CosmicAI Seminar: Surrogate Modeling — Omar Ghattas

Learned stand-ins for costly simulations

CosmicAI — the NSF-Simons AI Institute for Cosmic Origins, based at UT Austin — develops AI and machine learning methods for astronomy and cosmology. The first half of the fall seminar series concentrates on surrogate modelling of astrophysical systems: training models that reproduce the behaviour of physical simulations at a fraction of the computational cost, and characterising when they can and cannot be relied on.

This session is presented by Omar Ghattas, whose work sits at the intersection of computational science, inverse problems, and uncertainty quantification — a natural fit for the theme.

Practical details

• When: Wednesday, September 16, 1:00–2:00 p.m. Central
• Where: UT Austin, hybrid (in person plus online). The building and room are not published on the series page; check the CosmicAI events page before travelling, or attend online.
• Cost: Free, no ticketing — the organizers publish Google Calendar and ICS links rather than a registration form.

The specific talk title had not been announced when this listing was written. The date, time, and speaker are confirmed. The session is pitched at a research audience.