Invited Session Series on Learning-based Control

Organizers:
external page Angela P. Schoellig (University of Toronto, Canada)
external page Sebastian Trimpe (RWTH Aachen University, Germany)
Melanie N. Zeilinger (ETH Zürich, Switzerland)
external page Matthias Müller (Leibniz University Hannover, Germany)

Data science and machine learning have demonstrated tremendous success in the last decades in applications such as image recognition, recommender systems, or question answering. Compared to these applications, which can largely be subsumed as static problems, learning-based control systems take a special role within the world of statistical and machine learning. The coupling of a learning algorithm with a control loop requires a combined treatment as a dynamic process and raises fundamental questions about stability, robustness, and safety, which are generally less critical in most traditional application areas of machine learning. In order to leverage the potential of data-based and learning methods for control, we therefore believe that principled approaches integrating machine learning and control theory are needed, which extend beyond the methods and tools of the individual disciplines.

Based on the increasing interest in this domain, we have started a new Invited Session Series on Learning-based Control taking place yearly at the IEEE Conference on Decision and Control (CDC). The sessions in 2016 through 2023 were a great success and we are looking forward to your contributions and the next session at CDC'24. Please do not hesitate to contact us if you have any questions.