Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers [article]

Hadi Salman, Greg Yang, Jerry Li, Pengchuan Zhang, Huan Zhang, Ilya Razenshteyn, Sebastien Bubeck
2020 arXiv   pre-print
Recent works have shown the effectiveness of randomized smoothing as a scalable technique for building neural network-based classifiers that are provably robust to ℓ_2-norm adversarial perturbations. In this paper, we employ adversarial training to improve the performance of randomized smoothing. We design an adapted attack for smoothed classifiers, and we show how this attack can be used in an adversarial training setting to boost the provable robustness of smoothed classifiers. We demonstrate
more » ... through extensive experimentation that our method consistently outperforms all existing provably ℓ_2-robust classifiers by a significant margin on ImageNet and CIFAR-10, establishing the state-of-the-art for provable ℓ_2-defenses. Moreover, we find that pre-training and semi-supervised learning boost adversarially trained smoothed classifiers even further. Our code and trained models are available at http://github.com/Hadisalman/smoothing-adversarial .
arXiv:1906.04584v5 fatcat:whnwu6rcqvf33fuxolghsqsf3u