Action-Level Backdoor Attacks Against Deep Reinforcement Learning Systems via Adaptive Reward Exploration

Summary

Adapdoor is an attack framework that injects action-level backdoors into deep reinforcement learning models during training, so that adversaries can manipulate action outputs at deployment. It initializes the backdoor reward from benign reward statistics and iteratively fine-tunes it using performance feedback from benign and backdoor tasks. Across 3 DRL algorithms, 11 environments and 53 backdoor tasks, it outperforms existing baselines by 42.0% to 144.2%.

Mitigation

The source says the authors evaluate three potential defenses to explore pathways for mitigating this threat, but it does not name them or state a fix.

Topics

• Adversarial machine learning

Related items

• InfoSigned but Steerable: Measuring and Closing the Text-Vector Integrity Gap in Offline Retrieval-Augmented GenerationSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
• InfoSecuring next-generation edge intelligence: A survey from federated learning to autonomous agents in cyber-physical systemsSimilar attack · OpenAlex (peer-reviewed AI security)
• InfoPop Goes the Verifier: Phoneme-Targeted Data Poisoning of Voice Liveness DetectionSimilar attack · Arxiv (cs.RO + cs.CV security)
• InfoCharacterizing Recoverable Expert-Utilization Signatures in Fine-Tuned Mixture-of-Experts LLMsSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
• InfoDiscrete Unintentional Adversarial Perturbation Detection for DQN Traffic Signal ControlSimilar attack · OpenAlex (peer-reviewed AI security)