Current browse context:
cs.LG
Change to browse by:
References & Citations
Computer Science > Machine Learning
Title: Neural networks learn to magnify areas near decision boundaries
(Submitted on 26 Jan 2023 (this version), latest version 14 Oct 2023 (v3))
Abstract: We study how training molds the Riemannian geometry induced by neural network feature maps. At infinite width, neural networks with random parameters induce highly symmetric metrics on input space. Feature learning in networks trained to perform classification tasks magnifies local areas along decision boundaries. These changes are consistent with previously proposed geometric approaches for hand-tuning of kernel methods to improve generalization.
Submission history
From: Jacob Zavatone-Veth [view email][v1] Thu, 26 Jan 2023 19:43:16 GMT (20018kb,D)
[v2] Wed, 17 May 2023 17:07:01 GMT (18656kb,D)
[v3] Sat, 14 Oct 2023 22:52:45 GMT (22814kb,D)
Link back to: arXiv, form interface, contact.