Seminar: Explainable Reinforcement Learning on GPUs

In this seminar, you will learn the basics of GPU programming and its application to reinforcement learning, specifically Deep Q‑Learning. You will also become familiar with explainability in neural networks, which is crucial for understanding and interpreting decisions made by AI systems. You will be assigned a scientific paper to present, covering one of the topics listed below (GPU programming, reinforcement learning, or explainability). In addition to the presentations, you will work on a practical project to implement and apply the concepts you've learned. The project culminates in a final presentation in which you showcase your work and findings, with a specific focus on your assigned topic.

• Offered by: UMTL (Chair of Prof. Dr. Antonio Krüger)
• Lecturers: Julian Groß, Gian‑Luca Kiefer
• Location: DFKI, D3 2 Room -2.07 (Barwise)
• Time: Tuesdays, 14:15–16:00 (CET/CEST)
• ECTS credits: 7
• Language: English
• Seats: 10

 

Important Dates

 

Note: Dates and topics are subject to change. Updates will be sent via email and posted on the seminar website.

 

Paper Presentation

At the beginning of the seminar, you will be assigned one scientific paper in GPU programming, reinforcement learning, or explainability. Your presentation should last 20 minutes, followed by a 10‑minute discussion.

 

Project Work

During the seminar, you will work on a practical project to explore GPU programming, reinforcement learning, and explainability. You are expected to work independently; your submitted code will be checked for plagiarism. You will give a 10‑minute project presentation focusing on the topic of your key takeaways.

Paper Topic Preferences

You can give your paper topic preferences here until 22.10.2025, 2pm (14:00). The next day, you will get your assigned topic by mail.

Grading

• Paper presentation: 40%
• Practical project: 50%
• Project presentation: 10%

 

Passing the Seminar

You must achieve at least 50% in each component (presentations and practical project). Each component is graded independently and must be passed separately.

 

Attendance and Submissions

Attendance at all meetings is expected; exceptions require an official document (e.g., a doctor’s certificate). You must submit the current state of your practical project each week. The submission deadline is every Tuesday at 13:59 (CET/CEST). Plagiarism is taken very seriously. If identical or shared code is submitted, only one submission may be graded; other students will receive 0 points for that submission.

 

Requirements

You do not need a physical GPU device to attend and pass the seminar. We provide a software simulator and access to a hardware evaluation machine.

You should have passed Programming II, the Software Praktikum, Concurrent Programming (Nebenläufige Programmierung), and Mathematics for Computer Scientists I–II (or Analysis I and Linear Algebra I). Passing the Software Engineering core lecture and a course on Machine Learning or Neural Networks is a plus. You are expected to be fluent in a C‑like programming language.

We will use C# in this seminar; you are expected to acquire basic familiarity before the seminar starts. We will use ILGPU to develop all GPU programs. Feel free to browse the website and samples to get an early overview.

 

Submission Information

• Please use the subject prefix “[XRL Seminar]” for all emails concerning this seminar (e.g., “[XRL Seminar] Question about XYZ”).
• Please use the subject prefix “[XRL Seminar] Submission” for all submissions (e.g., “[XRL Seminar] Submission: Topic 1”).
• Please always send emails to both lecturers: julian.gross@dfki.de and gian-luca.kiefer@dfki.de.

 

Julian Groß, Gian-Luca Kiefer, Prof. Dr. Antonio Krüger