Announcement Detail
Thursday, April 16, 2026
3pm EDT
Join via Zoom: https://us06web.zoom.us/j/
USACM UQ Virtual Seminar
Learning Stochastic Closure Models for Complex Dynamical Systems
Speaker
Jinlong Wu, University of Wisconsin-Madison
Abstract
Closure models are widely used in simulating complex multiscale dynamical systems such as turbulence and Earth’s climate, for which direct numerical simulation that resolves all scales is often too expensive. For those systems without a clear scale separation, deterministic and local closure models often lack enough generalization capability, which limits their performance in many real-world applications. In this talk, I will present a data-driven modeling framework for constructing stochastic and nonlocal closure models via generative AI, neural operators, and data assimilation techniques. In addition, an ensemble-based derivative-free optimization approach will be discussed to ensure that the modeled system not only mimics the short-term behaviors of the true system but also matches the long-term statistics. The results show that the proposed methodology provides a systematic approach via generative AI methods to constructing efficient and generalizable data-driven stochastic closure models for multiscale complex dynamical systems.