Sebastian Kaltenbach, M.Sc.
Technical University of Munich
Associate Professorship of Continuum Mechanics (Prof. Koutsourelakis)
Postal address
Postal:
Boltzmannstr. 15
85748 Garching b. München
- Phone: +49 (89) 289 - 16692
- Office hours: Dienstag, 10-11.30 Uhr
- Room: 5504.EG.438
- sebastian.kaltenbach@tum.de
Raum MW 0438
Research Focus
- Physics-aware Machine Learning
- Uncertainty Quantification
- Bayesian Strategies
- Coarse Graining
- Stochastic Partial Differential Equations
- UQ in multiscale problems (e.g. biomedical systems)
Publications
- Sebastian Kaltenbach, Phaedon-Stelios Koutsourelakis: Incorporating physical constraints in a deep probabilistic machine learning framework for coarse-graining dynamical systems (Journal of Computational Physics)
- Sebastian Kaltenbach, Phaedon-Stelios Koutsourelakis: Physics-aware, data-driven discovery of slow and stable coarse-grained dynamics for high-dimensional multiscale systems (NeurIPS Workshop Paper)
- Sebastian Kaltenbach, Phaedon-Stelios Koutsourelakis: Physics-aware, probabilistic model order reduction with guaranteed stability (ICLR Conference Paper)
Conference contributions
- SIAM CSE 2019: Physics-aware and Sparse Coarse-graining of Multiscale Dynamics
- SIAM UQ 2020: Deep probabilistic learning of reduced dynamics of multiscale systems in the Small Data regime
- WCCM & ECCOMAS 2020: Physics-aware, deep probabilistic modeling of multiscale dynamics in the Small Data regime
- SIAM CSE 2021: Deep Probabilistic Coarse-Graining of Multiscale Dynamical Systems with Long-Term Stability
- ICLR 2021: Physics-aware, probabilistic model order reduction with guaranteed stability
Seminars and Workshops
- Computational Statistics and Data-Driven Models, ICERM Virtual Workshop, 2020, Poster Presenter
- Deep Learning and Reinforcement Learning Summer School 2020, CIFAR & Mila
- Interpretable Inductive Biases and Physically Structured Learning, NeurIPS 2020
Supervised Student Projects
- Jonas Eichelsdörfer: Physics Informed Machine Learning and Hamiltonian Neural Networks, Master's Thesis (supervised together with Atul Agrawal)
- Zhiyi He: Bayesian GANs for medical image analysis, Master's Thesis (supervised together with Hongwei Li)
- Simon Jarvers: Machine Learning of ODE with Gaussian Processes, Bachelor's Thesis
- Tim Beyer: Neural Ordinary Differential Equations for Physical Problems, Bachelor's Thesis
- Tobias Pielok: Residual Enhanced Probabilistic Koopman-based Representation Learning, Master's Thesis
- Martin Kronthaler: Physics Enhanced Neural Networks for the Prediction of Dynamical Systems, Term Project
- Zhiyi He: Magnetic Resonance Images Reconstruction using Generative Adversarial Networks and Uncertainty Analysis, Term Project (supervised together with Hongwei Li)
- Tobias Pielok: Bayesian Coarse Graining with Memory, Term Project
- Fan Wang: Learning evolution laws from complete and incomplete data, Research Internship