NHR PerfLab Seminar: Generative AI & Probabilistic Learned Solvers (hybrid)
Topic: Generative AI & Probabilistic Learned Solvers
Speaker: Prof. Nils Thürey, Technical University of Munich (TUM)
Date and time: Tuesday, November 11, 2025, 10:30 p.m. CET
Location: seminar room 02.049 (RRZE, Martensstraße 1, 91058 Erlangen)
Or online via Zoom: https://go-nhr.de/perflab-seminar
Abstract:
In this talk, I’ll explain into how generative AI methods—like diffusion models and flow matching—can be used to build improved, data-driven simulators. These models don’t just produce out a single best guess (the “mean”); they learn full probability distributions, letting us draw different samples and explore the range of possible outcomes. By combining these AI techniques with traditional numerical methods, we can even build powerful inverse solvers that are both fast and accurate. What’s especially exciting is that these probabilistic models can capture uncertainty and give us deeper insight into how the systems we model actually behave.
Short bio:
Nils Thürey is a professor at the Technical University of Munich (TUM) since 2013. He did his PhD at the FAU Erlangen-Nürnberg, and after a post-doc at ETH Zurich and time in industry (at Scanline VFX), he now focuses on deep-learning methods for physical systems, with a focus on fluids. Together with his research group, he targets new methods for the tight integration of AI and learning algorithms with classical numerical methods. Among others, this includes diffusion modeling, large-scale surrogates, and differentiable simulations.
For a list of past and upcoming NHR PerfLab seminar events, please see: https://hpc.fau.de/research/nhr-perflab-seminar-series/