In charge of an internal reinforcement learning training program open to researchers and engineers at EDF R&D. The program covers fundamental concepts, algorithms, and practical applications of reinforcement learning in industrial settings. The course follows Emmanuel Rachelson’s MVA class material.

Approximate dynamic programming

• Markov Decision Processes
• Value iteration
• Q-learning
• Deep Q-Networks

Policy Gradient methods

• Policy gradient theorem
• REINFORCE
• Actor-Critic methods
• Proximal Policy Optimization