Jul 2026
Contrastive-Collapsed Loss for Flexible and Geometrically Optimal Embeddings and Faster Convergence
A theoretical analysis positioning CoCo with respect to related objectives shows that the new proposed loss benefits from closer initialization to the optimal configuration, more informative gradients, and stronger incentives for class-wise representation collapse.
Blanca Cano-Camarero, 'Angela Fern'andez-Pascual, José R. Dorronsoro
· arXiv.org · 0 citations