| 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. |