# Asynchronous Sharpness-Aware Minimization For Fast and Accurate Deep Learning

arXiv: https://arxiv.org/abs/2503.11147 · HTML: https://mlsanity.com/paper/2503.11147/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 34 references that are in the map.
- First posted on arXiv 14 Mar 2025.

## Builds on (most cited first, 23 of 23)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Wide Residual Networks](https://mlsanity.com/paper/1605.07146/) (2016)
- [Sharpness-Aware Minimization for Efficiently Improving Generalization](https://mlsanity.com/paper/2010.01412/) (2020)
- [Visualizing the Loss Landscape of Neural Nets](https://mlsanity.com/paper/1712.09913/) (2017)
- [Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition](https://mlsanity.com/paper/1804.03209/) (2018)
- [SGD: General Analysis and Improved Rates](https://mlsanity.com/paper/1901.09401/) (2019)
- [Asynchronous Decentralized Parallel Stochastic Gradient Descent](https://mlsanity.com/paper/1710.06952/) (2017)
- [Generalized Federated Learning via Sharpness Aware Minimization](https://mlsanity.com/paper/2206.02618/) (2022)
- [Revisiting Distributed Synchronous SGD](https://mlsanity.com/paper/1702.05800/) (2017)
- [A Bayesian Perspective on Generalization and Stochastic Gradient Descent](https://mlsanity.com/paper/1710.06451/) (2017)
- [Towards Efficient and Scalable Sharpness-Aware Minimization](https://mlsanity.com/paper/2203.02714/) (2022)
- [Penalizing Gradient Norm for Efficiently Improving Generalization in Deep Learning](https://mlsanity.com/paper/2202.03599/) (2022)
- [Asynchronous Parallel Stochastic Gradient for Nonconvex Optimization](https://mlsanity.com/paper/1506.08272/) (2015)
- [Asynchronous Stochastic Gradient Descent with Delay Compensation](https://mlsanity.com/paper/1609.08326/) (2016)
- [An Adaptive Policy to Employ Sharpness-Aware Minimization](https://mlsanity.com/paper/2304.14647/) (2023)
- [No One Left Behind: Inclusive Federated Learning over Heterogeneous Devices](https://mlsanity.com/paper/2202.08036/) (2022)
- [Distributed Hybrid CPU and GPU training for Graph Neural Networks on Billion-Scale Graphs](https://mlsanity.com/paper/2112.15345/) (2021)
- [Normalization Layers Are All That Sharpness-Aware Minimization Needs](https://mlsanity.com/paper/2306.04226/) (2023)
- [Recycling Model Updates in Federated Learning: Are Gradient Subspaces Low-Rank?](https://mlsanity.com/paper/2202.00280/) (2022)
- [DISTREAL: Distributed Resource-Aware Learning in Heterogeneous Systems](https://mlsanity.com/paper/2112.08761/) (2021)
- [Taming Convergence for Asynchronous Stochastic Gradient Descent with Unbounded Delay in Non-Convex Learning](https://mlsanity.com/paper/1805.09470/) (2018)
- [Embracing Federated Learning: Enabling Weak Client Participation via Partial Model Training](https://mlsanity.com/paper/2406.15125/) (2024)

## Built on by (most cited first, 2 of 2)

- [Evolution of Optimization Methods: Algorithms, Scenarios, and Evaluations](https://mlsanity.com/paper/2604.12968/) (2026)
- [LSAM: Asynchronous Distributed Training with Landscape-Smoothed Sharpness-Aware Minimization](https://mlsanity.com/paper/2509.03110/) (2025)

## Abstract (arXiv)

Sharpness-Aware Minimization (SAM) is an optimization method that improves the generalization performance of machine learning models. Despite its effectiveness, it has not been widely adopted in real-world applications due to its high computational cost. Existing approaches reduce this overhead by approximating or reusing gradients in the perturbation step, but still assume sequential perturbation and update steps. In this study, we introduce AsyncSAM, a variant of SAM that enables concurrent model perturbation and parameter updates across multiple compute resources. By adjusting the perturbation batch size in a system-aware manner, our asynchronous scheme fully overlaps perturbation with updates, eliminating the overhead and allowing even CPUs to perform perturbations without extra time. We also provide theoretical analysis showing that AsyncSAM achieves the same order of convergence as conventional SAM. Extensive experiments demonstrate that AsyncSAM delivers superior generalization performance and the highest training throughput among state-of-the-art SAM variants. For example, in ViT fine-tuning on CIFAR-100, it achieves accuracy comparable to SAM while maintaining training time nearly identical to SGD.