State Adversarial

State adversarial research focuses on enhancing the robustness of reinforcement learning (RL) agents against attacks that subtly manipulate their input observations (states). Current work investigates this problem across various RL architectures, including Q-learning and multi-agent systems like QMIX, often employing adversarial training techniques or exploring alternative policy optimization methods to improve resilience. This research is crucial for deploying RL agents in real-world scenarios where adversarial manipulations could have significant consequences, impacting the reliability and safety of autonomous systems and other applications.

Papers

October 15, 2024

Adversarially Guided Stateful Defense Against Backdoor Attacks in Federated Deep Learning
Hassan Ali, Surya Nepal, Salil S. Kanhere, Sanjay Jha
Backdoor Attack Adversarial Perturbation Federated Deep Ensemble Defense State Adversarial

April 16, 2024

Self-playing Adversarial Language Game Enhances LLM Reasoning
Pengyu Cheng, Tianhao Hu, Han Xu, Zhisong Zhang, Yong Dai, Lei Han, Nan Du
Self Play Adversarial Game State Adversarial

February 20, 2024

Beyond Worst-case Attacks: Robust RL with Adaptive Defense via Non-dominated Policies
Xiangyu Liu, Chenghao Deng, Yanchao Sun, Yongyuan Liang, Furong Huang
Reinforcement Learning Sublinear Regret Robust Reinforcement Learning Diverse Policy Adversarial Bandit State Adversarial Adaptive Defense

February 3, 2024

Towards Optimal Adversarial Robust Q-learning with Bellman Infinity-error
Haoran Li, Zicheng Zhang, Wang Luo, Congying Han, Yudong Hu, Tiande Guo, Shichen Liao
Deep Reinforcement Learning Q Learning Robust Policy Bellman Error Optimal Robust Policy State Adversarial

July 3, 2023

Enhancing the Robustness of QMIX against State-adversarial Attacks
Weiran Guo, Guanjun Liu, Ziyuan Zhou, Ling Wang, Jiacun Wang
Deep Reinforcement Learning Native Robustness Multi Agent Reinforcement Learning Robust Reinforcement Learning Soft QMIX State Adversarial

May 4, 2023

Toward Evaluating Robustness of Reinforcement Learning with Adversarial Policy
Xiang Zheng, Xingjun Ma, Shengjie Wang, Xinyu Wang, Chao Shen, Cong Wang
Reinforcement Learning Robustness Evaluation Adversarial Agent Adversarial Policy State Adversarial

November 7, 2022

Are AlphaZero-like Agents Robust to Adversarial Perturbations?
Li-Cheng Lan, Huan Zhang, Ti-Rong Wu, Meng-Yu Tsai, I-Chen Wu, Cho-Jui Hsieh
Adversarial Example Adversarial Perturbation Agent Smith State Adversarial