English(EN) Convergence of Sharpness-Aware Minimization Algorithms using Increasing Batch Size and Decaying Learning Rate

新研究详解锐度感知最小化算法的收敛性

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-07 04:00

一篇新研究论文探讨了锐度感知最小化（SAM）算法的收敛性，特别是其差距引导SAM（GSAM）变体。该研究从理论上证明，采用增大批量大小或衰减学习率（如余弦退火或线性衰减）可以实现收敛。数值比较表明，与使用恒定批量大小和学习率相比，增大批量大小的GSAM在最坏情况下的自适应锐度更低。 AI

影响为深度神经网络训练优化提供了理论和数值见解，可能提高泛化能力。

排序理由在arXiv上发表的研究论文，详细介绍了优化算法的理论和数值发现。[lever_c_demoted from research: ic=1 ai=1.0]

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 7 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 在arXiv上发表的研究论文，详细介绍了优化算法的理论和数值发现。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, other; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

arXiv cs.LG TIER_1 English(EN) · Hinata Harada, Hideaki Iiduka · 2026-10-07 04:00

使用增大批量大小和衰减学习率的锐度感知最小化算法的收敛性

arXiv:2409.09984v2 Announce Type: replace Abstract: The sharpness-aware minimization (SAM) algorithm and its variants, including gap guided SAM (GSAM), have been successful at improving the generalization capability of deep neural network models by finding flat local minima of th…