Welcome to our GitHub repository dedicated to significant research papers in the vibrant field of Reinforcement Learning (RL)! Here, we curate essential papers accepted at elite academic conferences such as AAAI, IJCAI, NeurIPS, ICML, ICLR, ICRA, and AAMAS. This resource hub is designed to keep you updated on the latest developments in reinforcement learning, facilitate your exploration of research trends, and introduce you to cutting-edge algorithms and methodologies.

News and Updates

• 2023/11/12: Related repository added.
• 2023/08/19: Papers accepted at AAAI23, IJCAI23, ICRA23, ICML23, ICLR23 and NeurIPS22 added.
• 2023/01/16: Repository created.

Contributing to Our Repository

We invite you to contribute! Format your additions using **Markdown**:

• Paper Name. [[pdf](link)] [[code](link)] – Author 1, Author 2, Author 3. *conference, year*.

Feel free to contact us or add a pull request if you’d like to contribute. For any questions, don’t hesitate to reach out!

Table of Contents

• 1. Multi-Agent Reinforcement Learning (MARL)
• 2. Meta Reinforcement Learning (Meta RL)
• 3. Hierarchical Reinforcement Learning (HRL)
• 4. Multi-Task Reinforcement Learning (Multi-Task RL)
• 5. Offline Reinforcement Learning (Offline RL)
• 6. Inverse Reinforcement Learning (IRL)
• 7. Reinforcement Learning with Large Language Models

1. Multi-Agent Reinforcement Learning

Explore the dynamic realm of multi-agent reinforcement learning where multiple agents cooperate or compete to achieve their individual goals. Think of it as a bustling marketplace where vendors (agents) negotiate, sell their goods (strategies), and adapt based on customer responses (rewards).

• Online Tuning for Offline Decentralized Multi-Agent Reinforcement Learning. [[pdf](https://doi.org/10.1609/aaai.v37i7.25973)] – Jiechuan Jiang, Zongqing Lu. *AAAI 2023*.
• Reward Poisoning Attacks on Offline Multi-Agent Reinforcement Learning. [[pdf](https://doi.org/10.1609/aaai.v37i9.26240)] – Young Wu et al. *AAAI 2023*.
• Models as Agents: Optimizing Multi-Step Predictions of Interactive Local Models in Model-Based Multi-Agent Reinforcement Learning. [[pdf](https://doi.org/10.1609/aaai.v37i9.26241)] – Zifan Wu et al. *AAAI 2023*.

For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

2. Meta Reinforcement Learning

This section aims at equipping agents with the ability to adapt their learning process based on variations in tasks, similar to a shape-shifting actor who adjusts their performance according to audience reactions and director’s feedback.

For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

3. Hierarchical Reinforcement Learning

The hierarchical approach mimics a company’s management structure, where each level of management focuses on different strategies while aligned towards common goals. Executives set the vision while mid-managers handle the operations.

4. Multi-Task Reinforcement Learning

This is akin to a professional multitasker who juggles various responsibilities efficiently, striking a balance while still performing well in every area without dropping any task.

5. Offline Reinforcement Learning

Imagine a student studying from previous exams to prepare for the next—analyzing past data to learn strategies without engaging dynamically in a live environment. They learn solely from the feedback of completed tests.

6. Inverse Reinforcement Learning

Like a detective piecing together clues to understand the motivations behind actions, inverse reinforcement learning focuses on deducing the rewards that drive observed behavior.

7. Reinforcement Learning with Large Language Models

In this section, we look at how language models use reinforcement learning to better understand human preferences, much like personalizing content to match user interests based on their feedback.

Troubleshooting

Should you face any issues while navigating through papers or accessing resources, ensure that your internet connection is stable and try refreshing your browser. If the problem persists, feel free to reach out for assistance. For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.