Constructing Unrestricted Adversarial Examples with Generative Models

• Main Authors: Song, Yang, Shu, Rui, Kushman, Nate, Ermon, Stefano
• Format: Journal Article
• Language: English
• Published: 21.05.2018
• Subjects: Computer Science - Artificial Intelligence; Computer Science - Computer Vision and Pattern Recognition; Computer Science - Cryptography and Security; Computer Science - Learning; Statistics - Machine Learning

Adversarial examples are typically constructed by perturbing an existing data point within a small matrix norm, and current defense methods are focused on guarding against this type of attack. In this paper, we propose unrestricted adversarial examples, a new threat model where the attackers are not...