Improving the Transferability of Adversarial Examples by Feature Augmentation

TL;DR

This paper introduces FAUG, a feature augmentation attack that enhances adversarial transferability by injecting noise into intermediate features, improving attack success across models without extra computation.

Contribution

The proposed FAUG method improves transferability of adversarial examples by feature augmentation, compatible with existing attacks, and validated on ImageNet with significant performance gains.

Findings

Achieves +26.22% transfer success on input transformation attacks

Achieves +5.57% transfer success on combined attacks

Effective across CNN and transformer models

Abstract

Despite the success of input transformation-based attacks on boosting adversarial transferability, the performance is unsatisfying due to the ignorance of the discrepancy across models. In this paper, we propose a simple but effective feature augmentation attack (FAUG) method, which improves adversarial transferability without introducing extra computation costs. Specifically, we inject the random noise into the intermediate features of the model to enlarge the diversity of the attack gradient, thereby mitigating the risk of overfitting to the specific model and notably amplifying adversarial transferability. Moreover, our method can be combined with existing gradient attacks to augment their performance further. Extensive experiments conducted on the ImageNet dataset across CNN and transformer models corroborate the efficacy of our method, e.g., we achieve improvement of +26.22% and +5.57% on input transformation-based attacks and combination methods, respectively.