Fast Adversarial Training with Adaptive Step Size

TL;DR

This paper introduces ATAS, an adaptive step size method for adversarial training that mitigates catastrophic overfitting and improves robustness on multiple datasets by adjusting step sizes based on instance gradient norms.

Contribution

The paper proposes a novel adaptive step size approach for adversarial training that is theoretically faster and empirically more robust than existing methods.

Findings

ATAS reduces catastrophic overfitting across datasets.

ATAS achieves higher robust accuracy on CIFAR10, CIFAR100, and ImageNet.

Theoretical analysis confirms faster convergence of ATAS.

Abstract

While adversarial training and its variants have shown to be the most effective algorithms to defend against adversarial attacks, their extremely slow training process makes it hard to scale to large datasets like ImageNet. The key idea of recent works to accelerate adversarial training is to substitute multi-step attacks (e.g., PGD) with single-step attacks (e.g., FGSM). However, these single-step methods suffer from catastrophic overfitting, where the accuracy against PGD attack suddenly drops to nearly 0% during training, destroying the robustness of the networks. In this work, we study the phenomenon from the perspective of training instances. We show that catastrophic overfitting is instance-dependent and fitting instances with larger gradient norm is more likely to cause catastrophic overfitting. Based on our findings, we propose a simple but effective method, Adversarial Training with Adaptive Step size (ATAS). ATAS learns an instancewise adaptive step size that is inversely proportional to its gradient norm. The theoretical analysis shows that ATAS converges faster than the commonly adopted non-adaptive counterparts. Empirically, ATAS consistently mitigates catastrophic overfitting and achieves higher robust accuracy on CIFAR10, CIFAR100 and ImageNet when evaluated on various adversarial budgets.