Robust Policy

Robust policy learning in reinforcement learning (RL) focuses on developing agents capable of performing well despite uncertainties and disturbances in their environment, a crucial step for deploying RL in real-world settings. Current research emphasizes techniques like adversarial training, curriculum learning, and the use of factored state representations to improve robustness, often employing model architectures such as deep Q-networks, actor-critic methods, and transformers. These advancements are significant because they address the critical challenge of generalization and reliability in RL, paving the way for safer and more dependable autonomous systems across various applications.

Papers

December 20, 2021

Learning Robust Policy against Disturbance in Transition Dynamics via State-Conservative Policy Optimization
Yufei Kuang, Miao Lu, Jie Wang, Qi Zhou, Bin Li, Houqiang Li
Robust Policy Unknown Disturbance Transition Dynamic Conservative Policy

December 2, 2021

Sample Complexity of Robust Reinforcement Learning with a Generative Model
Kishan Panaganti, Dileep Kalathil
Generative Modeling Sample Complexity Robust Policy Robust Reinforcement Learning Optimal Robust Policy Robust Markov Decision

November 6, 2021

Robust Deep Reinforcement Learning for Quadcopter Control
Aditya M. Deshpande, Ali A. Minai, Manish Kumar
Deep Reinforcement Learning Robust Policy Quadrotor Control Drone Control Robust Deep Reinforcement Learning

December 8, 2011

The Extended UCB Policies for Frequentist Multi-armed Bandit Problems
Keqin Liu, Tianshuo Zheng, Haoran Chen
Multi Armed Bandit Robust Policy Sequential Decision Making UCB Algorithm Heavy Tailed Reward

Robust Policy

Papers

Learning Robust Policy against Disturbance in Transition Dynamics via State-Conservative Policy Optimization

Sample Complexity of Robust Reinforcement Learning with a Generative Model

Robust Deep Reinforcement Learning for Quadcopter Control

The Extended UCB Policies for Frequentist Multi-armed Bandit Problems