arXivDaily arXiv每日学术速递 周一至周五更新
看起来今天可以划水了
arXiv 2102.12002cs.LGcs.AIcs.CRstat.ML
收藏

非一致扰动下的对抗鲁棒性

Adversarial Robustness with Non-uniform Perturbations

Imperial College London(帝国理工学院);Amazon Web Services(亚马逊云服务)

Ecenaz Erdemir, Jeffrey Bickford, Luca Melis, Sergul Aydore

摘要:

机器学习模型的鲁棒性对于与安全相关的应用至关重要,其中现实世界的对手唯一地专注于逃避基于神经网络的检测器。以前的工作主要集中在制作对抗性示例(AE),这些示例具有跨特征的小的均匀范数有界扰动,以保持不可感知性的要求。然而,均匀扰动不会在诸如恶意软件、金融和社交网络等领域中产生真实的AE。对于这些类型的应用程序,功能通常具有一些语义上有意义的依赖关系。我们提出的方法的关键思想是启用非均匀扰动,可以在对抗训练期间充分表示这些特征依赖性。我们建议使用经验数据分布的特征,无论是在特征之间的相关性和特征本身的重要性。使用恶意软件分类,信用风险预测和垃圾邮件检测的实验数据集,我们表明,我们的方法是更强大的现实世界的攻击。最后,我们提出了利用非均匀扰动界的鲁棒性认证,并表明非均匀界实现更好的认证。

英文摘要:

Robustness of machine learning models is critical for security related applications, where real-world adversaries are uniquely focused on evading neural network based detectors. Prior work mainly focus on crafting adversarial examples (AEs) with small uniform norm-bounded perturbations across features to maintain the requirement of imperceptibility. However, uniform perturbations do not result in realistic AEs in domains such as malware, finance, and social networks. For these types of applications, features typically have some semantically meaningful dependencies. The key idea of our proposed approach is to enable non-uniform perturbations that can adequately represent these feature dependencies during adversarial training. We propose using characteristics of the empirical data distribution, both on correlations between the features and the importance of the features themselves. Using experimental datasets for malware classification, credit risk prediction, and spam detection, we show that our approach is more robust to real-world attacks. Finally, we present robustness certification utilizing non-uniform perturbation bounds, and show that non-uniform bounds achieve better certification.

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。

↑