Tactics of Adversarial Attack on Deep Reinforcement Learning Agents

TL;DR

This paper introduces two novel adversarial attack tactics on deep reinforcement learning agents, demonstrating effective reward reduction and target state luring with minimal attack frequency and high success rates.

Contribution

The paper proposes strategically-timed and enchanting attack methods that effectively compromise deep RL agents with reduced detectability and high success, advancing adversarial attack strategies.

Findings

Strategically-timed attack reduces reward as much as uniform attack with fewer attacks.

Enchanting attack achieves over 70% success in luring agents to target states.

Attacks are effective on state-of-the-art deep RL algorithms like DQN and A3C.

Abstract

We introduce two tactics to attack agents trained by deep reinforcement learning algorithms using adversarial examples, namely the strategically-timed attack and the enchanting attack. In the strategically-timed attack, the adversary aims at minimizing the agent's reward by only attacking the agent at a small subset of time steps in an episode. Limiting the attack activity to this subset helps prevent detection of the attack by the agent. We propose a novel method to determine when an adversarial example should be crafted and applied. In the enchanting attack, the adversary aims at luring the agent to a designated target state. This is achieved by combining a generative model and a planning algorithm: while the generative model predicts the future states, the planning algorithm generates a preferred sequence of actions for luring the agent. A sequence of adversarial examples is then crafted to lure the agent to take the preferred sequence of actions. We apply the two tactics to the agents trained by the state-of-the-art deep reinforcement learning algorithm including DQN and A3C. In 5 Atari games, our strategically timed attack reduces as much reward as the uniform attack (i.e., attacking at every time step) does by attacking the agent 4 times less often. Our enchanting attack lures the agent toward designated target states with a more than 70% success rate. Videos are available at http://yenchenlin.me/adversarial_attack_RL/