Characterizing Attacks on Deep Reinforcement Learning

Xinlei Pan, Chaowei Xiao, Warren He, Shuang Yang, Jian Peng, Mingjie Sun, Mingyan Liu, Bo Li, Dawn Song

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Recent studies show that Deep Reinforcement Learning (DRL) models are vulnerable to adversarial attacks, which attack DRL models by adding small perturbations to the observations. However, some attacks assume full availability of the victim model, and some require a huge amount of computation, making them less feasible for real world applications. In this work, we make further explorations of the vulnerabilities of DRL by studying other aspects of attacks on DRL using realistic and efficient attacks. First, we adapt and propose efficient black-box attacks when we do not have access to DRL model parameters. Second, to address the high computational demands of existing attacks, we introduce efficient online sequential attacks that exploit temporal consistency across consecutive steps. Third, we explore the possibility of an attacker perturbing other aspects in the DRL setting, such as the environment dynamics. Finally, to account for imperfections in how an attacker would inject perturbations in the physical world, we devise a method for generating a robust physical perturbations to be printed. The attack is evaluated on a real-world robot under various conditions. We conduct extensive experiments both in simulation such as Atari games, robotics and autonomous driving, and on real-world robotics, to compare the effectiveness of the proposed attacks with baseline approaches. To the best of our knowledge, we are thefirst to apply adversarial attacks on DRL systems to physical robots.

Original languageEnglish (US)
Title of host publicationInternational Conference on Autonomous Agents and Multiagent Systems, AAMAS 2022
PublisherInternational Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS)
Number of pages9
ISBN (Electronic)9781713854333
StatePublished - 2022
Externally publishedYes
Event21st International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2022 - Auckland, Virtual, New Zealand
Duration: May 9 2022May 13 2022

Publication series

NameProceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS
ISSN (Print)1548-8403
ISSN (Electronic)1558-2914


Conference21st International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2022
Country/TerritoryNew Zealand
CityAuckland, Virtual


  • Adversarial Machine Learning
  • Reinforcement Learning
  • Robotics

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software
  • Control and Systems Engineering


Dive into the research topics of 'Characterizing Attacks on Deep Reinforcement Learning'. Together they form a unique fingerprint.

Cite this