mlsanityData to 7 Oct 2026

Rethinking Adversarial Attacks in Reinforcement Learning from Policy Distribution Perspective

arXiv 2501.03562 ↗First posted 7 Jan 2025cs.LG cs.AIcited by 10

Not on the map. Why?

Abstract (arXiv)

Deep Reinforcement Learning (DRL) suffers from uncertainties and inaccuracies in the observation signal in realworld applications. Adversarial attack is an effective method for evaluating the robustness of DRL agents. However, existing attack methods targeting individual sampled actions have limited impacts on the overall policy distribution, particularly in continuous action spaces. To address these limitations, we propose the Distribution-Aware Projected Gradient Descent attack (DAPGD). DAPGD uses distribution similarity as the gradient perturbation input to attack the policy network, which leverages the entire policy distribution rather than relying on individual samples. We utilize the Bhattacharyya distance in DAPGD to measure policy similarity, enabling sensitive detection of subtle but critical differences between probability distributions. Our experiment results demonstrate that DAPGD achieves SOTA results compared to the baselines in three robot navigation tasks, achieving an average 22.03% higher reward drop compared to the best baseline.

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Builds on 29 of its 40 references that are in the map. First posted on arXiv 7 Jan 2025.

Builds on

29 of 40 references.

References · 29

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  2. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  3. 2015Trust Region Policy Optimizationcited by 2,503
  4. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,510
  5. 2017Constrained Policy Optimizationcited by 928
  6. 2018Improving Transferability of Adversarial Examples with Input Diversitycited by 591
  7. 2017One pixel attack for fooling deep neural networkscited by 579
  8. 2019Nesterov Accelerated Gradient and Scale Invariance for Adversarial Attackscited by 352
  9. 2017Adversarial Attacks on Neural Network Policiescited by 314
  10. 2020Robust Deep Reinforcement Learning against Adversarial Perturbations on State Observationscited by 235
  11. 2019Adversarial Attacks and Defenses in Images, Graphs and Text: A Reviewcited by 234
  12. 2017Tactics of Adversarial Attack on Deep Reinforcement Learning Agentscited by 163
17 more
  1. 2017Robust Deep Reinforcement Learning with Adversarial Attackscited by 152
  2. 2023Safety-Gymnasium: A Unified Safe Reinforcement Learning Benchmarkcited by 135
  3. 2021Robust Reinforcement Learning on State Observations with Learned Optimal Adversarycited by 133
  4. 2020CRPO: A New Approach for Safe Reinforcement Learning with Convergence Guaranteecited by 128
  5. 2018SparseFool: a few pixels make a big differencecited by 94
  6. 2020Robust Deep Reinforcement Learning through Adversarial Losscited by 73
  7. 2018Adv-BNN: Improved Adversarial Defense through Robust Bayesian Neural Networkcited by 68
  8. 2020Stealthy and Efficient Adversarial Attacks against Deep Reinforcement Learningcited by 62
  9. 2020A Survey of Reinforcement Learning Algorithms for Dynamically Varying Environmentscited by 62
  10. 2019Online Robustness Training for Deep Reinforcement Learningcited by 26
  11. 2022Pixle: a fast and effective black-box attack based on rearranging pixelscited by 20
  12. 2023Adversarial Robust Deep Reinforcement Learning Requires Redefining Robustnesscited by 18
  13. 2022CUP: A Conservative Update Policy Algorithm for Safe Reinforcement Learningcited by 16
  14. 2020Deep Reinforcement learning for real autonomous mobile robot navigation in indoor environmentscited by 13
  15. 2023Towards Building More Robust Models with Frequency Biascited by 12
  16. 2019Reinforcement Learning for Robotic Manipulation using Simulated Locomotion Demonstrationscited by 10
  17. 2020Robotic Arm Control and Task Training through Deep Reinforcement Learningcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 9; 2026: 1.

10 papers.

  1. 2025LLM-Driven Stationarity-Aware Expert Demonstrations for Multi-Agent Reinforcement Learning in Mobile Systemscited by 8
  2. 2025State-Aware Perturbation Optimization for Robust Deep Reinforcement Learningcited by 7
  3. 2025Robust Deep Reinforcement Learning in Robotics via Adaptive Gradient-Masked Adversarial Attackscited by 7
  4. 2025Rethinking Membership Inference Attacks Against Transfer Learningcited by 5
  5. 2025Defensive Adversarial CAPTCHA: A Semantics-Driven Framework for Natural Adversarial Example Generationcited by 3
  6. 2025Sample Efficient Experience Replay in Non-stationary Environmentscited by 2
  7. 2025Secure Resource Allocation via Constrained Deep Reinforcement Learningcited by 1
  8. 2025SFL-LEO: Asynchronous Split-Federated Learning Design for LEO Satellite-Ground Network Frameworkcited by 1
  9. 2025DP-TRAE: A Dual-Phase Merging Transferable Reversible Adversarial Example for Image Privacy Protectioncited by 0
  10. 2026CIVA: Critic-Induced Value-Subspace Attacks on Visual World-Model Agentscited by 0