Frequency domain regularization for iterative adversarial attacks

Adversarial examples have attracted more and more attentions with the prosperity of convolutional neural networks. The transferability of adversarial examples is an important property that makes black-box attacks possible in real-world applications. On the other side, many adversarial defense methods have been proposed to improve the robustness, leading to the requirement for more transferable adversarial examples. Inspired by the regularization term for network parameters at training process, we treat adversarial attacks as training process of inputs and propose regularization constraint for inputs to prevent adversarial examples from overfitting the white-box networks and enhance the transferability. Specifically, we find a universal attribute that the outputs of convolutional neural networks have consistency to the low frequencies of inputs, and based on this, we construct a frequency domain regularization to inputs for iterative attacks. In this way, our method is compatible with existing iterative attack methods and can learn more transferable adversarial examples. Extensive experiments on ImageNet validate the superiority of our method, and compared with several attacks, we achieve attack success rate improvements of 8.0% and 11.5% on average to normal models and defense methods respectively. (c) 2022 Published by Elsevier Ltd.

Automated summary

Adversarial examples have gained increasing attention and the transferability of such examples is crucial for black-box attacks. To enhance transferability and prevent overfitting, this study proposes a regularization constraint for inputs in adversarial attacks. By exploiting the consistency between the outputs of convolutional neural networks and low frequencies of inputs, a frequency domain regularization is constructed. Experimental results on ImageNet demonstrate the superiority of the proposed method, achieving significant improvements in attack success rate compared to other attacks and defense methods.