1 result for “zero-shot-learning”
Zero-shot learning is a machine learning paradigm in which a model makes accurate predictions on categories it has never seen during training by leveraging semantic descriptions or attribute representations.