Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers

Code

• github.com/Hadisalman/smoothing-adversarial
  OfficialIn paperpytorch★ 0
• github.com/akshaymehra24/poisoning_certified_defenses
  tf★ 14
• github.com/azshue/boosting_robust
  pytorch★ 0

Abstract

Recent works have shown the effectiveness of randomized smoothing as a scalable technique for building neural network-based classifiers that are provably robust to _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 through extensive experimentation that our method consistently outperforms all existing provably _2-robust classifiers by a significant margin on ImageNet and CIFAR-10, establishing the state-of-the-art for provable _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 .

Tasks

Reproductions