Contrastive Learning

What is Contrastive Learning?

A self-supervised learning technique where a model learns representations by pulling positive pairs (e.g., two different crops of the same image) closer together in the embedding space and pushing negative pairs (e.g., different images) apart.

Where did the term "Contrastive Learning" come from?

Popularized by frameworks like SimCLR (Google) and MoCo (Facebook AI Research).

How is "Contrastive Learning" used today?

The basis for powerful unsupervised vision models (like CLIP) and modern text embedding models.

Related Terms