Abstract:Ensuring the security of reinforcement learning (RL) models is critical, particularly when they are trained by third parties and deployed in real-world systems. Attackers can implant backdoors into these models, causing them to behave normally under typical conditions, but execute malicious behaviors when specific triggers are activated. In this work, we propose Plan2Cleanse, a test-time detection and mitigation framework that adapts Monte Carlo Tree Search to efficiently identify and neutralize RL backdoor attacks without requiring model retraining. Our approach recasts backdoor detection as a planning problem, enabling systematic exploration of temporally extended trigger sequences while maintaining black-box access to the target policy. By leveraging the detection results, Plan2Cleanse can further achieve efficient mitigation through tree-search preventive replanning. We evaluated our method in competitive MuJoCo environments, simulated O-RAN wireless networks, and Atari games. Plan2Cleanse achieves substantial improvements, increasing trigger detection success rates by more than 61.4 percentage points in stealthy O-RAN scenarios and improving win rates from 35\% to 53\% in competitive Humanoid environments. These results demonstrate the effectiveness of our test-time defense approach and highlight the importance of proactive defenses against backdoor threats in RL deployments. Our implementation is publicly available at this https URL.
| Comments: | Published in Transactions on Machine Learning Research (TMLR) |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.09638 [cs.LG] |
| (or arXiv:2605.09638v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.09638 arXiv-issued DOI via DataCite (pending registration) |
From: Ping-Chun Hsieh [view email]
[v1]
Sun, 10 May 2026 16:34:16 UTC (772 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。