Efficient Adversarial Training with Transferable Adversarial Examples
This work addresses the efficiency bottleneck in adversarial training for machine learning practitioners, offering a novel approach that is incremental but provides substantial speed-ups and robustness gains.
The paper tackles the high computational cost of adversarial training by introducing ATTA, a method that leverages transferable adversarial examples across epochs to improve efficiency, achieving up to 7.2% higher adversarial accuracy on CIFAR10 and reducing training time by 12-14x on MNIST and CIFAR10.
Adversarial training is an effective defense method to protect classification models against adversarial attacks. However, one limitation of this approach is that it can require orders of magnitude additional training time due to high cost of generating strong adversarial examples during training. In this paper, we first show that there is high transferability between models from neighboring epochs in the same training process, i.e., adversarial examples from one epoch continue to be adversarial in subsequent epochs. Leveraging this property, we propose a novel method, Adversarial Training with Transferable Adversarial Examples (ATTA), that can enhance the robustness of trained models and greatly improve the training efficiency by accumulating adversarial perturbations through epochs. Compared to state-of-the-art adversarial training methods, ATTA enhances adversarial accuracy by up to 7.2% on CIFAR10 and requires 12~14x less training time on MNIST and CIFAR10 datasets with comparable model robustness.