Episodic Control-Based Adversarial Policy Learning in Two-player Competitive Games
Keywords: Reinforcement Learning; Adversarial Policy Training; Episodic Control
Abstract: Training adversarial agents to attack neural network policies has proven to be both effective and practical. However, we observe that existing methods can be further enhanced by distinguishing between states leading to win or lose and encouraging the policy training to prioritize winning states. In this paper, we address this gap by introducing an episodic control-based approach for adversarial policy training. Our method extracts the historical evaluations for states from historical experiences with an episodic memory, and then incorporating these evaluations into the rewards to improve the adversarial policy optimization. We evaluate our approach using two-player competitive games in MuJoCo simulation environments, demonstrating that our method establishes the most promising attack performance and defense difficulty against the victims among the existing adversarial policy training techniques.
Primary Area: reinforcement learning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 5808
Loading