1 result for “representation learning”
A machine learning training paradigm in which a model generates its own supervisory signal from unlabelled data by solving pretext tasks, learning rich representations without human-annotated labels.