Weight-feature equality conjecture at full-rank stationary points

From papers

Let W\boldsymbol{W} and H\boldsymbol{H} be the classifier and feature matrices in the unconstrained feature model. Assume that both matrices have full row rank and satisfy the first-order optimality condition referred to in the source.

Weight-feature equality conjecture. Then

W=H.\boldsymbol{W}=\boldsymbol{H}.

This conjecture links first-order optimality with equality of the classifier and feature matrices. Its resolution is not specified in the source.

Progress summary

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Sources & referencesView supporting material

Primary source

Jiachen Jiang, Jinxin Zhou, Peng Wang, Qing Qu, Dustin Mixon, Chong You and Zhihui Zhu, “Generalized Neural Collapse for a Large Number of Classes”, arXiv:2310.05351 (2023).

Solutions 0

No solutions have been posted yet.