Data Augmentation Can Improve Robustness
share
Adversarial training suffers from robust overfitting, a phenomenon where the robust test accuracy starts to decrease during training. In this paper, we focus on reducing robust overfitting by using common data augmentation schemes. We demonstrate that, contrary to previous findings, when combined with model weight averaging, data augmentation can significantly boost robust accuracy. Furthermore, we compare various augmentations techniques and observe that spatial composition techniques work the best for adversarial training. Finally, we evaluate our approach on CIFAR-10 against ℓ_∞ and ℓ_2 norm-bounded perturbations of size ϵ = 8/255 and ϵ = 128/255, respectively. We show large absolute improvements of +2.93 robust accuracy compared to previous state-of-the-art methods. In particular, against ℓ_∞ norm-bounded perturbations of size ϵ = 8/255, our model reaches 60.07 also achieve a significant performance boost with this approach while using other architectures and datasets such as CIFAR-100, SVHN and TinyImageNet.
READ FULL TEXT