Robust Deep Reinforcement Learning Through Adversarial Attacks and Training: A Survey

L Schott, J Delas, H Hajri, E Gherbi, R Yaich… - arxiv preprint arxiv …, 2024 - arxiv.org
Deep Reinforcement Learning (DRL) is a subfield of machine learning for training
autonomous agents that take sequential actions across complex environments. Despite its …

Improving robustness by action correction via multi-step maximum risk estimation

Q Chen, K Ding, X Zhang, H Zhang, F Zhu - Neural Networks, 2025 - Elsevier
Certifying robustness against external uncertainties throughout the control process to reduce
the risk of instability is very important. Most existing approaches based on adversarial …

Robustness Evaluation of Offline Reinforcement Learning for Robot Control Against Action Perturbations

S Ayabe, T Otomo, H Kera, K Kawamoto - arxiv preprint arxiv:2412.18781, 2024 - arxiv.org
Offline reinforcement learning, which learns solely from datasets without environmental
interaction, has gained attention. This approach, similar to traditional online deep …

Wolfpack Adversarial Attack for Robust Multi-Agent Reinforcement Learning

S Lee, J Hwang, Y Jo, S Han - arxiv preprint arxiv:2502.02844, 2025 - arxiv.org
Traditional robust methods in multi-agent reinforcement learning (MARL) often struggle
against coordinated adversarial attacks in cooperative scenarios. To address this limitation …