most citedHarnessing Orthogonality to Train Low-Rank Neural Networks

1 citations · 1 across the 1 of their papers we have counts for

collaborators

1 paper

cs.LG2024★ 1 cited

Harnessing Orthogonality to Train Low-Rank Neural Networks

Daniel Coquelin, Katharina Flügel, Marie Weiel +4

This study explores the learning dynamics of neural networks by analyzing the singular value decomposition (SVD) of their weights throughout training. Our investigation reveals tha…