MLLGFeb 19, 2023

mSAM: Micro-Batch-Averaged Sharpness-Aware Minimization

arXiv:2302.09693v212 citationsh-index: 38
AI Analysis

This work addresses generalization issues in deep learning for practitioners, offering an incremental improvement over SAM with flexible implementation.

The paper tackles the problem of improving generalization in over-parameterized deep learning models by proposing mSAM, a variant of Sharpness-Aware Minimization that aggregates adversarial perturbations across micro-batches, achieving flatter minima and superior generalization performance compared to SAM across various image classification and NLP tasks.

Modern deep learning models are over-parameterized, where different optima can result in widely varying generalization performance. The Sharpness-Aware Minimization (SAM) technique modifies the fundamental loss function that steers gradient descent methods toward flatter minima, which are believed to exhibit enhanced generalization prowess. Our study delves into a specific variant of SAM known as micro-batch SAM (mSAM). This variation involves aggregating updates derived from adversarial perturbations across multiple shards (micro-batches) of a mini-batch during training. We extend a recently developed and well-studied general framework for flatness analysis to theoretically show that SAM achieves flatter minima than SGD, and mSAM achieves even flatter minima than SAM. We provide a thorough empirical evaluation of various image classification and natural language processing tasks to substantiate this theoretical advancement. We also show that contrary to previous work, mSAM can be implemented in a flexible and parallelizable manner without significantly increasing computational costs. Our implementation of mSAM yields superior generalization performance across a wide range of tasks compared to SAM, further supporting our theoretical framework.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes