M2 Artificial Intelligence - Reinforcement Learning

Lectures by L. Mathelin, practical sessions by R. Hosseinkhan-Boucher, M2 Artificial Intelligence, Université Paris-Saclay.

General Resources

Excellent resources for Reinforcement Learning:


2026-2027

The practical sessions take place on Fridays at the PUIO, room E105.

Lecture Slides

Practical Sessions

Practical Session 1 (20 November 2026): Markov Decision Processes and Finite-Horizon Dynamic Programming


Archive: 2025-2026

Several statements of these notebooks contain errors, corrected in the 2026-2027 exercise sheets as they are released.

Lecture Slides

Notebooks

Practical Session 1: Markov Decision Processes and Finite-Horizon Dynamic Programming

Practical Session 2: Discounted Infinite-horizon Markov Decision Problems and Iterative Methods

Practical Session 3: Stochastic Approximation and Reinforcement Learning

Practical Session 4: Approximation Methods in Reinforcement Learning