semi-supervised learning
A broad category of machine learning training techniques that leverages a small amount of labeled data with a larger quantity of unlabeled data to improve the accuracy of a model, unlike supervised learning and unsupervised learning.