Sie sind hier: Startseite Teaching SS2023 Seminar on Current Works in Deep …

Seminar on Current Works in Deep Reinforcement Learning


Prof. Dr. Joschka Boedecker


Jasper Hoffmann and Yannick Vogt


Deep Learning and Deep Reinforcement Learning have shown tremendous potential in Medicine, Robotics, or Optimal Control by detecting tumors, understanding the surrounding environment, or even controlling nuclear fusion reactors. However, there are still a lot of challenges when applying Deep Learning and especially Deep Reinforcement Learning in the real world. The goal of this seminar is to allow the participants to dive into current research tackling these challenges and to get an insight into what the Neurorobotics lab is currently interested in.

The format is the following: Each participant can express interest in research papers on the topics listed below. The student will then read it and discuss it with a supervisor. After understanding it, the student will give a presentation followed by a discussion afterward. Finally, the student will write a short report on the paper (4 pages). Additionally, we will discuss general aspects of scientific presentations and academic honesty in the beginning.

Seminar Paper Example


Course information

Course number: 11LE13S-7334-MB
Places: 8

Lectures will be in a hybrid format:

Meeting room 42 on the 1st floor of the new IMBIT building (Georges-Köhler-Allee 201). This is right next to the kitchen area.

Zoom session details (can be used for all meeting if you cannot attend in person):

Meeting ID: 638 2089 0181
Passcode: 9hfd7YfF2

Course Schedule:

Introduction:  Thursday, April 20th, 16:00 c.t.

Lecture on how to present: Thursday, May 25th, 14:00

RL Introduction (optional): If desired, will provide you with relevant reads and lecture recordings.

Deadline for Poster Print: Monday, July 17th at midnight. If you are too late, you will need to print the poster on your own.

(Updated!) Poster Session: Thursday, July 20th

Deadline Report: Friday, September 1st

Requirements: Prior knowledge in Reinforcement Learning and Deep Learning is recommended.
Selection Procedure: Candidates are prioritised if they are present at the introduction event or write a motivational mail to both co-supervisors until Monday, April 24th 12am (updated !!!). We will also provide a recording of the introduction. After you been confirmed, you will then bid on a topic by ranking all the listed topics below, until Friday, May 26th.
Resources: Poster guideline as pptx or pdf



  1. RT-1: Robotics Transformer
  2. Hierarchical Policy Blending As Optimal Transport
  3. Mastering the game of Go with deep neural networks and tree search
  4. Mastering Atari, Go, chess and shogi by planning with a learned model
  5. SQIL: Imitation Learning via Reinforcement Learning with Sparse Rewards
  6. Segment Anything
  7. DayDreamer: World Models for Physical Robot Learning
  8. Real-time estimation of electric fields induced by transcranial magnetic stimulation with deep neural networks
  9. Maximum a Posteriori Policy Optimisation
  10. Is Conditional Generative Modeling all you need for Decision-Making?
  11. Efficient Online Reinforcement Learning with Offline Data
  12. Making Sense of Reinforcement Learning and Probabilistic Inference
  13. Formal Algorithms for Transformers