ML technique where a model learns to recognize objects, tasks, or concepts it has never seen during training.
Generality: 660
