STALDYS · Statistical Learning for Dynamical Systems

The proposed research aims to establish groundbreaking new computational methods for the analysis of deterministic, control and random dynamical systems by using tools from the field of Statistical Learning Theory (a.k.a. Machine Learning). These three related system types constitute a very active research area with a wide range of applications, including climate studies, power systems, autonomous vehicle control, aircraft control, robotics, finance and chemical engineering. In many applications the understanding of such systems requires the availability of efficient, high-performance algorithms and a key innovative aspect of this proposal is the development of a unified theory for computational dynamics that has the potential to establish a new field at the intersection of machine learning and dynamical systems theory.