MS 6: Machine Learning and Simulation Science


Organizers & Moderation
Gael Rigaud
Mathias Niepert

The goal of the project network is the development and integration of methods and algorithms from the field of machine learning for the problem of solving problems in the simulation sciences, and to learn from the simulation sciences to build better machine learning models. Important examples are the design of surrogate models for complex physical models from fluid mechanics or structural mechanics. The Machine Learning and Simulation Science mini-symposium aims to provide an informal, inter-disciplinary, and leading-edge venue for research and discussions at the interface of machine learning (ML) and the simulation sciences. This interface spans (1) applications of ML in simulation science (ML for Sim), (2) developments in ML motivated by insights from the simulation sciences (Sim for ML), and most recently (3) convergence of ML and simulations sciences (Sim with ML) which leads to questions about what scientific exploration and understanding are in a world where AI and ML-supported science becomes more ubiquitous.

Thursday, 5 October 2023, 4:00-5:30 pm

Learning under requirements

Luiz Chamon  
(University of Stuttgart)

4:00 - 4:15 pm

Active Learning and Enhanced Sampling for Developing Interatomic Neural Network Potentials Viktor Zaverkin
(University of Stuttgart)
4:15 - 4:30 pm
Learning Physical Law Reliably - CANCELLED Gitta Kutyniok
(LMU Munich)
4:30 - 4:50 pm
ClimaX: A foundation model for weather and climate Johannes Brandstetter
(JKU Linz)
4:50 - 5:10 pm
PDEBENCH: An Extensive Benchmark for Scientific Machine Learning Makoto Takamoto
(NEC Laboratories Europe GmbH)
5:10 - 5:30 pm
To the top of the page