Adversarial Imitation Attack

Mingyi Zhou, Jing Wu, Yipeng Liu +4

Deep learning models are known to be vulnerable to adversarial examples. A practical adversarial attack should require as little as possible knowledge of attacked models. Current s…