Provably Efficient Black-Box Action Poisoning Attacks Against Reinforcement Learning

Abstract

Due to the broad range of applications of reinforcement learning (RL), understanding the effects of adversarial attacks against RL model is essential for the safe applications of this model. Prior theoretical works on adversarial attacks against RL mainly focus on either observation poisoning attacks or environment poisoning attacks. In this paper, we introduce a new class of attacks named action poisoning attacks, where an adversary can change the action signal selected by the agent. Compared with existing attack models, the attacker's ability in the proposed action poisoning attack model is more restricted, which brings some design challenges. We study the action poisoning attack in both white-box and black-box settings. We introduce an adaptive attack scheme called LCB-H, which works for most RL agents in the black-box setting. We prove that the LCB-H attack can force any efficient RL agent, whose dynamic regret scales sublinearly with the total number of steps taken, to choose actions according to a policy selected by the attacker very frequently, with only sublinear cost. In addition, we apply LCB-H attack against a popular model-free RL algorithm: UCB-H. We show that, even in the black-box setting, by spending only logarithm cost, the proposed LCB-H attack scheme can force the UCB-H agent to choose actions according to the policy selected by the attacker very frequently.

References in corpus (8)

• Certified Adversarial Robustness via Randomized Smoothing
• Robust Adversarial Reinforcement Learning
• Delving into adversarial attacks on deep policies
• Interpreting Adversarially Trained Convolutional Neural Networks
• Policy Poisoning in Batch Reinforcement Learning and Control
• Data Poisoning Attacks on Stochastic Bandits
• Logarithmic Regret for Reinforcement Learning with Linear Function Approximation
• Reward Poisoning in Reinforcement Learning: Attacks Against Unknown Learners in Unknown Environments