Institute Homepage
Institute Homepage DE Sign In

Back

Talk Upcoming Biography
11 December 2025 at 11:00 - 12:00 | MPI-IS Tübingen, N0.002

Special Talk: Towards Physical AI

ORGANIZERS
Thumb ticker sm 20241104 hardt moritz 12 cleaned kleiner
Social Foundations of Computation
Moritz Hardt
  • Director
Thumb ticker sm laemmerhirt eva 2 02
Scientific Coordination Office
Eva Lämmerhirt
Referentin der Geschäftsleitung | Institute Management Officer
Thumb ticker sm 20240912 tyagi nisha 1 3
Scientific Coordination Office
Nisha Tyagi
Thumb ticker xxl website banner st andreas g

Physical AI is considered the next major frontier of AI. It refers to artificial intelligence embodied in physical systems that can sense, learn and act autonomously in the real world, bridging perception, cognition and action. In the first part of this talk, I will argue that spatial perception is relevant for Physical AI systems, but doesn’t necessarily emerge on its own. I will review some of our work that led to the revolution of coordinate-based 3D reconstruction and novel view synthesis methods. In particular, I will show recent optimization-based and feed-forward methods for both tasks. In the second part, I will demonstrate progress on a specific instance of Physical AI which has always fascinated me: Self-Driving Cars. In particular, I will review progress on imitation and reinforcement-based methods and demonstrate the relevance of open-loop and closed-loop benchmarks. In the final part of my talk, I will share with you my personal lessons learned from 17 years of research in AI as well as my predictions about the future.

Speaker Biography

Andreas Geiger (University of Tübingen)

Group Leader and Head of the Department of Computer Science

Andreas Geiger is a full professor at the University of Tübingen and the Tübingen AI Center. He currently serves as the head of the department of computer science. Prior to 2018, he was a visiting professor at ETH Zürich and an independent W2 group leader at the MPI for Intelligent Systems, Tübingen. He studied at KIT, EPFL and MIT, and received his PhD degree in 2013 from KIT. His research interests are at the intersection of computer vision, machine learning, NLP and robotics. He has created some of the most influential benchmarks in computer vision, including KITTI and KITTI-360. He has been an ELLIS founding board member and created the ELLIS PhD and PostDoc program, the largest and most successful AI/ML network in Europe. His work has been recognized with the Longuet-Higgins Prize, the Mark Everingham Prize, the IEEE PAMI Young Investigator Award, the Heinz Maier Leibnitz Prize and the German Pattern Recognition Award. He received 7 best paper awards and 2 ERC grants.