AI & Machine Learning · Intermediate

Self-Supervised Learning (SSL)

Learning in which the training signal is generated automatically from the data itself, without human labels.

Definition

Self-supervised learning creates pretext tasks such as predicting a masked word or the next token, so the data supplies its own labels. It is how large language models and many vision encoders are pretrained on massive unlabeled corpora. The learned representations are then adapted to downstream tasks.

Related terms