Fixing Data Augmentation to Improve Adversarial Robustness
This addresses the problem of robust overfitting in adversarial training for machine learning practitioners, offering incremental improvements over prior methods.
The paper tackles robust overfitting in adversarial training by improving data augmentation methods, showing that combining model weight averaging with augmentation boosts robust accuracy and using generative models to expand training data further enhances robustness. On CIFAR-10, it achieves absolute improvements of +7.06% and +5.88% in robust accuracy against specific perturbations, reaching 64.20% robust accuracy without external data.
Adversarial training suffers from robust overfitting, a phenomenon where the robust test accuracy starts to decrease during training. In this paper, we focus on both heuristics-driven and data-driven augmentations as a means to reduce robust overfitting. First, we demonstrate that, contrary to previous findings, when combined with model weight averaging, data augmentation can significantly boost robust accuracy. Second, we explore how state-of-the-art generative models can be leveraged to artificially increase the size of the training set and further improve adversarial robustness. Finally, we evaluate our approach on CIFAR-10 against $\ell_\infty$ and $\ell_2$ norm-bounded perturbations of size $ε= 8/255$ and $ε= 128/255$, respectively. We show large absolute improvements of +7.06% and +5.88% in robust accuracy compared to previous state-of-the-art methods. In particular, against $\ell_\infty$ norm-bounded perturbations of size $ε= 8/255$, our model reaches 64.20% robust accuracy without using any external data, beating most prior works that use external data.