Adversarial Training and Provable Defenses: Bridging the Gap

Anonymous

Sep 25, 2019 ICLR 2020 Conference Blind Submission readers: everyone Show Bibtex
  • TL;DR: We propose a novel combination of adversarial training and provable defenses which produces a model with state-of-the-art accuracy and certified robustness on CIFAR-10.
  • Abstract: We propose a new method to train neural networks based on a novel combination of adversarial training and provable defenses. The key idea is to model training as a procedure which includes both, the verifier and the adversary. In every iteration, the verifier aims to certify the network using convex relaxation while the adversary tries to find inputs inside that convex relaxation which cause verification to fail. We experimentally show that this training method is promising and achieves the best of both worlds – it produces a model with state-of-the-art accuracy (74.8%) and certified robustness (55.9%) on the challenging CIFAR-10 dataset with a 2/255 L-infinity perturbation. This is a significant improvement over the currently known best results of 68.3% accuracy and 53.9% certified robustness, achieved using a 5 times larger network than our work.
  • Keywords: adversarial examples, adversarial training, provable defense, convex relaxations, deep learning
0 Replies

Loading