Deep Reinforcement Learning

Deep Reinforcement Learning

News

15.10.2026: Self-study until 22.10. On Thursday 15.10., there is no regular exercise. Until 22.10., please do the following on your own:

  1. Get your FRIZ account for the computer pool: register on OPAL, then pick up your login data at the FRIZ user service in room A11.202 (Mon–Fri 10:30–11:30 or 13:00–14:00).
  2. Set up Python in the pool: follow the page Python setup on the course website (TU login).
  3. Work through the notebooks of exercise 1 (Python) and exercise 2 (Numpy and Matplotlib) from the course website. The appendix at the end of each notebook is optional.

The pool is open Mon–Fri 07:00–20:30 unless another course is using it (occupancy plan), but the FRIZ user service is there only at the times given in step 1, so pick up your login data then.

Problems? Post the exact error message, or a screenshot, in the forum of our OPAL course, or come to the optional help session on Thursday 22.10., 07:30, in room A11.202. From 29.10. on, the exercises take place every Thursday at 07:30 in A11.202.

Winter semester 2026/27:

Lecture (English): Tuesday, 9:15 - 10:45, A10.375, (O. Maith)
Exercise (English): Thursday, 07:30 - 09:00 A11.202, (O. Maith)

First lecture: Tuesday 13.10., 09:15, room A10.375

First exercise: Thursday 15.10., 07:30, room A11.202

General Information

Suggested prerequisites: Mathematics I to IV, Neurocomputing, basic knowledge in Python.

Exam: written exam (90 minutes), English & German, 5 ECTS.

Contact:

Language: English. The exam is in English & German.

Registration (for receiving emails): Opal

Registration (for exercise computer pool): Opal

Content

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.

Materials

All materials are available at: https://www.tu-chemnitz.de/informatik/KI/edu/deeprl/materials

TUCaktuell