English(EN) One Size Does Not Fit All: Curvature-Aware Adaptive Perturbation for SAM

新的CAP-SAM方法增强了神经网络的泛化能力

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-15 04:00

研究人员引入了一种名为曲率感知自适应扰动用于SAM（CAP-SAM）的新方法，以提高深度神经网络的泛化能力。该技术通过考虑损失景观的局部曲率来修改Sharpness-Aware Minimization（SAM）中使用的扰动半径。通过自适应地调整扰动，CAP-SAM旨在更有效地找到鲁棒的最小值，在CIFAR-10、CIFAR-100和ImageNet等基准数据集上优于标准的SAM及其变体。 AI

影响增强了深度神经网络的泛化能力，有望在各种应用中实现更鲁棒的AI模型。

排序理由研究论文，详细介绍了提高深度神经网络泛化能力的新方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

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: 14 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 研究论文，详细介绍了提高深度神经网络泛化能力的新方法。[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, model release; 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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.LG TIER_1 English(EN) · Alessia Rossi, Felipe Campelo, Gabriel L. Oliveira · 2026-10-15 04:00

一刀切不可取：面向SAM的感知曲率自适应扰动

arXiv:2610.20415v1 Announce Type: new Abstract: Poor generalisation in deep neural networks is frequently associated with convergence to sharp minima. Sharpness-Aware Minimisation (SAM) improves generalisation of deep neural networks by minimising a combination of the training lo…

本文如何被排名

报道来源 [1]

一刀切不可取：面向SAM的感知曲率自适应扰动

相关实体

相关话题