This repository serves as a curated collection of research papers on Reinforcement Learning (RL), primarily focusing on single-agent RL. It aims to provide researchers and practitioners with a structured overview of key advancements, from classic methods to the latest conference contributions, covering various subfields like model-free, model-based, offline RL, meta-RL, and adversarial RL.

How It Works

The repository organizes papers by sub-topic, presenting them in a tabular format that includes title, method, conference, and a brief description. This structure allows users to quickly identify relevant research and understand the core contributions of each paper. The content is updated with recent papers from top conferences, ensuring coverage of current trends and techniques.

Quick Start & Requirements

This repository is a curated list of papers and does not require installation or execution. It serves as a reference guide.

Highlighted Details

• Comprehensive coverage of RL subfields: Model-Free (Online/Offline), Model-Based, Meta-RL, Adversarial RL, Generalization, Transformers, and more.
• Detailed tables for each subfield, listing paper title, method, conference, and description.
• Regular updates with papers from major conferences like ICLR, ICML, and NeurIPS.
• Includes foundational RL concepts and tutorials.

Maintenance & Community

The repository is maintained by yingchengyang. Further community engagement details are not specified in the README.

Licensing & Compatibility

The repository itself is not software and thus does not have a software license. The papers listed are subject to their respective publication licenses.

Limitations & Caveats

The repository focuses on papers the maintainer found insightful, meaning it may not be exhaustive. The primary focus is on single-agent RL, with limited coverage of multi-agent systems.