Are Generative Classifiers More Robust to Adversarial Attacks? [article]

Yingzhen Li, John Bradshaw, Yash Sharma
2019 arXiv   pre-print
There is a rising interest in studying the robustness of deep neural network classifiers against adversaries, with both advanced attack and defence techniques being actively developed. However, most recent work focuses on discriminative classifiers, which only model the conditional distribution of the labels given the inputs. In this paper, we propose and investigate the deep Bayes classifier, which improves classical naive Bayes with conditional deep generative models. We further develop
more » ... ion methods for adversarial examples, which reject inputs with low likelihood under the generative model. Experimental results suggest that deep Bayes classifiers are more robust than deep discriminative classifiers, and that the proposed detection methods are effective against many recently proposed attacks.
arXiv:1802.06552v3 fatcat:tyjy55zu7bfkvexaj5vh6u2diq