Deep Reinforcement Learning
|Vorlesung:||Aufzeichnung,||OPAL||(Dr. J. Vitay)|
|Übung (für Konsultationen):||Donnerstag,||15:30 - 17:00||(Dr. J. Vitay)|
Suggested prerequisites: Mathematics I to IV, Neurocomputing, basic knowledge in Python.
Exam: written examination (90 minutes), 5 ECTS.
Contact: julien dot vitay at informatik dot tu-chemnitz dot de.
Language: English. The exam can of course be done in German.
The course will dive into the field of deep reinforcement learning. It starts with the basics of reinforcement learning (Sutton and Barto, 2017) before explaining modern model-free (DQN, DDPG, PPO) and model-based (I2A, World models, AlphaGo) architectures making use of deep neural networks for function approximation. More "exotic" forms of RL are then presented (successor representations, hierarchical RL, inverse RL, etc).
The different algorithms presented during the lectures will be studied in more details during the exercises, through implementations in Python.
The plan of the course is:
- Reinforcement Learning (MDP, dynamic programming, Monte-Carlo methods, temporal difference)
- Value-based deep RL (DQN)
- Policy gradient methods (A3C, DDPG, TRPO, PPO)
- Model-based RL (Dyna Q, AlphaGo, I2A)
- Successor representations
- Hierarchical RL, Inverse RL, Multi-agent RL
- How do I register for the course?
You can register on OPAL: https://bildungsportal.sachsen.de/opal/auth/RepositoryEntry/21637267457.
- How do I register for the exam?
Registration on SBService happens in December. Only registered students can participate to the exam.
- Who can assist to this course and write the exam?
Students of the Masters Informatik, angewandte Informatik, Neurorobotik and Data Science can write the exam. Others are welcome to assist to the lectures.