Self-Supervised Learning

What is Self-Supervised Learning?

Self-supervised learning is a training method where the model learns from unlabeled data by generating its own labels, such as predicting the next word in a sentence or filling in missing parts of an input.

Where did the term "Self-Supervised Learning" come from?

A key driver behind the success of modern NLP and computer vision models.

How is "Self-Supervised Learning" used today?

Enables training on internet-scale datasets without the need for expensive human annotation.

Related Terms