论文标题
通过局部线性化稳定有效的对抗训练
Stable and Efficient Adversarial Training through Local Linearization
论文作者
论文摘要
最近的单步对抗训练激增，因为它表现出稳健性和效率。 However, a phenomenon referred to as ``catastrophic overfitting" has been observed, which is prevalent in single-step defenses and may frustrate attempts to use FGSM adversarial training. To address this issue, we propose a novel method, Stable and Efficient Adversarial Training (SEAT), which mitigates catastrophic overfitting by harnessing on local properties that distinguish a robust model from that of a灾难性的模型具有强大的理论理由，使座椅损失最小化可以表现出平滑的经验风险，从而导致稳健性。在强烈的PGD-50攻击下，半径$ 8/255 $的扰动与仅3％的计算成本的10步迭代对抗训练的性能匹配。
There has been a recent surge in single-step adversarial training as it shows robustness and efficiency. However, a phenomenon referred to as ``catastrophic overfitting" has been observed, which is prevalent in single-step defenses and may frustrate attempts to use FGSM adversarial training. To address this issue, we propose a novel method, Stable and Efficient Adversarial Training (SEAT), which mitigates catastrophic overfitting by harnessing on local properties that distinguish a robust model from that of a catastrophic overfitted model. The proposed SEAT has strong theoretical justifications, in that minimizing the SEAT loss can be shown to favour smooth empirical risk, thereby leading to robustness. Experimental results demonstrate that the proposed method successfully mitigates catastrophic overfitting, yielding superior performance amongst efficient defenses. Our single-step method can reach 51% robust accuracy for CIFAR-10 with $l_\infty$ perturbations of radius $8/255$ under a strong PGD-50 attack, matching the performance of a 10-step iterative adversarial training at merely 3% computational cost.
