arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2009.11397cs.LGcs.CRcs.CVstat.ML
收藏

基于反攻击的迭代对抗攻击检测

Detection of Iterative Adversarial Attacks via Counter Attack

University of Wuppertal(伍珀塔尔大学);ENSEEIHT Toulouse(图卢兹ENSEEIHT学院);University of Novi Sad(诺威萨德大学)

Matthias Rottmann, Kira Maag, Mathis Peyron, Natasa Krejic, Hanno Gottschalk

摘要:

深度神经网络(DNN)已被证明是处理非结构化数据的强大工具。然而,对于高维数据,如图像,它们天生容易受到对抗攻击。添加到输入中的小的几乎不可见的扰动可以用来欺骗DNN。近年来,已经引入了各种攻击、加固方法和检测方法。众所周知,通过迭代最小化计算的Carlini-Wagner(CW)类型攻击属于最难检测的攻击。在这项工作中,我们概述了一个数学证明,CW攻击本身可以用作检测器。也就是说,在某些假设下,在攻击迭代的限制下,该检测器提供了原始图像和攻击图像的渐近最优分离。在数值实验中,我们通过实验验证了这一说法,并在CIFAR 10和ImageNet上获得了高达99.73%的AUROC值。这是目前最先进的CW攻击检测率的上半部分。

英文摘要:

Deep neural networks (DNNs) have proven to be powerful tools for processing unstructured data. However for high-dimensional data, like images, they are inherently vulnerable to adversarial attacks. Small almost invisible perturbations added to the input can be used to fool DNNs. Various attacks, hardening methods and detection methods have been introduced in recent years. Notoriously, Carlini-Wagner (CW) type attacks computed by iterative minimization belong to those that are most difficult to detect. In this work we outline a mathematical proof that the CW attack can be used as a detector itself. That is, under certain assumptions and in the limit of attack iterations this detector provides asymptotically optimal separation of original and attacked images. In numerical experiments, we experimentally validate this statement and furthermore obtain AUROC values up to 99.73% on CIFAR10 and ImageNet. This is in the upper part of the spectrum of current state-of-the-art detection rates for CW attacks.

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。

↑