Lecture: Multi-agent deep reinforcement learning

• Lecture
• MADRL

Maxime Toquebiau

AI for Robotics Master, Sorbonne Université, 2024-2025, 2025-2026

Abstract. This is a lecture for Master's students on multi-agent deep reinforcement learning (MADRL). It comprises a 2-hour course presenting the context of multi-agent learning, important MADRL algorithms, and current trends of research. A practical course is associated, with the implementation of some MADRL algorithms and training in a simple environment.

Slides Practical works (.html) Download practical works (.ipynb)

Article: An Introduction to Deep Reinforcement Learning

• Pre-print
• Deep RL

Maxime Toquebiau, Jae-Yun Jun Kim, Faïz Ben Amar, Nicolas Bredeche

2025

Abstract. This article presents an introduction to the field of deep reinforcement learning. It is primarily intended for students taking their first steps in this field. First, a general introduction to reinforcement learning is given, defining the approach and describing its origins and current trends. Then, the essential foundations of reinforcement learning algorithms - i.e., value, policy, and model - are described. Then the main model-free deep reinforcement learning algorithms are explained. Finally, current directions of research on deep reinforcement learning are introduced.

@TechReport{Toquebiau2025_IntroDRL,
author = {Toquebiau, Maxime and Jun Kim, Jae-Yun and Ben Amar, Faïz and Bredeche, Nicolas},
institution = {Sorbonne University},
title = {An Introduction to Deep Reinforcement Learning},
year = {2025},
url = {https://maxtoq.github.io/docs/Preprint_Intro_DRL.pdf},
}

Pre-print