Saved in:
Bibliographic Details
Main Authors: Tiwary, Nalin, Aananth, Siddarth
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.01714
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916466180227072
author Tiwary, Nalin
Aananth, Siddarth
author_facet Tiwary, Nalin
Aananth, Siddarth
contents Sharpness-Aware Minimization (SAM) is an optimization technique designed to improve generalization by favoring flatter loss minima. To achieve this, SAM optimizes a modified objective that penalizes sharpness, using computationally efficient approximations. Interestingly, we find that more precise approximations of the proposed SAM objective degrade generalization performance, suggesting that the generalization benefits of SAM are rooted in these approximations rather than in the original intended mechanism. This highlights a gap in our understanding of SAM's effectiveness and calls for further investigation into the role of approximations in optimization.
format Preprint
id arxiv_https___arxiv_org_abs_2411_01714
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle 1st-Order Magic: Analysis of Sharpness-Aware Minimization
Tiwary, Nalin
Aananth, Siddarth
Machine Learning
Sharpness-Aware Minimization (SAM) is an optimization technique designed to improve generalization by favoring flatter loss minima. To achieve this, SAM optimizes a modified objective that penalizes sharpness, using computationally efficient approximations. Interestingly, we find that more precise approximations of the proposed SAM objective degrade generalization performance, suggesting that the generalization benefits of SAM are rooted in these approximations rather than in the original intended mechanism. This highlights a gap in our understanding of SAM's effectiveness and calls for further investigation into the role of approximations in optimization.
title 1st-Order Magic: Analysis of Sharpness-Aware Minimization
topic Machine Learning
url https://arxiv.org/abs/2411.01714