BIMSA >
BIMSA-Tsinghua Seminar on Machine Learning and Differential Equations
Data-Driven Modeling and Learning Using the Weak Form
Data-Driven Modeling and Learning Using the Weak Form
Organizers
Speaker
David Bortz
Time
Thursday, October 20, 2022 8:50 AM - 12:15 PM
Venue
1129B
Online
Zoom 537 192 5549
(BIMSA)
Abstract
Data-driven modeling approaches have proven highly successful in a wide range of fields in science and engineering. In this talk, I will discuss several ubiquitous challenges with the conventional model development / discretization / parameter inference / model revision loop (illustrating the issues with examples from Microbial Electrolysis Cells, chemotaxis, and HIV viral dynamics). I will present our Weak form Sparse Identification of Nonlinear Dynamics (WSINDy) framework which addresses several of these challenges and offers substantial advantages in terms of computational efficiency, noise robustness, and modest data needs (in an online learning context). Lastly, I will illustrate applications of WSINDy to several benchmark problems as well as to the inference of interacting particle systems, specifically learning anisotropic interaction rules from individual trajectories in a heterogeneous cellular population.
Speaker Intro
Prof. Bortz earned his PhD in 2002 with H.T. Banks at North Carolina State University. He was then a postdoc at the University of Michigan in Mathematics and joined the faculty in Applied Math at the University of Colorado in 2006. The core of his research interests are in scientific computation methodologies for data-driven modeling and inverse problems at the intersection of applied math and statistics. His research has received support from NSF, NIH, DOE, DOD, and NVIDIA.