cs.AINov 13, 2026

When Cooperation Must Change: Semantic Strategy Unlearning in Multi-Agent Reinforcement Learning

Authors: Wenhan Wu, Linghan Chen, Yudong Gao, Fan Yang, Honglong Chen, Kai Yin, Jiawei Jiang, Chuang Hu

Organizations: School of Computer Science, Wuhan University, Wuhan, Hubei, China · School of Computer Science and Information Technology, University of Adelaide, Adelaide, South Australia, Australia · Department of Computer Science and Engineering, Hong Kong University of Science and Technology, Hong Kong SAR · School of Electromechanical and Automation, Huaqiao University, Xiameng, Fujian, China · 

Abstract

Cooperative multi-agent reinforcement learning (MARL) supports coordinated decision-making in distributed Web services, multi-robot control, and collaborative games through learned coordination conventions. When deployment requirements change, a previously useful convention may become obsolete or unsafe. For example, revised cache-sharing permissions in Web content delivery may require caches to abandon a learned division of storage and serving roles while maintaining efficient delivery. Teams must therefore remove the specified convention while preserving useful cooperation. This is challenging because the same relational pattern can be reproduced by different agents across different histories. Runtime shielding leaves the underlying coordination preference intact and can impair task progress, while existing reinforcement unlearning and trajectory-deletion methods do not directly target role-invariant conventions. This paper formalizes this problem as semantic coordination-strategy unlearning and proposes CARVE.CARVE uses causal coordination attribution to identify influential interactions and retain-aware subspace updates to limit interference with unrelated behaviors. Counterfactual re-coordination helps agents learn productive alternatives, while adversarial auditing tests whether the target convention can still be recovered. Accepted updates are merged into decentralized actors for direct deployment. Evaluation across five cooperative MARL benchmarks shows that CARVE reduces mean nominal convention-completion and worst-case elicitation rates by 91.4% and 83.3%, respectively.The median utility loss is 5.9%, and measured inference latency matches that of the original policies.

Figures & tables

Appendix figures & tables2 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Position: Beyond Training Time, Test-Time Coordination is Essential for Cooperative MARL

    Oct 12, 2026Dongsu Lee, Sooraj Sathish, Joonkyung Kim +2Centralized Training with Decentralized ExecutionCooperative MARL

  2. Interaction-Breaking Adversarial Learning Framework for Robust Multi-Agent Reinforcement Learning

    May 18, 2026Sunwoo Lee, Mingu Kang, Yonghyeon Jo +1Adversarial TrainingMulti-Agent Coordination