All community members are invited to attend our 9th interview symposium featuring scientific applicant talks, several talks from our current scholars, a faculty keynote, and more.
Each year, IMPRS-IS hosts an interview symposium to interview and recruit new doctoral candidates.
The 2025 event will feature scientific talks from the applicants, as well as a keynote from Prof. Georg Martius from the University of Tübingen.
Date: Wednesday, February 19 Time: 15:00 - 16:00 CET Location: https://url.is.mpg.de/interviews-2025 Intrinsic Motivations in Reinforcement Learning
Abstract: I will present our recent work on developing intrinsically motivated exploration strategies for robots to playfully learn to manipulate objects and control high-dimensional musculoskeletal models of ostriches and humans. We will see that exploration is crucial for achieving good results. In the talk, I will present a line of research on model-based reinforcement learning, which enables zero-shot generalization to new tasks and is a promising route to efficient learning on real robots. The control of high-dimensional systems is relevant in robotics and in understanding human motor control. Recently, we have achieved learning of natural behavior in realistic musculoskeletal simulations using reinforcement learning with a novel exploration strategy. Besides explaining some of the technical details, I will also show you a few vivid and funny videos.
Bio: Georg Martius is a full professor at the University of Tübingen in Computer Science since 2023. From 2017 to 2024, he led a research group on Autonomous Learning at the Max Planck Institute for Intelligent Systems in Tübingen. Before joining the MPI in Tübingen, he was a postdoctoral fellow at the IST Austria and was a postdoctoral researcher at the Max Planck Institute for Mathematics in the Sciences in Leipzig. He received his Ph.D. from the University of Göttingen and his diploma in computer science from the University of Leipzig. His research focuses on machine learning for robotics, including internal model learning, reinforcement learning, intrinsic motivation, representation learning, differentiable combinatorial optimization, and haptics.
Other Event Details
A detailed event schedule is available to community members now from our interviews website: https://interviews.imprs.is.mpg.de/
All applicants, IMPRS-IS scholars, faculty, and associated faculty are granted access to the website link above. Since this is a closed event, only community members can view our interviews website and must first register for an account before gaining access.
For details about this event or questions about gaining access to the interviews website, please email us (imprs@is.mpg.de).
Photo credit: MPI für Intelligent Systeme
More information