EVENT DETAILS
Title: Learning Structured Representations: From PDE Solution Operators to Constitutive Closures
Speaker: Dr. Bahador Bahmani, Northwestern University
Abstract: A central challenge in data-driven modeling of physical systems, particularly when observations are sparse or incomplete, is deciding what to learn and how to represent it. Inspired by classical numerical methods, approximation theory, and physical principles, my research explores how carefully formulated learning problems and structured, data-adaptive representations can improve computational efficiency, generalizability, and interpretability. In this talk, I will discuss two complementary directions. The first focuses on learning solution operators and dynamics of PDE-governed systems through continuous and latent representations that accommodate irregular spatial and temporal observations. The second focuses on discovering constitutive closures in continuum mechanics, where physically constrained representations facilitate the symbolic discovery of interpretable material laws. Together, these developments illustrate how encoding essential mathematical structures while learning the remaining complexity from data can provide a foundation for more principled scientific computing.
Bio: TBA
Zoom: https://northwestern.zoom.us/j/94084398588
---------
To subscribe to the Applied Mathematics Colloquia List send a message to LISTSERV@LISTSERV.IT.NORTHWESTERN.EDU with the command:
SUBSCRIBE esam-seminar FirstName LastName
TIME Tuesday December 1, 2026 at 11:00 AM - 12:00 PM
LOCATION M416, Technological Institute map it
ADD TO CALENDAR&group= echo $value['group_name']; ?>&location= echo htmlentities($value['location']); ?>&pipurl= echo $value['ppurl']; ?>" class="button_outlook_export">
CONTACT Natalie Masini natalie.masini@northwestern.edu
CALENDAR McCormick-Engineering Sciences and Applied Mathematics (ESAM)