Type I Attack for Generative Models

Chengjin Sun, Sizhe Chen, Jia Cai +1

Generative models are popular tools with a wide range of applications. Nevertheless, it is as vulnerable to adversarial samples as classifiers. The existing attack methods mainly f…