Agnostic Reinforcement Learning

Agnostic reinforcement learning (RL) aims to develop RL agents that can effectively learn and adapt to new tasks or environments with minimal prior knowledge about the underlying dynamics or reward structure. Current research focuses on improving sample efficiency through techniques like causal exploration and incorporating diverse strategies via novel diversity measures and iterative learning algorithms. This approach is significant because it addresses the limitations of traditional RL methods that often require extensive task-specific data or hand-engineered features, paving the way for more robust and generalizable AI agents applicable to diverse real-world problems such as robotics and satellite control.

Papers

October 31, 2024

Progressive Safeguards for Safe and Model-Agnostic Reinforcement Learning
Nabil Omi, Hosein Hasanbeig, Hiteshi Sharma, Sriram K. Rajamani, Siddhartha Sen
LeArning Abstract Meta Learning Safe Reinforcement Learning Finite State Machine Agnostic Meta learnIng Progressive Alignment Agnostic Reinforcement Learning

September 9, 2024

Forward KL Regularized Preference Optimization for Aligning Diffusion Policies
Zhao Shan, Chenyou Fan, Shuang Qiu, Jiyuan Shi, Chenjia Bai
Policy OpTimization Policy Learning Direct Preference Optimization Preference Optimization Agnostic Reinforcement Learning

July 30, 2024

Boosting Efficiency in Task-Agnostic Exploration through Causal Knowledge
Yupei Yang, Biwei Huang, Shikui Tu, Lei Xu
Causal Knowledge Enhancing Efficiency Causal Insight Agnostic Exploration Agnostic Reinforcement Learning Causal Exploration

October 23, 2023

Iteratively Learn Diverse Strategies with State Distance Information
Wei Fu, Weihua Du, Jingwei Li, Sunli Chen, Jingzhao Zhang, Yi Wu
Reinforcement Learning Multi Agent Diversity Metric Distance Information Diversity Reinforcement Learning Diverse Strategy Agnostic Reinforcement Learning

October 9, 2023

When is Agnostic Reinforcement Learning Statistically Tractable?
Zeyu Jia, Gene Li, Alexander Rakhlin, Ayush Sekhari, Nathan Srebro
Optimal Policy Agnostic Learning PAC Learnability Agnostic PAC Agnostic Reinforcement Learning

October 1, 2023

Learning to Make Adherence-Aware Advice
Guanting Chen, Xiaocheng Li, Chunlin Sun, Hanzhao Wang
LeArning Abstract Artificial Intelligence Policy Human Artificial Intelligence Agnostic Reinforcement Learning

October 6, 2022

Low-Thrust Orbital Transfer using Dynamics-Agnostic Reinforcement Learning
Carlos M. Casas, Belen Carro, Antonio Sanchez-Esguevillas
Reinforcement Learning Trajectory Optimization Model Free Reinforcement Learning Agnostic Reinforcement Learning Low Thrust Transfer Low Thrust Trajectory

August 31, 2022

Style-Agnostic Reinforcement Learning
Juyong Lee, Seokjun Ahn, Jaesik Park
Adversarial Training Adversarial Learning Style Representation Agnostic Reinforcement Learning Adversarial Style Perturbation