Phase Transition from Clean Training to Adversarial Training

Yue Xing · Qifan Song · Guang Cheng

Hall J #206

Keywords: [ adversarial training ] [ Adversarial Robustness ]

[ Abstract ]
[ Paper [ Poster [ OpenReview
Thu 1 Dec 9 a.m. PST — 11 a.m. PST


Adversarial training is one important algorithm to achieve robust machine learning models. However, numerous empirical results show a great performance degradation from clean training to adversarial training (e.g., 90+\% vs 67\% testing accuracy on CIFAR-10 dataset), which does not match the theoretical guarantee delivered by the existing studies. Such a gap inspires us to explore the existence of an (asymptotic) phase transition phenomenon with respect to the attack strength: adversarial training is as well behaved as clean training in the small-attack regime, but there is a sharp transition from clean training to adversarial training in the large-attack regime. We validate this conjecture in linear regression models, and conduct comprehensive experiments in deep neural networks.

Chat is not available.