Dr. rer. nat. Daniel Lehmberg
Technische Universität München
TUM School of CIT
Department of Computer Science
Office: MI 02.05.042
Email: d.lehmberg (at) tum.de
Office Hours: by arrangement (email)
- data-driven modeling
- operator-based machine learning (Koopman and Laplace-Beltrami operators)
- dynamical systems
- scientific software
I am maintainer and developer of the open source Python software datafold.
A main objective of datafold is to transfer methods that are based on operator theory to a machine learning perspective. With datafold you can, for example, extract infromative geometric coordinates for nonlinear dimension reduction or build predictive models from high-dimensional time series.
The software is published at the Journal of Open Source Software (JOSS).
List available at ORCID.