AI RESEARCH

Neural Collapse by Design: Learning Class Prototypes on the Hypersphere

arXiv CS.LG

ArXi:2605.20302v1 Announce Type: new Supervised classification has a theoretical optimum, Neural Collapse (NC), yet neither of its two dominant paradigms reaches it in practice. Cross entropy (CE) leaves radial degrees of freedom unconstrained and converges to a degenerate geometry, while supervised contrastive learning (SCL) drives features toward NC during pre