Robust Deep Reinforcement Learning against Adversarial Perturbations on State Observations

Abstract

A deep reinforcement learning (DRL) agent observes its states through observations, which may contain natural measurement errors or adversarial noises. Since the observations deviate from the true states, they can mislead the agent into making suboptimal actions. Several works have shown this vulnerability via adversarial attacks, but existing approaches on improving the robustness of DRL under this setting have limited success and lack for theoretical principles. We show that naively applying existing techniques on improving robustness for classification tasks, like adversarial training, is ineffective for many RL tasks. We propose the state-adversarial Markov decision process (SA-MDP) to study the fundamental properties of this problem, and develop a theoretically principled policy regularization which can be applied to a large family of DRL algorithms, including proximal policy optimization (PPO), deep deterministic policy gradient (DDPG) and deep Q networks (DQN), for both discrete and continuous action control problems. We significantly improve the robustness of PPO, DDPG and DQN agents under a suite of strong white box adversarial attacks, including new attacks of our own. Additionally, we find that a robust policy noticeably improves DRL performance even without an adversary in a number of environments. Our code is available at https://github.com/chenhongge/StateAdvDRL.

Huan Zhang and Hongge Chen contributed equally

References in corpus (14)

• Dota 2 with Large Scale Deep Reinforcement Learning
• Theoretically Principled Trade-off between Robustness and Accuracy
• Solving Rubik's Cube with a Robot Hand
• Robust Adversarial Reinforcement Learning
• Safe, Multi-Agent, Reinforcement Learning for Autonomous Driving
• Implementation Matters in Deep Policy Gradients: A Case Study on PPO and TRPO
• Delving into adversarial attacks on deep policies
• Action Robust Reinforcement Learning and Applications in Continuous Control
• Certified Adversarial Robustness for Deep Reinforcement Learning
• Whatever Does Not Kill Deep Reinforcement Learning, Makes It Stronger
• Adversary A3C for Robust Reinforcement Learning
• Optimal Attacks on Reinforcement Learning Policies
• Deep Robust Kalman Filter
• Online Robustness Training for Deep Reinforcement Learning

Cited by in corpus (4)

• Robust Multi-agent Communication via Multi-view Message Certification
• Structure in Deep Reinforcement Learning: A Survey and Open Problems
• Robustness Testing for Multi-Agent Reinforcement Learning: State Perturbations on Critical Agents
• Optimal Attack and Defense for Reinforcement Learning