Generative Adversarial Perturbations

Omid Poursaeed, Isay Katsman, Bicheng Gao +1

In this paper, we propose novel generative models for creating adversarial examples, slightly perturbed images resembling natural images but maliciously crafted to fool pre-trained…