Engineering the Neural Collapse Geometry of Supervised-Contrastive Loss

Supervised-contrastive loss (SCL) is an alternative to cross-entropy (CE) for classification tasks that makes use of similarities in the embedding space to allow for richer representations.

Data and Resources

Cite this as

Jaidev Gill, Vala Vakilian, Christos Thrampoulidis (2024). Dataset: Engineering the Neural Collapse Geometry of Supervised-Contrastive Loss. https://doi.org/10.57702/arr5v9yv

DOI retrieved: December 3, 2024

Additional Info

Field Value
Created December 3, 2024
Last update December 3, 2024
Defined In https://doi.org/10.48550/arXiv.2310.00893
Author Jaidev Gill
More Authors
Vala Vakilian
Christos Thrampoulidis
Homepage https://arxiv.org/abs/2303.12345