

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

118

J. How NOT To Evaluate Your Dialogue System: An Empirical Study of Unsupervised Evaluation Metrics for Dialogue Response Generation.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

119

26 Iulian Vlad Serban, Alessandro Sordoni, Ryan Lowe, Laurent Charlin, Joelle Pineau, Aaron Courville, Yoshua Bengio. A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialogues. arXiv preprint arXiv:1605.06069 Luan, Y., Ji, Y., & Ostendorf, M. (2016). LSTM based Conversation Models. arXiv preprint arXiv:1603.09457. Kiros, R., Zhu, Y., Salakhutdinov, R. R., Zemel, R., Urtasun, R., Torralba, A., & Fidler, S. (2015). Skip-thought vectors. In Advances in neural information processing systems (pp. 3294-3302). Hochreiter, S., & Schmidhuber, J. (1997). Long

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

120

systems (pp. 3294-3302). Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780. Sukhbaatar, S., Szlam, A., Weston, J., & Fergus, R. (2015). Weakly supervised memory networks. CoRR, abs/1503.08895, 2. M. Iyyer, V. Manjunatha, J. Boyd-Graber, and H. Daumé III. Deep unordered composition rivals syntactic methods for text classification. In Proceedings of ACL , 2015. Guo, F., Metallinou, A., Khatri, C., Raju, A., Venkatesh, A., & Ram, A. (2018). Topic-based evaluation for conversational bots. In NIPS, 2017. Raju,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

121

Ram, A. (2018). Topic-based evaluation for conversational bots. In NIPS, 2017. Raju, A., Hedayatnia, B., Liu, L., Gandhe, A., Khatri, C., Metallinou, A., A., Venkatesh & Rastrow, A. (2018). Contextual Language Model Adaptation for Conversational Agents. Proc. Interspeech 2018, 3333-3337. Mehri, S., Hedayatnia, B., Raju, A., Gandhe, A., Khatri, C., Rastrow, A., Gabriel, R. & Mandal, A. (2018). Contextual Neural Language Models for Speech Recognition in Conversational Agents. Manuscript submitted for publication. Khatri, C., Goel, R., Hedayatnia, B., Metanillou, A., Venkatesh,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

122

Khatri, C., Goel, R., Hedayatnia, B., Metanillou, A., Venkatesh, A., Gabriel, R., & Mandal, A. (2018a). Contextual Topic Modeling For Dialog Systems. Manuscript accepted for publication. Available on arxiv. Khatri, C., Goel, R., Hedayatnia, B., Venkatesh, A., Gabriel, R., & Mandal, A. (2018b). Detecting Offensive Content in Open-domain Conversations using Two Stage Semi-supervision. Manuscript submitted for publication. Yi, S., Khatri, C., Goel, R., Hedayatnia, B., Venkatesh, A., Gabriel, R., & Mandal, A. (2018). Towards Coherent and Engaging Spoken Dialog Response Generation Using Automatic Conversation Evaluators. Manuscript

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

123

Towards Coherent and Engaging Spoken Dialog Response Generation Using Automatic Conversation Evaluators. Manuscript in preparation. 2018 Jigsaw. (2018.) Toxic comment classification challenge. Retrieved from https://kaggle.com/c/ 259 jigsaw-toxic-comment-classification-challenge. [Accessed: 2018-07-01]. Venkatesh, A., Khatri, C., Ram, A., Guo, F., Gabriel, R., Nagar, A., Prasad, R., Cheng, M., Hedayatnia, B., Metallinou, A. & Goel, R. (2018). On Evaluating and Comparing Conversational Agents. In NIPS, 2017. Surowiecki, J. (2004). The wisdom of crowds: Why the many are smarter than the few and how collective wisdom shapes business. Economies, Societies

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

124

Why the many are smarter than the few and how collective wisdom shapes business. Economies, Societies and Nations, 296. Flask Python Tool. http://flask.pocoo.org/ [Accessed: 2017-10-28]. Allen, J. (1995). Natural language understanding. Pearson.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

125

27 Pennington, J., Socher, R., & Manning, C. (2014). Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP) (pp. 1532-1543). Danescu-Niculescu-Mizil, C., & Lee, L. (2011, June). Chameleons in imagined conversations: A new approach to understanding coordination of linguistic style in dialogs. In Proceedings of the 2nd Workshop on Cognitive Modeling and Computational Linguistics (pp. 76-87). Association for Computational Linguistics. Sutskever, I., Martens, J., & Hinton, G. E. (2011). Generating text with recurrent neural networks. In Proceedings of the 28th International Conference on

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

126

(2011). Generating text with recurrent neural networks. In Proceedings of the 28th International Conference on Machine Learning (ICML-11) (pp. 1017-1024). Iyyer, M., Manjunatha, V., Boyd-Graber, J., & Daumé III, H. (2015). Deep unordered composition rivals syntactic methods for text classification. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) (Vol. 1, pp. 1681-1691). Klein, G., Kim, Y., Deng, Y., Senellart, J., & Rush, A. M. (2017). Opennmt: Open-source toolkit for neural machine translation. arXiv preprint

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

127

A. M. (2017). Opennmt: Open-source toolkit for neural machine translation. arXiv preprint arXiv:1701.02810. Sutskever, I., Vinyals, O., & Le, Q. V. (2014). Sequence to sequence learning with neural networks. In Advances in neural information processing systems (pp. 3104-3112). Curry, A. C., Papaioannou, I., Suglia, A., Agarwal, S., Shalyminov, I., Xu, X., Dušek, O., Eshghi, A.Yu, Y., & Lemon, O. (2018). Alana v2: Entertaining and Informative Open-domain Social Dialogue using Ontologies and Entity Linking. Alexa Prize Proceedings, 2018. Pichl, J., Marek, P., Konrád, J.,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

128

Alexa Prize Proceedings, 2018. Pichl, J., Marek, P., Konrád, J., Matulík, M., & Šedivý, J. (2018). Alquist 2.0: Alexa Prize Socialbot Based on Sub-Dialogue Model. Alexa Prize Proceedings, 2018. Fulda, N., Etchart, T., Myers, W., Ricks, D., Brown, Z., Szendre, J., Murdoch, B., Carr, A & Wingate, D. (2018). BYU-EVE: Mixed Initiative Dialog via Structured Knowledge Graph Traversal and Conversational Scaffolding. Alexa Prize Proceedings, 2018. Jonell, P., Bystedt, M., Dŏgan, F. I., Fallgren, Per., Ivarsson, J., Slukova, M., Wennberg, U.,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

129

F. I., Fallgren, Per., Ivarsson, J., Slukova, M., Wennberg, U., Lopes, José., Boye, Johan & Skantze, G. (2018). Fantom: A Crowdsourced Social Chatbot using an Evolving Dialog Graph. Alexa Prize Proceedings, 2018. Chen, C., Yu, D., Wen, W., Yang, Y. M., Zhang, J., Zhou, M., Jesse, K., Chau, A., Bhowmick, A., Iyer, S., Sreenivasulu, G., Cheng, R., Bhandare, A & Yu, Z. (2018). Gunrock: Building A Human-Like Social Bot By Leveraging Large Scale Real User Data. Alexa Prize Proceedings, 2018. Ahmadvand, A., Choi, I., Sahijwani, H., Schmidt,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

130

Proceedings, 2018. Ahmadvand, A., Choi, I., Sahijwani, H., Schmidt, J., Sun, M., Volokhin, S., Wang, Z & Agichtein, E. (2018). Emory IrisBot: An Open-Domain Conversational Bot for Personalized Information Access. Alexa Prize Proceedings, 2018. Bowden, K. K., Wu, J., Cui, W., Juraska, J., Harrison, V., Schwarzmann, B., Santer, N & Walker, M. (2018). SlugBot: Developing a Computational Model and Framework of a Novel Dialogue Genre. Alexa Prize Proceedings, 2018. Larionov, G., Kaden, Z., Dureddy, H. V., Kalejaiye, G. B. T., Kale, M., Potharaju, S. P.,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

131

V., Kalejaiye, G. B. T., Kale, M., Potharaju, S. P., Shah, A. P., & Rudnicky, A. I. (2018). Tartan: A retrieval-based socialbot powered by a dynamic finite-state machine architecture. Alexa Prize Proceedings, 2018.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

132

28 Pichl, J., Marek, P., Konrád, J., Matulík, M., Nguyen, H. L., & Šedivý, J. (2017). Alquist: The Alexa Prize Socialbot. Alexa Prize Proceedings, 2017. Bowden, K. K., Wu, J., Oraby, S., Misra, A., & Walker, M. (2018). Slugbot: An application of a novel and scalable open domain socialbot framework. Alexa Prize Proceedings, 2017. Prabhumoye, S., Botros, F., Chandu, K., Choudhary, S., Keni, E., Malaviya, C., Pasumarthi, R., Poddar, S., Ravichander, A. & Yu, Z. (2017). Building CMU Magnus from User Feedback. Alexa Prize Proceedings,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

133

& Yu, Z. (2017). Building CMU Magnus from User Feedback. Alexa Prize Proceedings, 2017. Damonte, B. K. M., Dobre, M., Duma, D., Fainberg, J., Fancellu, F., Kahembwe, E., Kahembwe, E., Cheng, J. & Webberz, B. (2017). Edina: Building an Open Domain Socialbot with Self-dialogues. Alexa Prize Proceedings, 2017. Serban, I. V., Sankar, C., Saizheng Zhang, Z. L., Subramanian, S., Kim, T., Chandar, S., Ke, N.R., Mudumba, S., de Brebisson, A., Sotelo, J.M. & Suhubdy, D. (2017). The Octopus Approach to the Alexa Competition: A Deep Ensemble-based

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

134

Suhubdy, D. (2017). The Octopus Approach to the Alexa Competition: A Deep Ensemble-based Socialbot. Alexa Prize Proceedings 2017. Fang, H., Cheng, H., Clark, E., Holtzman, A., Sap, M., Ostendorf, M.,Choi, Y. & Smith, N. A. (2017). Sounding board–university of washington’s alexa prize submission. Alexa Prize Proceedings, 2017. Adewale, O., Beatson, A., Buniatyan, D., Ge, J., Khodak, M., Lee, H., Prasad, N., Saunshi, N., Seff, A., Singh, K. & Suo, D. (2017). Pixie: A Social Chatbot. Alexa Prize Proceedings, 2017. Guss, W. H., Bartlett, J.,

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

135

A Social Chatbot. Alexa Prize Proceedings, 2017. Guss, W. H., Bartlett, J., Kuznetsov, P., & Patil, P. (2017). Eigen: A Step Towards Conversational AI. Alexa Prize Proceedings. Liu, H., Lin, T., Sun, H., Lin, W., Chang, C. W., Zhong, T., & Rudnicky, A. (2017). RubyStar: A Non-Task-Oriented Mixture Model Dialog System. Alexa Prize Proceedings, 2017. Papaioannou, I., Curry, A. C., Part, J. L., Shalyminov, I., Xu, X., Yu, Y., Dušek, O., Rieser, V. & Lemon, O. (2017). Alana: Social dialogue using an ensemble model and a ranker trained on

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

1812.10757

136

Lemon, O. (2017). Alana: Social dialogue using an ensemble model and a ranker trained on user feedback. Alexa Prize Proceedings, 2017. Williams, Jason D., Kavosh Asadi, and Geoffrey Zweig. (2017). Hybrid code networks: practical and efficient end-to-end dialog control with supervised and reinforcement learning. arXiv preprint arXiv:1702.03274, 2017.

1812.10757

Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize

Building open domain conversational systems that allow users to have engaging conversations on topics of their choice is a challenging task. Alexa Prize was launched in 2016 to tackle the problem of achieving natural, sustained, coherent and engaging open-domain dialogs. In the second iteration of the competition in 2018, university teams advanced the state of the art by using context in dialog models, leveraging knowledge graphs for language understanding, handling complex utterances, building statistical and hierarchical dialog managers, and leveraging model-driven signals from user responses. The 2018 competition also included the provision of a suite of tools and models to the competitors including the CoBot (conversational bot) toolkit, topic and dialog act detection models, conversation evaluators, and a sensitive content detection model so that the competing teams could focus on building knowledge-rich, coherent and engaging multi-turn dialog systems. This paper outlines the advances developed by the university teams as well as the Alexa Prize team to achieve the common goal of advancing the science of Conversational AI. We address several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation. These collaborative efforts have driven improved experiences by Alexa users to an average rating of 3.61, the median duration of 2 mins 18 seconds, and average turns to 14.6, increases of 14%, 92%, 54% respectively since the launch of the 2018 competition. For conversational speech recognition, we have improved our relative Word Error Rate by 55% and our relative Entity Error Rate by 34% since the launch of the Alexa Prize. Socialbots improved in quality significantly more rapidly in 2018, in part due to the release of the CoBot toolkit.

http://arxiv.org/pdf/1812.10757

2018 Alexa Prize Proceedings

cs.CL

20181227

20181227

2211.00053

0

Preprint. Under review. GENERATING SEQUENCES BY LEARNING TO [SELF-]CORRECT Sean Welleck1,3,*Ximing Lu1,* Peter West3,yFaeze Brahman1,y Tianxiao Shen3Daniel Khashabi2Yejin Choi1,3 1Allen Institute for Artificial Intelligence 2Center for Language and Speech Processing, Johns Hopkins University 3Paul G. Allen School of Computer Science & Engineering, University of Washington ABSTRACT Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present SELF -CORRECTION ,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

1

their parameters for task-specific adaptation. We present SELF -CORRECTION , an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that SELF CORRECTION improves upon the base generator in three diverse generation tasks– mathematical program synthesis, lexically-constrained generation, and toxicity control– even when the corrector is much smaller than the base generator. 1 I NTRODUCTION The standard practice for natural language generation tasks is inherently single-pass: applying a decoding procedure to either a few-shot prompted language model or one tuned for a given task, then considering the generation as “finished” (e.g. Radford et al. (2019); Brown et al. (2020); Chen et al. (2021)). Powerful generation models often meet most of the task requirements, yet miss a few (e.g., omitting a subset of keywords), or generate incorrect hypotheses that nevertheless provide

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

2

few (e.g., omitting a subset of keywords), or generate incorrect hypotheses that nevertheless provide useful structure (e.g., a correct problem solving strategy with a missing step). However, after generating even a slightly sub-optimal sequence, the single-pass paradigm requires models to “start from scratch”, effectively discarding work already done. A more natural, intuitive approach is leveraging the generation as a useful starting point to refine into a higher quality output. To formalize this intuition, we introduce Self-Correction for Sequence Generation . Figure 1 demonstrates its central principle: a generation model is re-framed as a base generator , which produces a reasonable initial hypothesis but does not need to solve the task in one pass, and a second module–the corrector –trained to make up the difference between the hypothesis and an optimal solution. Neither the generator nor the corrector must solve the full task in one pass, and the corrector can be applied multiple times to iteratively improve the output (§3.6). We propose a simple, general procedure for training the corrector (Figure 2) by pairing generator outputs with carefully selected targets. The result is a system which self-corrects, producing outputs through multiple generation passes and breaking the task into steps that can be solved by dedicated and efficient sub-systems.

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

3

passes and breaking the task into steps that can be solved by dedicated and efficient sub-systems. First authors, contributed equally. ySecond authors, contributed equally. 1arXiv:2211.00053v1 [cs.CL] 31 Oct 2022 Preprint. Under review. Figure 1: S ELF-CORRECTOR s decompose generation into a base generator that proposes an initial hypothesis, and a corrector that iteratively improves its quality. We find that Self-Correction is broadly applicable. Training a corrector model improves the base generator on 3 diverse tasks: mathematical program synthesis (§3.1), lexically constrained generation (§3.2), and toxicity reduction (§3.3). The trained corrector model can even be applied to a larger generator with similar performance to training a new corrector (§3.4), showing that the subtask of correction is transferable, even to stronger generators. Finally, we explore the prospect of introducing a third module to the Self-Correction system (§3.5)–explicitly using natural language feedback to guide corrections–with promising results. Self-Correction offers an exciting opportunity to build on existing generation models and the sequences they generate, with efficient, effective, and transferable corrector networks.

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

4

and transferable corrector networks. 2 S ELF-CORRECTING SEQUENCE GENERATORS A typical autoregressive text generator (e.g. GPT-3 (Brown et al., 2020)) maps an input prompt to a distribution over outputs using a single parameterized module (e.g. a large transformer), p0(yjx). We explore an alternative that decomposes into two modules, a base generator , and a corrector , p(yjx) =X y0p0(y0jx)|{z} generatorp(yjy0;x)|{z} corrector(1) where the generator provides an initial hypothesis that is refined by the corrector. In practice, the corrector can be applied multiple times, p(yTjx) =P y0P y1


P yT1p0(y0jx)Q tp(yt+1jyt;x). Since a model of this form can both generate and correct its generations, we call it a Self-Corrector . Self-correctors have several unique properties compared to typical generators. First, a self-corrector decouples generation and correction, allowing us to freely parameterize each module – for instance,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

5

Self-correctors have several unique properties compared to typical generators. First, a self-corrector decouples generation and correction, allowing us to freely parameterize each module – for instance, by prompting a single language model or using two different language models. In this paper, we develop a framework to train a separate corrector model (§2.1). We find that the resulting selfcorrector improves upon the generator alone (§3), even when the corrector is much smaller (§3.4). Second, since the generator and the corrector are separated, we can keep the generator as a generalpurpose language model and train the corrector with different objectives for different task requirements. In §2.1, we propose a training algorithm for the corrector that is dedicated to improving generations, where the improvement can be in any aspect, measured by scalar values. Third, the corrector can receive explicit feedback about intermediate generations to guide subsequent generations. Formally, p(yjx) =P y0p0(y0jx)p(yjy0;x;f (y0)), wherefis the feedback. The feedback can be of many forms, e.g. a sentence, a compiler trace, etc. In contrast, a typical generator that generates in a single pass does not leverage feedback on its own generation. In this paper, we

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

6

that generates in a single pass does not leverage feedback on its own generation. In this paper, we show that the corrector can learn to exploit explicit natural language feedback to achieve better performance (§3.5). Next, we describe our training framework of the corrector. 2.1 L EARNING A CORRECTOR Our goal is to have the generator generate an initial hypothesis, then improve the hypothesis with the corrector (Eq. 1). We train the corrector to improve the quality of a hypothesis, while staying as 2 Preprint. Under review. Figure 2: S ELF-CORRECTIVE LEARNING iteratively trains a corrector by generating hypotheses and corrections, forming value-improving pairs, and selecting those with high similarity for learning. Algorithm 1 Self-corrective learning input Generatorp0, correctorp, promptsX, valuev(
), feedbackf(
) Initialize datapool Dby sampling from p0 .Initialization: Eq. 2 foriteration2f1;2;:::gdo forx2Xdo Sample hypotheses yfrom datapool D Generate corrections y0p(
jy;x;f (y)) Add all (x;y0;v(y0);f(y0))to the datapool D . Exploration: Eq. 5

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

7

Add all (x;y0;v(y0);f(y0))to the datapool D . Exploration: Eq. 5 Form value-improving pairs PfromD . Pairing: Eq. 3 forstep in 1;2;:::;M do Sample a batch of value-improving pairs from Pusing Eq. 4 Compute the loss and update using gradient descent .Learning close as possible to the original hypothesis. Here, quality is measured with a scalar value function v(y)which we assume is accessible at training time (e.g. a classifier). Since direct supervision on how to improve hypotheses is not available, we design a new algorithm to train the corrector, which we refer to as self-corrective learning. The algorithm collects a pool of generations, groups them and selects pairs of generation that increase in value and are nearby, then updates the corrector on these pairs. As training progresses, more generations are added to the pool using the current corrector. Algorithm 1 summarizes self-corrective learning, detailed below. Initialization. Self-corrective learning begins with a generator p0(y0jx), a corrector p(y0jy;x),

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

8

Initialization. Self-corrective learning begins with a generator p0(y0jx), a corrector p(y0jy;x), a set of training prompts X, and a value function v:Y!R. Optionally, we can use additional feedbackf:Y!F and learnp(y0jy;x;f (y)), whereFis arbitrary. The algorithm initializes a datapool of (input, output, value, feedback) examples by using the generator to generate multiple outputs for each input. Formally, Dx=f(x;y;v (y);f(y))jfor ally2y1:Nq(p0(
jx))g; D =[ x2XDx; (2) wherey1:NdenotesNoutputs generated with decoding algorithm q(e.g. temperature sampling). When available, (x;y;v (y);f(y))examples from another source (e.g. a dataset) can also be added. Pairing. Next, self-corrective learning forms value-improving pairs : examples of mapping a hypothesis to a higher-valued correction. We use the datapool Dto form a set of (input, hypothesis,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

9

correction) pairs. A pair is formed when an output has a higher value than another1: Px=f(x;y;y0)jv(y)<v(y0)for ally;y02DxDxg; P =[ x2XPx; (3) 1We also store the value and feedback for yandy0along with (x;y;y0), which we omit to reduce clutter. 3 Preprint. Under review. Learning. Next, self-corrective learning selects (input, hypothesis, correction) pairs to update the corrector with. We sample a (x;y;y0)pair proportional to its improvement in value as well as the proximity between the hypothesis yand the correction y0: P[(x;y;y0)]/exp 
(v(y0)v(y))|{z} improvement+
s(y;y0)|{z} proximity
=Z(y); (4) wheres(y;y0)is a similarity function and Z(y)normalizes over the available corrections for yin Px. Increasing the hyperparameter 2R0puts more weight on targets that add more value, while

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

10

Px. Increasing the hyperparameter 2R0puts more weight on targets that add more value, while increasing2R0retains more similar targets. We update the corrector using the cross-entropy lossL() =logp(y0jy;x;f (y))on batches sampled in this way. Exploration. During exploration, self-corrective learning adds new generations to the datapool by generating from the current corrector: D0 x=f(x;y0;v(y0);f(y0))jfor ally02y01:Nq(p(
jy;x;f (y))g; D0=[ x2XD0 x (5) and updating the datapool D D[D0. The hypotheses yto correct can come from any source, e.g. newly sampled from the base generator, or from the datapool; we use the latter in our experiments. Inference. We use the trained corrector along with a generator to generate a trajectory y0;y1;:::;yT, and consider yTthe final output. Since marginalizing over the intermediate generations in Eq. 1 is intractable, we approximate each summation with a single sequence generated

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

11

with a decoding algorithm q(
). That is, we decode from the generator, then repeatedly from the corrector: • Generation: y0q(p0(y0jx)); • Correction: yt+1q(p(yt+1jyt;x;f (yt))),t= 0;1;:::;T1. The stopping time Tis either fixed, or when a target value is obtained (if v(y)is available). 3 E XPERIMENTS We evaluate SELF -CORRECTION on a diversity of tasks: mathematical program synthesis , in which generations are strictly correct or incorrect, and generators typically have low performance; lexically-constrained generation , which allows for partial credit, and generators usually give partially-correct solutions (e.g. matching 3 out of 5 constraints); and toxicity control , where ‘correctness’ is more loosely defined, and the output space is much more open-ended. Our experiments are organized to study three settings: 1. Using self-correctors to improve upon generators (§3.1,3.2,3.3). 2. Correcting generators that are much larger than the corrector (§3.4). 3. Leveraging explicit feedback during training and inference (§3.5).

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

12

2. Correcting generators that are much larger than the corrector (§3.4). 3. Leveraging explicit feedback during training and inference (§3.5). Next, we describe the self-correction setup and baselines for each task, along with their results.2 3.1 M ATHEMATICAL PROGRAM SYNTHESIS First, we consider mathematical program synthesis (Austin et al., 2021; Mishra et al., 2022). Given a natural language problem specification x, the task is to generate a program ythat upon execution returns the correct answer to x. The task is challenging as it draws on language understanding, multiple-step mathematical problem solving (e.g. identifying a solution strategy, decomposing a problem), and leveraging symbolic tools (e.g. built-in operations, variables). Furthermore, the task demands a high level of precision, e.g. a single misplaced operation makes the program incorrect. Experimental setup. As the corrector we use GPT-Neo 1.3B (Black et al., 2021), an open-source autoregressive language model. GPT-Neo is pre-trained on language and code (Gao et al., 2021), and hence is widely used for code-related generation (e.g. Chen et al. (2021); Ni et al. (2022);

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

13

and hence is widely used for code-related generation (e.g. Chen et al. (2021); Ni et al. (2022); Mishra et al. (2022)). We consider two settings for the initial generator: (1) a separate fine-tuned 2Code will be publicly available upon acceptance. 4 Preprint. Under review. Dataset Model Correct Multiarith GPT-NEO 1.3B 60.00 +SELF-CORRECT 98.33 +SELF-CORRECT 99.17 Multitask GPT-NEO 1.3B 49.02 +SELF-CORRECT 73.53 +SELF-CORRECT 78.24Dataset Model Params Correct GSM OpenAI 3B [5] 3B 15.50 OpenAI 6B [5] 6B 20.00 GPT-NEO [28] 2.7B 18.80 NEO FCP+PCP [28] 2.7B 19.50 GPT-NEO 1.3B 8.57 +SELF-CORRECT 1.3B 21.26 +SELF-CORRECT 1.3B 24.22 Table 1: Evaluation results of mathematical program synthesis experiments. GPT-NEO (1.3B) is

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

14

+SELF-CORRECT 1.3B 24.22 Table 1: Evaluation results of mathematical program synthesis experiments. GPT-NEO (1.3B) is the initial generator for S ELF-CORRECT . SELF-CORRECTmeans only applying the corrector to incorrect outputs. Italicized : original non-program version of GSM. Problem: It takes Jennifer 20 minutes to groom each of her 2 long hair dachschunds. If she grooms her dogs every day, how many hours does she spend grooming her dogs in 30 days? Generator: a=20*2 b=a*30 answer=b print (answer)Corrector: a=20*2 b=a*30 c=b/60 #fix answer=c print (answer)Problem: Mrs. Wilsborough saved $500 to buy concert tickets for her family. She bought 2 VIP tickets at $100 each and 3 regular tickets at $50 each. How much of her savings does Mrs. Wilsborough have after she buys the tickets? Generator: a=2*100 b=3*50 c=a+b answer=c print (answer)Corrector: a=2*100 b=3*50

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

15

b=3*50 c=a+b answer=c print (answer)Corrector: a=2*100 b=3*50 c=500-a-b #fix answer=c print (answer) Figure 3: Grade-school-math (GSM) self-corrections. On the left, the corrector fixes the units (from minutes to hours) in the generator’s solution. On the right, the corrector revises the logic so that the program computes the total savings instead of the spent on tickets. We add #fixhere to indicate the change. See Figure 7 and Figure 8 for additional examples. instance of GPT-Neo 1.3B, and (2) few-shot prompted GPT-3 (Brown et al., 2020). For GPT-3, we evaluate the davinci and text-davinci-002 engines, representative of large ( 175B3) generators that are state-of-the-art in related tasks (Wei et al., 2022). See the Appendix for additional details. Self-correction setup. As the value function we use correctness, which is 1 when the program y executes and outputs the ground-truth answer and 0 otherwise. Our main experiments do not use

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

16

executes and outputs the ground-truth answer and 0 otherwise. Our main experiments do not use explicit feedback, i.e. f(y) =;. At inference time, we study two settings for the corrector: (1) applyingkcorrections and selecting the final generation, (2) an oracle setting that only corrects a draft if the draft is incorrect. We use greedy decoding for the generator and corrector, and k= 1. Datasets. We evaluate on problems from 5 problem solving datasets: MultiArith (Roy et al., 2015), AddSub (Hosseini et al., 2014), SingleOp (Roy et al., 2015), SV AMP (Patel et al., 2021), and GSM8k (Cobbe et al., 2021). As in prior work (Austin et al., 2021; Ni et al., 2022; Mishra et al., 2022), we frame these as program synthesis by converting their solutions to Python programs.4We separate our experiments into three increasingly difficult settings: 1.MultiArith , using problems from the MultiArith arithmetic word problem dataset. 2.Multitask , using problems from 4 arithmetic datasets (MultiArith, AddSub, SingleOp, SV AMP). 3.GSM , using problems from the challenging GSM8k dataset.

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

17

3.GSM , using problems from the challenging GSM8k dataset. For the MultiArith and Multitask settings, we make train/valid/test splits using 60/20/20% of the respective datasets. Similar to Ni et al. (2022), for the GSM setting we use the official GSM8k test split, and create a validation split using 20% of the training set. Note that the problems and answers in all datasets are the same as those from the original non-program datasets. 3Estimated size of davinci (https://blog.eleuther.ai/gpt3-model-sizes ). Further details not available. 4We use data from the Lila benchmark ( https://github.com/allenai/Lila ). 5 Preprint. Under review. Method Runtime CIDER Constraints NeuroLogic [22] 2.04s 14.70 97.70 NeuroLogic-A* [24] 19.24s 15.20 97.80 GPT-2 0.20s 14.97 91.38 SELF-CORRECT 0.80s 15.30 94.58 +NeuroLogic 2.24s 15.28 97.80Method Fluency Constraints Prefix-Tuning [17] 2.96 91.16

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

18

+NeuroLogic 2.24s 15.28 97.80Method Fluency Constraints Prefix-Tuning [17] 2.96 91.16 NeuroLogic [22] 2.80 96.91 NeuroLogic-A* [24] 2.85 96.97 GPT-2 2.94 91.50 SELF-CORRECT 2.98 98.77 Table 2: Lexically-constrained generation. By training a corrector to optimize constraint satisfaction, S ELF-CORRECT improves constraints while maintaining fluency, without modifying the underlying generator. Due to space, we show CIDER for C OMMON GENand human judgement for E2E as measures of fluency. Other metrics show similar trends and can be found in the Appendix. Baselines. We compare SELF -CORRECT with its baseline generator (GPT-Neo 1.3B) in all three settings. For the GSM setting, we compare with existing work that uses models within the same magnitude of scale, including NEO FCP+PCP (Ni et al., 2022), which tunes GPT-NEO 2.7B with additional self-sampled programs, and their fine-tuned GPT-NEO 2.7B baseline. We also report 3B

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

19

additional self-sampled programs, and their fine-tuned GPT-NEO 2.7B baseline. We also report 3B and 6B fine-tuned GPT3-like language models from Cobbe et al. (2021), which were trained on the non-program version of GSM8k. We evaluate larger models later in (§3.4). Results. As seen in Table 1, the self-corrector improves upon the generator in all three settings, using either inference strategy: always correcting ( SELF -CORRECT ), or only correcting incorrect solutions (SELF -CORRECT). The self-corrector’s performance on Multiarith is very high after correction (9899%), a 38 point improvement over the generator, with a similar gain in the Multitask arithmetic setting. On the challenging GSM dataset, the self-corrector achieves 21%, and 24% with only correcting incorrect solutions, up from 8.57% for the generator. Notably, this is higher than previous work based on the larger 2.7B GPT-Neo, or larger models tuned on the language version of GSM. The results show that self-corrective learning can improve task performance via training a corrector. Qualitatively, the self-corrector can correct values in a correctly structured solution, fix the order of

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

20

Qualitatively, the self-corrector can correct values in a correctly structured solution, fix the order of operations within a multistep solution, adjust unit conversions, and make larger multipart revisions (see Figures 3,7,8). Notably, these are learned automatically through self-corrective learning. 3.2 L EXICALLY CONSTRAINED GENERATION Next, we consider lexically constrained generation. Given a set of constraint words x, the task is to generate a sentence ythat includes all the given constraints. Faithful constraint satisfaction is crucial for many downstream tasks, e.g., those that require converting information to text (McKeown, 1985). Datasets and Metrics. We experiment on C OMMON GEN(Lin et al., 2020) and E2E (Novikova et al., 2017). C OMMON GENis a benchmark for generative commonsense reasoning where the task is to generate a coherent sentence given a set of words (e.g., dog, catch). E2E involves converting structured inputs into natural language. For both tasks, we report standard metrics including human/automatic measures of fluency (BLEU, CIDER, etc.) as well as constraint coverage. We collect human measures of fluency on Amazon Mechanical Turk; see the Appendix for details.

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

21

collect human measures of fluency on Amazon Mechanical Turk; see the Appendix for details. Setup. We parameterize the base generator with GPT-2 Radford et al. (2019) (large-size for C OMMON GENand medium-size for E2E). We fine-tuned the generator for each task. As the value function for self-corrective learning we use coverage, i.e. the percentage of constraints that are present in the output. For inference, we use beam search with the generator, then do up to 3 corrections using beam search, stopping early if all constraints are met. See the Appendix for additional details. Results. Table 2 shows the evaluation results. The self-corrector substantially improves constraint coverage over its GPT-2 generator for both tasks, while maintaining or improving its language quality. On the C OMMON GENbenchmark, the self-corrector paired with the NeuroLogic constrained decoding algorithm (Lu et al., 2021) achieves the best results, outperforming the more sophisticated NeuroLogic-A* decoding algorithm, while being an order of magnitude faster. Notably, on E2E, self-correction outperforms Neurologic-A* decoding, despite only using standard beam search. This suggests that a corrector can be viewed as an alternative to using a more sophisticated decoding

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

22

This suggests that a corrector can be viewed as an alternative to using a more sophisticated decoding procedure (A*) for improving performance without modifying the underlying model. See Figure 9 for qualitative examples. 6 Preprint. Under review. Toxicity Fluency Diversity Avg. Max .Prob. Perplexity dist-2 dist-3 GPT-2 0.527 0.520 11.31 0.85 0.85 PPLM [6] 0.520 0.518 32.58 0.86 0.86 GeDi [14] 0.363 0.217 43.44 0.84 0.83 DExpert [21] 0.314 0.128 25.21 0.84 0.84 DAPT [12] 0.428 0.360 31.22 0.84 0.84 PPO [23] 0.218 0.044 14.27 0.79 0.82 Quark [23] 0.196 0.035 12.47 0.80 0.84 SELF-CORRECT 0.171 0.026 11.81 0.80 0.83 Table 3: Toxicity reduction. GPT-2 is the base generator. Figure 4: Applying multiple corrections reduces toxicity. 3.3 T OXICITY REDUCTION

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

23

Table 3: Toxicity reduction. GPT-2 is the base generator. Figure 4: Applying multiple corrections reduces toxicity. 3.3 T OXICITY REDUCTION Next, we consider the task of toxicity reduction (Gehman et al., 2020; Liu et al., 2021). Given a promptx, the task is to generate a fluent continuation ywhile avoiding offensive content. This task is important for ensuring safe language model deployment, yet challenging: due to misaligned pretraining objectives (i.e. modeling internet text vs. non-toxic text), language models are susceptible to generating toxic completions, even when prompted with seemingly innocuous text (Gehman et al., 2020). Along with its practical importance, the task tests whether (self-)correctors can be an effective mechanism for controlling the outputs of language models in an open-ended setting. Datasets and Metrics. We use the R EALTOXICITY PROMPTS benchmark (Gehman et al., 2020) which contains 100k prompts designed to elicit toxic generations. Following the experimental setup of Liu et al. (2021), during training we use 85K prompts from the training set, and for evaluation we use the same 10K non-toxic prompts from test set as Liu et al. (2021). We use Perspective API to

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

24

use the same 10K non-toxic prompts from test set as Liu et al. (2021). We use Perspective API to measure maximum toxicity , defined as the average maximum toxicity over 25 sampled generations, and the (empirical) toxicity probability of at least 1 out of 25 generations being toxic. Baselines. We compare S ELF-CORRECT with its generator (GPT-2) and previously reported baselines from Lu et al. (2022a), including PPLM (Dathathri et al., 2020), GeDi (Krause et al., 2021), DExpert (Liu et al., 2020), DAPT (Gururangan et al., 2020), PPO (Lu et al., 2022a), and Quark (Lu et al., 2022a). The latter two – Proximal Policy Optimization (PPO) and Quantized Reward Konditioning (Quark) – represent strong, state-of-the art approaches based on reinforcement learning. Setup. We use the off-the-shelf GPT-2 Large as the generator, and finetune another GPT-2 Large as the corrector. During inference, we use nucleus sampling with p= 0:9to generate 25 samples for all

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

25

the corrector. During inference, we use nucleus sampling with p= 0:9to generate 25 samples for all baselines. As the value function, we use the Perspective API score, v(y)2[0;1], which measures the toxicity of the completed sequence. We do up to three corrections with the corrector model. Results. Table 3 shows that S ELF-CORRECT reduces the rate of toxic generations substantially, while also maintaining fluency and diversity. S ELF-CORRECT outperforms all baselines. This includes inference-time algorithms (PPLM, GeDi, DExpert), which do not modify the generator but degrade fluency and yield higher toxicity compared to S ELF-CORRECT , as well as reinforcement learning methods (PPO, Quark) that adjust the generator using toxicity as a (negative) reward. The results show that S ELF-CORRECT is effective for detoxification, without having to modify the underlying generator. We study implications of this latter property further in the next section. 3.4 C HANGING MODULES – C ORRECTING GPT-3 Next, we show that a self-corrector can improve the outputs of a generator that is much larger than the corrector. We consider two cases: (1) training with a small generator, then swapping in the larger

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

26

the corrector. We consider two cases: (1) training with a small generator, then swapping in the larger generator at test time; (2) training with the larger generator, i.e. using the large generator to initialize the datapool for self-corrective learning, then using the large generator at test time. Toxicity. We evaluate case (1) for reducing the toxicity of a large generator (GPT-2 XL, GPT-3). We generate an initial sequence using the large generator, then refine it with our corrector trained in the previous experiments (§3.3). Table 4 shows that the resulting self-corrector (large generator 7 Preprint. Under review. Task Dataset Generator (train) Generator (test) Generator Self-corrector Math Synthesis"Neo 1.3B GPT-3 46.70 80.00 Multitask Neo 1.3B GPT-3 Instruct 84.90 90.90 GPT-3 Instruct GPT-3 Instruct 84.90 92.75 Neo 1.3B GPT-3 6.96 24.30 GSM Neo 1.3B GPT-3 Instruct 36.80 45.00 GPT-3 Instruct GPT-3 Instruct 36.80 45.92

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

27

GSM Neo 1.3B GPT-3 Instruct 36.80 45.00 GPT-3 Instruct GPT-3 Instruct 36.80 45.92 Detoxification#GPT2-L GPT2-XL 0.383 0.027 RTPrompts GPT2-L GPT-3 0.182 0.025 GPT2-L GPT-3 Instruct 0.275 0.023 Table 4: Modularity (program synthesis and detoxification). Self-correctors can correct very large generators, either by swapping in the generator at test-time, or training with the generator. For math synthesis, the corrector is GPT-Neo 1.3B, and here we only correct incorrect outputs. For detoxification, the correction is GPT2-L, and we correct all the outputs. + corrector) has substantially reduced toxicity compared to the large generator. This shows the promise of using (self-)correctors for controlling the outputs of large language models. Math program synthesis. Table 4 shows results for math. Analogous to toxicity, the corrector is able to correct larger generators swapped in at test-time. For instance, the GPT-3 Instruct generator

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

28

able to correct larger generators swapped in at test-time. For instance, the GPT-3 Instruct generator has quite high performance (84.90 Multitask, 36.80 GSM), which improves to 90.90 and 45.00, respectively, by adding in a corrector. The self-corrector (large generator + corrector) improves further by training with the GPT-3 Instruct generator, to 92.75 and 45.92, respectively. 3.5 L EVERAGING EXPLICIT FEEDBACK Next, we demonstrate S ELF-CORRECT ’s capacity to incorporate explicit natural language feedback. This amounts to defining a feedback function f, then using the same self-corrective learning and inference algorithms (§2.1) as in our preceding experiments (in those experiments, freturned;). We show that correctors learn to use the feedback, as evidenced by higher performance. Toxicity. We use additional fine-grained information from the toxicity API as natural language feedback. Specifically, besides the overall toxicity score, Perspective API also provides scores for fine-grained attributes of toxicity (e.g. identity attack, profanity, flirtation, etc.). At training time,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

29

fine-grained attributes of toxicity (e.g. identity attack, profanity, flirtation, etc.). At training time, we compare the attribute scores from a hypothesis and its selected correction, and use the attribute with the largest decrease as natural language feedback (e.g. "decrease toxicity in profanity "). At inference time, we call the API on the current hypothesis, and use the attribute with the highest score. Here we use the API at inference time, which is notrequired in our previous experiments. Lexical constraints. In training time, we generate natural language feedback for every example pair (x;y;y0)by elaborating the extra lexical constraints satisfied by y0but noty. e.g. “adding constraint word: read” . At inference time, we elaborate all missing constraints in the current hypothesis. Math program synthesis. Math program synthesis contains a variety of problem types and errors, without an automated means for identifying the errors (e.g. an API). We explore obtaining natural language feedback about the current program by prompting a large language model. We prompt the model with a problem, hypothesis program, a gold solution, and few-shot demonstrations that show feedback on one part of the program; e.g. In the initial guess, 3 should be subtracted. When

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

30

show feedback on one part of the program; e.g. In the initial guess, 3 should be subtracted. When the program is correct, the feedback is Correct. At inference time, we also use feedback from the language model. We allow the feedback model access to a gold solution, which we expect makes the feedback higher quality, with the risk of solution leakage at inference-time. Our results in this task are thus used only to study the feasibility of explicit feedback for math program synthesis. Setup. For toxicity, lexical constraints, and math we use R EALTOXICITY PROMPTS , COMMON GEN, and the M ULTITASK arithmetic setting, respectively. We follow the setup of each task’s previous experiments (§3.3,§3.2,§3.1), except for math we use 5 correction iterations (previously 1). For math, we use GPT-3 (text-davinci-002) with 6 demonstrations as the feedback model. 8 Preprint. Under review. Toxicity# Constrained Gen. " Math" Avg. Max .Prob. Fluency Fluency Constraints Correct Correct  Generator 0.527 0.520 11.31 14.97 91.38 49.02 49.02

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

31

Generator 0.527 0.520 11.31 14.97 91.38 49.02 49.02 SELF-CORRECT 0.171 0.026 11.81 15.30 94.58 74.31 79.80 +FEEDBACK 0.156 0.020 11.86 15.24 95.88 81.76 82.35 Table 5: Explicit natural language feedback. Correctmeans only correcting incorrect outputs. Problem: Melanie had 19 dimes in her bank. Her dad gave her 39 dimes and her mother gave her 25 dimes. How many dimes does Melanie have now? Generator (GPT-Neo): answer = 19 + 25 print (answer) Feedback (GPT-3): # In the initial guess, # 39 is not included. Corrector (GPT-Neo): answer = 19 + 25 + 39 print (answer)Problem: Lana’s favorite band was holding a concert where tickets were 6 dollars each. Lana bought 8 tickets for herself and her friends and 2 extra tickets in case anyone else wanted to go. How much did she spend? Generator (GPT-Neo): answer=(6.0*8.0) print (answer) Feedback (GPT-3): # In the initial guess,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

32

answer=(6.0*8.0) print (answer) Feedback (GPT-3): # In the initial guess, # 2 tickets are not included. Corrector (GPT-Neo): answer=(6.0*(8.0+2.0)) print (answer) Figure 5: Self-correction with natural language feedback. Results. Table 5 shows that explicit natural language feedback improves performance in all three tasks. For toxicity, this means that providing fine-grained attributes (e.g. identity attack, profanity, etc.) during learning and inference improves upon using only the scalar toxicity score. Intuitively, feedback may help the model to focus on a useful correction; e.g., see Figure 5. 3.6 A DDITIONAL ABLATIONS AND ANALYSIS Effect of multiple corrections. Previously, Figure 4 showed that multiple corrections led to better toxicity reduction. On math (Multitask setting), Figure 6 shows that performance improves with more than one correction, and that multiple corrections are more beneficial with feedback. Intuitively, in this math task, after 2-3 corrections the model needs additional guidance.

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

33

more than one correction, and that multiple corrections are more beneficial with feedback. Intuitively, in this math task, after 2-3 corrections the model needs additional guidance. Effect of pairing and proportional sampling. Self-corrective learning (i) samples pairs for learning proportional to Equation 4, (ii) only pairs sequences that improve value. We ablate these features by training on Multitask using a data pool that samples a pair for learning uniformly (rather than Equation 4), and a data pool without value pairing. Table 6 shows that both improve performance. Effect of exploration. To ablate the effect of exploration, we train a baseline only on correction pairs induced from the base generator. Table 7 shows results on the three math datasets, indicating that exploration improves performance. 4 R ELATED WORK Self-correction relates to recent works on editing text, including modeling Wikipedia edits (Reid & Neubig, 2022; Faltings et al., 2021; Schick et al., 2022), which relies on supervised edits, unsupervised methods (Miao et al., 2019; Liu et al., 2020) that perturb sequences with simple operations (e.g. insertion, deletion), editing with models trained on human-written critiques (Saunders et al.,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

34

(e.g. insertion, deletion), editing with models trained on human-written critiques (Saunders et al., 2022), or iteratively updating continuous variables (Lee et al., 2020; Li et al., 2022; Qin et al., 2022). 9 Preprint. Under review. Figure 6: Math: multiple corrections.Ablation Math C OMMON GEN SELF-CORRECT 78.24 94.55 7proportional sampling 77.25 93.49 7value pairing 62.35 91.76 Table 6: Effect of pairing and proportional sampling. Exploration Multiarith Multitask GSM8k 7 89.20 73.49 17.60 3 99.17 78.24 23.96 Table 7: Effect of exploration on program synthesis. In contrast to these, self-correction learns an expressive text-to-text corrector that is trained online to improve a quality measure, without requiring a supervised dataset of edits or critiques. Separately, denoising ground-truth sequences is a common pretraining objective (Devlin et al., 2019; Lewis et al., 2020; Raffel et al., 2020), while self-correction ‘denoises’ generations to improve a scalar quality measure. Scalar measures are often improved with reinforcement learning (RL) on a base

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

35

quality measure. Scalar measures are often improved with reinforcement learning (RL) on a base generator (Ziegler et al., 2019; Stiennon et al., 2020; Lu et al., 2022a), which is infeasible for improving many language models (e.g. those accessed through an API), and uses only scalar feedback. Moreover, self-correction learns the difference between a generation and solution, and is complementary to RL-tuned generators, which can be used within a self-corrector. Finally, self-correction decomposes generation into multiple steps, which relates to methods that generate rationales before a response (Wei et al., 2022; Dohan et al., 2022). Self-correction also produces intermediate steps, but each step is of the same form as the output, allowing for re-using previous generations. 5 C ONCLUSION We introduced self-correctors, a class of models that decompose generation into initial generation and correction steps. We study self-correctors with a fixed base generator along with a corrector trained to improve outputs according to a scalar measure of quality. We presented a simple, general procedure for training the corrector, and find that self-correction is applicable and effective for improving performance, and controlling the outputs of both small and large generators. Moreover,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

36

improving performance, and controlling the outputs of both small and large generators. Moreover, we found that self-correction along with our learning framework provides a promising mechanism for using natural language feedback to improve generation. These findings, along with exploring alternative self-correctors, open up many avenues that we leave for future work. ACKNOWLEDGMENTS This work was funded in part by the DARPA MCS program through NIWC Pacific (N66001-19-24031), and the Allen Institute for AI. REFERENCES Jacob Austin, Augustus Odena, Maxwell Nye, Maarten Bosma, Henryk Michalewski, David Dohan, Ellen Jiang, Carrie J. Cai, Michael Terry, Quoc V . Le, and Charles Sutton. Program synthesis with large language models. ArXiv , abs/2108.07732, 2021. Sid Black, Leo Gao, Phil Wang, Connor Leahy, and Stella Biderman. GPT-Neo: Large Scale Autoregressive Language Modeling with Mesh-Tensorflow. Zenodo , March 2021. doi: 10.5281/ zenodo.5297715. URL https://doi.org/10.5281/zenodo.5297715 . If you use this software, please cite it using these metadata.

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

37

zenodo.5297715. URL https://doi.org/10.5281/zenodo.5297715 . If you use this software, please cite it using these metadata. Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-V oss, Gretchen Krueger, T. J. Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeff Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, 10 Preprint. Under review. Scott Gray, Benjamin Chess, Jack Clark, Christopher Berner, Sam McCandlish, Alec Radford, Ilya Sutskever, and Dario Amodei. Language models are few-shot learners. ArXiv , abs/2005.14165, 2020. Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Henrique Ponde de Oliveira Pinto, Jared Kaplan, Harri Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, Alex Ray, Raul Puri,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

38

Kaplan, Harri Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, Alex Ray, Raul Puri, Gretchen Krueger, Michael Petrov, Heidy Khlaaf, Girish Sastry, Pamela Mishkin, Brooke Chan, Scott Gray, Nick Ryder, Mikhail Pavlov, Alethea Power, Lukasz Kaiser, Mohammad Bavarian, Clemens Winter, Philippe Tillet, Felipe Petroski Such, Dave Cummings, Matthias Plappert, Fotios Chantzis, Elizabeth Barnes, Ariel Herbert-V oss, William Hebgen Guss, Alex Nichol, Alex Paino, Nikolas Tezak, Jie Tang, Igor Babuschkin, Suchir Balaji, Shantanu Jain, William Saunders, Christopher Hesse, Andrew N. Carr, Jan Leike, Josh Achiam, Vedant Misra, Evan Morikawa, Alec Radford, Matthew Knight, Miles Brundage, Mira Murati, Katie Mayer, Peter Welinder, Bob McGrew, Dario Amodei, Sam McCandlish, Ilya Sutskever, and Wojciech Zaremba. Evaluating large language models trained on code. arXiv , 2021. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

39

Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, Christopher Hesse, and John Schulman. Training verifiers to solve math word problems, 2021. URL https://arxiv. org/abs/2110.14168 . Sumanth Dathathri, Andrea Madotto, Janice Lan, Jane Hung, Eric Frank, Piero Molino, Jason Yosinski, and Rosanne Liu. Plug and play language models: A simple approach to controlled text generation. ArXiv , abs/1912.02164, 2020. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) , pp. 4171–4186, Minneapolis, Minnesota, June 2019. Association for Computational Linguistics. doi: 10.18653/v1/N19-1423. URL https: //aclanthology.org/N19-1423 .

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

40

//aclanthology.org/N19-1423 . David Dohan, Winnie Xu, Aitor Lewkowycz, Jacob Austin, David Bieber, Raphael Gontijo Lopes, Yuhuai Wu, Henryk Michalewski, Rif A. Saurous, Jascha Narain Sohl-Dickstein, Kevin Murphy, and Charles Sutton. Language model cascades. ArXiv , abs/2207.10342, 2022. Felix Faltings, Michel Galley, Gerold Hintz, Chris Brockett, Chris Quirk, Jianfeng Gao, and Bill Dolan. Text editing by command. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies , pp. 5259–5274, Online, June 2021. Association for Computational Linguistics. doi: 10.18653/v1/ 2021.naacl-main.414. URL https://aclanthology.org/2021.naacl-main.414 . Leo Gao, Stella Biderman, Sid Black, Laurence Golding, Travis Hoppe, Charles Foster, Jason Phang, Horace He, Anish Thite, Noa Nabeshima, Shawn Presser, and Connor Leahy. The pile:

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

41

Phang, Horace He, Anish Thite, Noa Nabeshima, Shawn Presser, and Connor Leahy. The pile: An 800gb dataset of diverse text for language modeling, 2021. URL https://arxiv.org/ abs/2101.00027 . Samuel Gehman, Suchin Gururangan, Maarten Sap, Yejin Choi, and Noah A. Smith. RealToxicityPrompts: Evaluating neural toxic degeneration in language models. In Findings of the Association for Computational Linguistics: EMNLP 2020 , pp. 3356–3369, Online, November 2020. Association for Computational Linguistics. doi: 10.18653/v1/2020.findings-emnlp.301. URL https://aclanthology.org/2020.findings-emnlp.301 . Suchin Gururangan, Ana Marasovi ´c, Swabha Swayamdipta, Kyle Lo, Iz Beltagy, Doug Downey, and Noah A. Smith. Don’t stop pretraining: Adapt language models to domains and tasks. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics , pp.

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

42

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics , pp. 8342–8360, Online, July 2020. Association for Computational Linguistics. doi: 10.18653/v1/ 2020.acl-main.740. URL https://aclanthology.org/2020.acl-main.740 . Mohammad Javad Hosseini, Hannaneh Hajishirzi, Oren Etzioni, and Nate Kushman. Learning to solve arithmetic word problems with verb categorization. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) , pp. 523–533, Doha, Qatar, 11 Preprint. Under review. October 2014. Association for Computational Linguistics. doi: 10.3115/v1/D14-1058. URL https://aclanthology.org/D14-1058 . Ben Krause, Akhilesh Deepak Gotmare, Bryan McCann, Nitish Shirish Keskar, Shafiq Joty, Richard Socher, and Nazneen Fatema Rajani. GeDi: Generative discriminator guided sequence generation. InFindings of the Association for Computational Linguistics: EMNLP 2021 , pp. 4929–4952,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

43

InFindings of the Association for Computational Linguistics: EMNLP 2021 , pp. 4929–4952, Punta Cana, Dominican Republic, November 2021. Association for Computational Linguistics. doi: 10.18653/v1/2021.findings-emnlp.424. URL https://aclanthology.org/2021. findings-emnlp.424 . Jason Lee, Raphael Shu, and Kyunghyun Cho. Iterative refinement in the continuous space for non-autoregressive neural machine translation. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) , pp. 1006–1015, Online, November 2020. Association for Computational Linguistics. doi: 10.18653/v1/2020.emnlp-main.73. URL https://aclanthology.org/2020.emnlp-main.73 . Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Veselin Stoyanov, and Luke Zettlemoyer. BART: Denoising sequence-to-sequence pretraining for natural language generation, translation, and comprehension. In Proceedings of

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031

2211.00053

44

the 58th Annual Meeting of the Association for Computational Linguistics , pp. 7871–7880, Online, July 2020. Association for Computational Linguistics. doi: 10.18653/v1/2020.acl-main.703. URLhttps://aclanthology.org/2020.acl-main.703 . Xiang Lisa Li and Percy Liang. Prefix-tuning: Optimizing continuous prompts for generation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) , pp. 4582–4597, Online, August 2021. Association for Computational Linguistics. doi: 10.18653/ v1/2021.acl-long.353. URL https://aclanthology.org/2021.acl-long.353 . Xiang Lisa Li, John Thickstun, Ishaan Gulrajani, Percy Liang, and Tatsunori Hashimoto. Diffusionlm improves controllable text generation. ArXiv , abs/2205.14217, 2022. Bill Yuchen Lin, Wangchunshu Zhou, Ming Shen, Pei Zhou, Chandra Bhagavatula, Yejin Choi,

2211.00053

Generating Sequences by Learning to Self-Correct

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator.

http://arxiv.org/pdf/2211.00053

cs.CL

20221031

20221031