Collective Behaviour - Winter Seminar Series 2021/22
Machine learning for dynamical systems
Speaker:
Dr. Tobias Sutter, Computer Science Department, University of Konstanz
This event is part of an event series „CASCB Seminar Series“.
Understanding combined decision making & cultural transmissions through a reinforcement learning framework
Given the recent progress in information technology with real-time data being available at large scale, many complex tasks involving dynamical environments are addressed via tools from machine learning, control theory and optimization. While control theory in the past has mainly focused on model based design the advent of large scale data sets raises the possibility to analyse dynamical systems on the basis of data rather than analytical models. From a machine learning perspective, one of the main challenges going forward is to tackle problems involving dynamical systems which are beyond static pattern recognition problems. In this talk, I will give an overview about different problems lying in this intersection of dynamical systems, learning and control that I have worked on in the past. One such problem is inverse reinforcement learning, which basically aims to explain observed behaviour (e.g., of an animal or human) by inferring the underlying cost function describing the observed behaviour as a Markov Decision Process.
Tobias Sutter received a B.Sc. and M.Sc. degree in Mechanical Engineering in 2010 and 2012 from ETH Zürich, and a Ph.D. degree in Electrical Engineering at the Automatic Control Laboratory, ETH Zürich in 2017. He currently is an Assistant Professor at the Computer Science Department in Konstanz, Germany. Prior to joining University of Konstanz, he held a research and lecturer appointment with EPFL at the Chair of Risk Analytics and Optimization and at the Institute of Machine Learning at ETH Zürich. His research interests revolve around control, reinforcement learning and data-driven robust optimization. He was a recipient of the 2016 George S. Axelby Outstanding Paper Award from the IEEE Control Systems Society and received the ETH Medal for his outstanding Ph.D. thesis on approximate dynamic programming in 2018.
Datum: 2021-12-20