GLOSSARY TERM
What is Contrastive Learning?
A framework that learns representations by contrasting positive and negative pairs.
Contrastive learning pulls similar examples closer in the latent space while pushing dissimilar ones apart. It is highly effective for self-supervised representation learning, enabling robust feature extraction without extensive manual labeling.
Robust Representations
Train highly discriminative models on proprietary data securely.