Towards Understanding the Role of Sharpness-Aware Minimization Algorithms for Out-of-Distribution Generalization

TL;DR

This paper investigates the effectiveness of Sharpness-Aware Minimization (SAM) algorithms for out-of-distribution generalization, providing empirical comparisons, theoretical bounds, and insights into their performance in various OOD settings.

Contribution

It offers a comprehensive comparison of SAM variants for OOD generalization, introduces theoretical bounds, and explores SAM's role in gradual domain adaptation.

Findings

Original SAM outperforms Adam baseline by 4.76% in zero-shot OOD

Strongest SAM variants outperform Adam by 8.01% in zero-shot OOD

SAM outperforms Adam by 0.82% on average in GDA setting

Abstract

Recently, sharpness-aware minimization (SAM) has emerged as a promising method to improve generalization by minimizing sharpness, which is known to correlate well with generalization ability. Since the original proposal of SAM, many variants of SAM have been proposed to improve its accuracy and efficiency, but comparisons have mainly been restricted to the i.i.d. setting. In this paper we study SAM for out-of-distribution (OOD) generalization. First, we perform a comprehensive comparison of eight SAM variants on zero-shot OOD generalization, finding that the original SAM outperforms the Adam baseline by $4.76\%$ and the strongest SAM variants outperform the Adam baseline by $8.01\%$ on average. We then provide an OOD generalization bound in terms of sharpness for this setting. Next, we extend our study of SAM to the related setting of gradual domain adaptation (GDA), another form of OOD generalization where intermediate domains are constructed between the source and target domains, and iterative self-training is done on intermediate domains, to improve the overall target domain error. In this setting, our experimental results demonstrate that the original SAM outperforms the baseline of Adam on each of the experimental datasets by $0.82\%$ on average and the strongest SAM variants outperform Adam by $1.52\%$ on average. We then provide a generalization bound for SAM in the GDA setting. Asymptotically, this generalization bound is no better than the one for self-training in the literature of GDA. This highlights a further disconnection between the theoretical justification for SAM versus its empirical performance, with recent work finding that low sharpness alone does not account for all of SAM's generalization benefits. For future work, we provide several potential avenues for obtaining a tighter analysis for SAM in the OOD setting.