Toward Evaluating Robustness of Deep Reinforcement Learning with Continuous Control
Tsui-Wei Weng, Krishnamurthy (Dj) Dvijotham*, Jonathan Uesato*, Kai Xiao*, Sven Gowal*, Robert Stanforth*, Pushmeet Kohli
Abstract
Deep reinforcement learning has achieved great success in many previously difficult reinforcement learning tasks, yet recent studies show that deep RL agents are also unavoidably susceptible to adversarial perturbations, similar to deep neural networks in classification tasks. Prior works mostly focus on model-free adversarial attacks and agents with discrete actions. In this work, we study the problem of continuous control agents in deep RL with adversarial attacks and propose the first two-step algorithm based on learned model dynamics. Extensive experiments on various MuJoCo domains (Cartpole, Fish, Walker, Humanoid) demonstrate that our proposed framework is much more effective and efficient than model-free based attacks baselines in degrading agent performance as well as driving agents to unsafe states.
BibTeX
@inproceedings{
Weng2020Toward,
title={Toward Evaluating Robustness of Deep Reinforcement Learning with Continuous Control},
author={Tsui-Wei Weng and Krishnamurthy (Dj) Dvijotham* and Jonathan Uesato* and Kai Xiao* and Sven Gowal* and Robert Stanforth* and Pushmeet Kohli},
booktitle={International Conference on Learning Representations},
year={2020},
url={https://openreview.net/forum?id=SylL0krYPS}
}