Adversarial Reinforcement Learning

Adversarial reinforcement learning (ARL) focuses on training agents that can perform optimally even when facing adversaries attempting to disrupt their actions or observations. Current research emphasizes developing robust algorithms, often employing minimax optimization within frameworks like Stackelberg games, and exploring various model architectures including deep neural networks and decision transformers to handle complex environments and high-dimensional data. This field is crucial for building reliable AI systems in safety-critical applications like autonomous driving and robotics, where robustness against unexpected events or malicious attacks is paramount. The development of provably robust and efficient ARL algorithms is a key focus, addressing challenges like convergence guarantees and sample efficiency.

Papers

September 29, 2024

Learning Robust Policies via Interpretable Hamilton-Jacobi Reachability-Guided Disturbances
Hanyang Hu, Xilun Zhang, Xubo Lyu, Mo Chen
Deep Reinforcement Learning Robust Policy Adversarial Reinforcement Learning Unknown Disturbance Hamilton Jacobi Reachability

September 26, 2024

Cat-and-Mouse Satellite Dynamics: Divergent Adversarial Reinforcement Learning for Contested Multi-Agent Space Operations
Cameron Mehlman, Joseph Abramov, Gregory Falco
Multi Agent Reinforcement Learning Multi Agent Adversarial Reinforcement Learning MULTI Agent

September 20, 2024

MAGICS: Adversarial RL with Minimax Actors Guided by Implicit Critic Stackelberg for Convergent Neural Synthesis of Robot Safety
Justin Wang, Haimin Hu, Duy Phuong Nguyen, Jaime Fernández Fisac
Adversarial Reinforcement Learning Robot Safety Neural Synthesis

September 19, 2024

Defending against Reverse Preference Attacks is Difficult
Domenic Rosati, Giles Edkins, Harsh Raj, David Atanasov, Subhabrata Majumdar, Janarthanan Rajendran, Frank Rudzicz, Hassan Sajjad
Adversarial Reinforcement Learning Training Time Attack Adversarial Reward Preference Poisoning

August 31, 2024

Robust off-policy Reinforcement Learning via Soft Constrained Adversary
Kosuke Nakanishi, Akihiro Kubo, Yuji Yasui, Shin Ishii
Robust Learning Policy Reinforcement Learning Robust Reinforcement Learning Adversarial Reinforcement Learning Soft Constraint

July 25, 2024

Adversarially Robust Decision Transformer
Xiaohang Tang, Afonso Marques, Parameswaran Kamalaruban, Ilija Bogunovic
Native Robustness Decision Transformer Minimax Optimality Adversarial Reinforcement Learning Adversarial Environment Transformer Based Generative Adversarial

June 7, 2024

Probabilistic Perspectives on Error Minimization in Adversarial Reinforcement Learning
Roman Belaire, Arunesh Sinha, Pradeep Varakantham
Adversarial Attack Deep Reinforcement Learning Adversarial Perturbation Bayesian Perspective Adversarial Noise Adversarial Reinforcement Learning Error Minimizing Noise

June 5, 2024

Dynamically Expanding Capacity of Autonomous Driving with Near-Miss Focused Training Framework
Ziyuan Yang, Zhaoyang Li, Jianming Hu, Yi Zhang
Autonomous Driving Autonomous Vehicle Reward Signal Capacity Loss Adversarial Reinforcement Learning

March 1, 2024

Scale-free Adversarial Reinforcement Learning
Mingyu Chen, Xuezhou Zhang
Adversarial Reinforcement Learning Adversarial Markov Decision Process Adversarial Multi Armed Bandit

November 30, 2023

Optimal Attack and Defense for Reinforcement Learning
Jeremy McMahan, Young Wu, Xiaojin Zhu, Qiaomin Xie
Reinforcement Learning Stackelberg Game Adversarial Reinforcement Learning Defense Algorithm Observable Stochastic Game Optimal Attack

November 9, 2023

Counter-Empirical Attacking based on Adversarial Reinforcement Learning for Time-Relevant Scoring System
Xiangguo Sun, Hong Cheng, Hang Dong, Bo Qiao, Si Qin, Qingwei Lin
Credit Scoring Adversarial Reinforcement Learning Learning Problem Scoring Function

November 3, 2023

Robust Adversarial Reinforcement Learning via Bounded Rationality Curricula
Aryaman Reddi, Maximilian Tölle, Jan Peters, Georgia Chalvatzaki, Carlo D'Eramo
Reinforcement Learning Adversarial Reinforcement Learning Zero Sum Markov Game Bounded Rational

October 31, 2023

Amoeba: Circumventing ML-supported Network Censorship via Adversarial Reinforcement Learning
Haoyu Liu, Alec F. Diallo, Paul Patras
Adversarial Reinforcement Learning General Adversarial New Censorship Instance Global Bypass Consistency Flow Generative Adversarial Network

September 3, 2023

Deception Game: Closing the Safety-Learning Loop in Interactive Robot Autonomy
Haimin Hu, Zixu Zhang, Kensuke Nakamura, Andrea Bajcsy, Jaime F. Fisac
Autonomous Vehicle Learning Dynamic Safety Guarantee Robot Autonomy Adversarial Reinforcement Learning Strategic Deception Safety Learning

August 29, 2023

Adversarial Style Transfer for Robust Policy Optimization in Deep Reinforcement Learning
Md Masudur Rahman, Yexiang Xue
Reinforcement Learning Deep Reinforcement Learning Reinforcement Learning Agent Robust Policy Adversarial Reinforcement Learning Deep Reinforcement Learning Algorithm Adversarial Transfer

August 22, 2023

Robust Lagrangian and Adversarial Policy Gradient for Robust Constrained Markov Decision Processes
David M. Bossens
Reinforcement Learning Policy Gradient Robust Policy Lagrangian System Adversarial Reinforcement Learning Adversarial Policy

August 9, 2023

Adversarial Deep Reinforcement Learning for Cyber Security in Software Defined Networks
Luke Borchjes, Clement Nyirenda, Louise Leenen
Deep Reinforcement Learning Deep Q Network Software Defined Adversarial Reinforcement Learning Cyber Security Double Deep Q Network

July 31, 2023

C-DARL: Contrastive diffusion adversarial representation learning for label-free blood vessel segmentation
Boah Kim, Yujin Oh, Bradford J. Wood, Ronald M. Summers, Jong Chul Ye
Vessel Segmentation Adversarial Reinforcement Learning Contrastive Diffusion Blood Vessel Segmentation

June 13, 2023

Using Collision Momentum in Deep Reinforcement Learning Based Adversarial Pedestrian Modeling
Dianwei Chen, Ekim Yurtsever, Keith Redmill, Umit Ozguner
Reinforcement Learning Algorithm Adversarial Reinforcement Learning Driving Algorithm Pedestrian Simulation Social Force Pedestrian Behavior Prediction Impact Mechanic

May 15, 2023

Horizon-free Reinforcement Learning in Adversarial Linear Mixture MDPs
Kaixuan Ji, Qingyue Zhao, Jiafan He, Weitong Zhang, Quanquan Gu
Regret Bound Episodic Reinforcement Learning Adversarial Reinforcement Learning Adversarial Markov Decision Process Planning Horizon Model Free Policy

Adversarial Reinforcement Learning

Papers

Learning Robust Policies via Interpretable Hamilton-Jacobi Reachability-Guided Disturbances

Cat-and-Mouse Satellite Dynamics: Divergent Adversarial Reinforcement Learning for Contested Multi-Agent Space Operations

MAGICS: Adversarial RL with Minimax Actors Guided by Implicit Critic Stackelberg for Convergent Neural Synthesis of Robot Safety

Defending against Reverse Preference Attacks is Difficult

Robust off-policy Reinforcement Learning via Soft Constrained Adversary

Adversarially Robust Decision Transformer

Probabilistic Perspectives on Error Minimization in Adversarial Reinforcement Learning

Dynamically Expanding Capacity of Autonomous Driving with Near-Miss Focused Training Framework

Scale-free Adversarial Reinforcement Learning

Optimal Attack and Defense for Reinforcement Learning

Counter-Empirical Attacking based on Adversarial Reinforcement Learning for Time-Relevant Scoring System

Robust Adversarial Reinforcement Learning via Bounded Rationality Curricula

Amoeba: Circumventing ML-supported Network Censorship via Adversarial Reinforcement Learning

Deception Game: Closing the Safety-Learning Loop in Interactive Robot Autonomy

Adversarial Style Transfer for Robust Policy Optimization in Deep Reinforcement Learning

Robust Lagrangian and Adversarial Policy Gradient for Robust Constrained Markov Decision Processes

Adversarial Deep Reinforcement Learning for Cyber Security in Software Defined Networks

C-DARL: Contrastive diffusion adversarial representation learning for label-free blood vessel segmentation

Using Collision Momentum in Deep Reinforcement Learning Based Adversarial Pedestrian Modeling

Horizon-free Reinforcement Learning in Adversarial Linear Mixture MDPs