arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2405.01349cs.LGcs.CR
收藏

立场:迈向对抗样本的抗毁性研究

Position: Towards Resilience Against Adversarial Examples

Princeton University(普林斯顿大学)

Sihui Dai, Chong Xiang, Tong Wu, Prateek Mittal

AI总结:

该立场论文针对现有对抗样本防御仅适配单一攻击类型、易被未覆盖攻击绕过的问题,提出对抗抗毁性概念与定义,引入其子问题持续自适应鲁棒性,梳理其与已有相关研究的关联并给出开放方向。

摘要:

当前对抗样本防御研究主要聚焦于实现对单一攻击类型(如$\boldsymbol{\text{ℓ}}_2$或$\boldsymbol{\text{ℓ}}_{\boldsymbol{\text{∞}}}$界攻击)的鲁棒性。然而,可能的扰动空间远大于现有多数防御所考虑的范围,且难以进行数学建模,因此攻击者可通过使用防御未覆盖的攻击类型轻松绕过防御。在这篇立场论文中,我们提出除了鲁棒性之外,还应致力于开发具备对抗抗毁性(adversarially resilient)的防御算法——这类防御算法需提供一种机制,能快速调整受保护模型,使其对新攻击具备鲁棒性。我们给出了对抗抗毁性的定义,并概述了设计对抗抗毁性防御的考量因素。随后我们引入了对抗抗毁性的一个子问题,称之为持续自适应鲁棒性(continual adaptive robustness):防御者会随时间逐步获取可能扰动空间的构造知识,并可基于这些信息更新模型。此外,我们阐明了持续自适应鲁棒性与此前研究的多攻击鲁棒性、未预见攻击鲁棒性问题之间的关联,并概述了这些领域中有助于提升持续自适应鲁棒性与对抗抗毁性的开放研究方向。

英文摘要:

Current research on defending against adversarial examples focuses primarily on achieving robustness against a single attack type such as $\ell_2$ or $\ell_{\infty}$-bounded attacks. However, the space of possible perturbations is much larger than considered by many existing defenses and is difficult to mathematically model, so the attacker can easily bypass the defense by using a type of attack that is not covered by the defense. In this position paper, we argue that in addition to robustness, we should also aim to develop defense algorithms that are adversarially resilient -- defense algorithms should specify a means to quickly adapt the defended model to be robust against new attacks. We provide a definition of adversarial resilience and outline considerations of designing an adversarially resilient defense. We then introduce a subproblem of adversarial resilience which we call continual adaptive robustness, in which the defender gains knowledge of the formulation of possible perturbation spaces over time and can then update their model based on this information. Additionally, we demonstrate the connection between continual adaptive robustness and previously studied problems of multiattack robustness and unforeseen attack robustness and outline open directions within these fields which can contribute to improving continual adaptive robustness and adversarial resilience.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑