This class will be given during the second semester on Wednesday afternoon in Room S36 of Building B37. It starts at 2:00pm. The first class is on the 10 th of February. The teaching assistant for the class is Mr. Samy Aittahar (Email: eb.eg1585583169eilu@1585583169rahat1585583169tias1585583169 ).
Students should come to the class with their laptop.
Lesson 1. 05/02/2020. Speaker: Damien Ernst. Introduction to Reinforcement Rearning (RL). Understand how to build a RL agent for non-adversarial environment with discrete state/action spaces.
Lesson 2. 12/02/2020. Speaker: Damien Ernst. The Q-learning algorithm (see Slides Lesson 1).
Lesson 3. 19/02/2020. Speaker: Damien Ernst. Reinforcement learning for continuous state-action spaces (see Slides Lesson 1).
Lesson 4. 26/02/2020. Speaker: Damien Ernst. Discussion Research paper 1.
=== The program hereafter will be modified due to the coronavirus crisis ===
Lesson 8. 25/03/2020. Speaker: Nicolas Vecoven, Pascal Leroy, Amina Benzerga. A glimpse at the research done in RL at the Montefiore Institute. Discussion Research paper 7 .
Lesson 9. 01/04/2020. Reinforcement learning in the energy industry. Real applications. Bert Claessens (Restore).
Lesson 10. 22/04/2020. Speaker: Damien Ernst. Exploration/exploitation in Reinforcement Learning: The multi-armed bandit problems. Class based on a discussion of Research paper 9 (first 25 pages).
Assignment 1 – 05/02/2020. Section 1 to 4 need to be submitted for the 11/02/2020 midnight. Section 5 and 6 for the 18/02/2020. Link to the submission platform: https://submit.montefiore.ulg.ac.be/index.php/login
Results assignments 1
Assignment 2 – 26/02/2020. Section 1 to 4 need to be submitted for the 06/03/2020 midnight. Deadline for the final submission: 26/03/2020 midnight.
Assignment 3 – 17/03/2020 (final assignment). Deadline for the final submission: 30/04/2020. Assignment 3 is not mandatory. But based on the quality of the work, you will get a bonus.
3. Final project
The final project will be about designing your own intelligent agent to control a double inverted pendulum – a well-known but challenging, chaotic physical system – using any reinforcement learning algorithm which is compatible with continuous state and action spaces.
Deadline for the final report is on the 15/05/2020. Presentation of your results: to be disucssed.
Schedule for the project defense
There will be a total of 5 evaluations.
5. Access your results
6. Final exam
Schedule for the final exam