← 研踪 PaperTrailarXiv:2611.14594 · v1
Adversarial Risk in the Generative AI Era Necessitates Dropping the Small Epsilon Ball
首次提交 2026-11-10 18:16
版本更新 2026-11-10 18:16
雷达观测 2026-11-13 10:09
摘要
Adversarial Machine Learning research is currently centered on the search for the epsilon-ball: minimal perturbations designed to fool models while remaining imperceptible to humans. We argue that this fixation on norms is a metrological trap that communicates a distorted perception of risk to defenders, one that is unaligned with real-world system vulnerabilities. This paradigm misleads defenders into optimizing for narrow, mathematical robustness at the expense of systemic security, while simultaneously incentivizing attackers to exploit non-overlapping alternative adversarial pathways. Ultimately, we argue that the dominance of the ball does not just fail to reduce harm---it may actively facilitate it by communicating a false sense of security against the wrong threats.
作者备注:neurips 2026, 10 pages, position paper on adversarial risk, certified robustness
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-10 18:16 · 最新观测版本
Adversarial Risk in the Generative AI Era Necessitates Dropping the Small Epsilon Ball
Adversarial Machine Learning research is currently centered on the search for the epsilon-ball: minimal perturbations designed to fool models while remaining imperceptible to humans. We argue that this fixation on norms is a metrological trap that communicates a distorted perception of risk to defenders, one that is unaligned with real-world system vulnerabilities. This paradigm misleads defenders into optimizing for narrow, mathematical robustness at the expense of systemic security, while simultaneously incentivizing attackers to exploit non-overlapping alternative adversarial pathways. Ultimately, we argue that the dominance of the ball does not just fail to reduce harm---it may actively facilitate it by communicating a false sense of security against the wrong threats.
查看此版本解读 →查看此版本原文 ↗