Active Learning
Last reviewed by Moderation API
Active learning is a training strategy where the model itself selects the most informative unlabeled examples for humans to annotate next, typically the cases it is least certain about. It reduces labeling cost by focusing human effort on the examples that will improve the model the most.
