English(EN) Does More Compute Make Every Adversarial Attack Transferable?

研究：计算差距决定对抗性攻击的可迁移性

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-23 04:00

一篇新的研究论文探讨了计算资源与机器学习中对抗性攻击的可迁移性之间的关系。研究表明，攻击者和防御者之间显著的计算差距，特别是二次方差异，可能导致分类任务中普遍可迁移的攻击或鲁棒的防御。然而，在此计算阈值以下的生成任务中，这种二分法不成立。该研究还为增加推理时间计算可以增强模型鲁棒性的观点提供了理论支持。 AI

影响这项研究可能通过强调计算资源在对抗性鲁棒性中的关键作用，为开发更鲁棒的AI防御提供信息。

排序理由该条目是发表在arXiv上的研究论文，详细介绍了机器学习中对抗性攻击的理论发现。[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: 4 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该条目是发表在arXiv上的研究论文，详细介绍了机器学习中对抗性攻击的理论发现。[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, safety; 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。

报道来源 [1]

arXiv cs.LG TIER_1 English(EN) · Greg Gluch, Shafi Goldwasser, Anastasiia Struss · 2026-10-23 04:00

Does More Compute Make Every Adversarial Attack Transferable?

arXiv:2504.20310v3 Announce Type: replace Abstract: Transferable attacks, or transferable jailbreaks, are a particularly strong version of adversarial examples that fool all defenders in a class. In this work, we study when more compute enables transferable attacks. We show that,…

本文如何被排名

报道来源 [1]

Does More Compute Make Every Adversarial Attack Transferable?

相关实体

相关话题