arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.04025cs.LG
收藏

WAT:提高对抗训练中的最差鲁棒性

WAT: Improve the Worst-class Robustness in Adversarial Training

School of Computer Science, Wuhan University(武汉大学计算机学院)

Boqi Li, Weiwei Liu

摘要:

深度神经网络(DNN)已被证明容易受到对抗性示例的影响。对抗性训练(AT)是防御对抗攻击的一种流行而有效的策略。最近的作品(Benz等人,2020年; Xu等人,2021; Tian等人,2021)已经表明,通过AT良好训练的鲁棒模型在类别之间表现出显着的鲁棒性差异,并提出了各种方法来获得跨类别的一致的鲁棒精度。不幸的是,这些方法牺牲了大量的平均鲁棒精度。因此,本文提出了一种新的最差类对抗训练框架,并利用无悔动力学来解决这个问题。我们的目标是获得一个分类器具有很大的性能在最坏的类和牺牲只是一个小的平均鲁棒精度在同一时间。然后,我们严格分析了我们提出的算法的理论特性,和泛化误差界的最坏类的鲁棒风险。此外,我们提出了一个测量来评估所提出的方法的平均和最差的类的准确性。在各种数据集和网络上的实验表明,我们提出的方法优于最先进的方法。

英文摘要:

Deep Neural Networks (DNN) have been shown to be vulnerable to adversarial examples. Adversarial training (AT) is a popular and effective strategy to defend against adversarial attacks. Recent works (Benz et al., 2020; Xu et al., 2021; Tian et al., 2021) have shown that a robust model well-trained by AT exhibits a remarkable robustness disparity among classes, and propose various methods to obtain consistent robust accuracy across classes. Unfortunately, these methods sacrifice a good deal of the average robust accuracy. Accordingly, this paper proposes a novel framework of worst-class adversarial training and leverages no-regret dynamics to solve this problem. Our goal is to obtain a classifier with great performance on worst-class and sacrifice just a little average robust accuracy at the same time. We then rigorously analyze the theoretical properties of our proposed algorithm, and the generalization error bound in terms of the worst-class robust risk. Furthermore, we propose a measurement to evaluate the proposed method in terms of both the average and worst-class accuracies. Experiments on various datasets and networks show that our proposed method outperforms the state-of-the-art approaches.

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。

↑