PA-AD Adversarial Attack for Reinforcement Learning

Released code for ICLR 2022 publication: Who Is the Strongest Enemy? Towards Optimal and Efficient Evasion Attacks in Deep RL.

Please see code_atari and code_mujoco for the codebase of atari and mujoco and their running instructions.

Links

OpenReview: https://openreview.net/forum?id=JM2kFbJvvI

Arxiv: https://arxiv.org/abs/2106.05087

Webpage: https://ycsun2017.github.io/paad/index.html

Pretrained Models: https://drive.google.com/file/d/17PiVw9f59YjnbyLm0hn55ekYP_ElIvbo/view?usp=sharing