Sie sind hier: Startseite › Teaching › WS2026/27 › Reinforcement Learning Lab

Reinforcement Learning Lab

General Information

  • HISinOne: 11LE13P-7320
  • Kickoff meeting: TBA
  • Location: Intelligent Machine-Brain Interface Technology (IMBIT), Nexus Lab 
  • Further Informations, exercises and solutions will be posted on ILIAS.
  • Language: English
  • Email:

Overview

The Reinforcement Learning Lab (not to be confused with the Deep Learning Lab) is a practical course in which students  learn to program their own deep reinforcement learning (DRL) agents using state-of-the-art algorithms such as Deep Q-Learning (Mnih et al., 2013) and Soft Actor-Critic (Haarnoja et al., 2018). Students implement these methods in Python and PyTorch during regular exercises and later apply them to a self-chosen DRL problem in small groups. Finally they will present their results.

Learning Objective 

Students gain hands-on experience implementing, training, and evaluating deep reinforcement learning methods.

Format

  • Exercise Phase: Students work individually on exercises that cover important concepts and methods in DRL. These build on each other and prepare students for the project phase. This structure may change as AI coding tools, such as Large Language Models (LLMs), become more commonly used
  • Project Phase: In groups of three to four, students select their own DRL problem, apply and extend methods from the course, and present their results in an in-person final presentation


All important materials will be made available on our ILIAS course.

Prerequisites

Students should have attended the Reinforcement Learning lecture and have experience programming in Python. Experience with machine learning, deep learning, and PyTorch is recommended.