Timezone: »

 
Poster
Data Augmentation Can Improve Robustness
Sylvestre-Alvise Rebuffi · Sven Gowal · Dan Andrei Calian · Florian Stimberg · Olivia Wiles · Timothy A Mann

Tue Dec 07 08:30 AM -- 10:00 AM (PST) @
in Poster Session 1 »
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 $\ell_\infty$ and $\ell_2$ norm-bounded perturbations of size $\epsilon = 8/255$ and $\epsilon = 128/255$, respectively. We show large absolute improvements of +2.93% and +2.16% in robust accuracy compared to previous state-of-the-art methods. In particular, against $\ell_\infty$ norm-bounded perturbations of size $\epsilon = 8/255$, our model reaches 60.07% robust accuracy without using any external data. We also achieve a significant performance boost with this approach while using other architectures and datasets such as CIFAR-100, SVHN and TinyImageNet.
Keywords: Robustness · Adversarial Robustness and Security

Author Information

Sylvestre-Alvise Rebuffi (University of Oxford)
Sven Gowal (DeepMind)
Dan Andrei Calian (DeepMind)
Florian Stimberg (DeepMind)
Olivia Wiles (DeepMind)
Timothy A Mann (DeepMind)

More from the Same Authors