ChBE Fall Seminar: Scientific Machine Learning for Knowledge Discovery

Time

-

Locations

Perlstein Hall, Room 108 10 West 33rd Street Chicago, IL 60616

Social Media Links

Fall 2022 Seminar Series

Co-Sponsored by the Center for Complex Systems and Dynamics

Ming Zhong

The Department of Chemical and Biological Engineering at Illinois Institute of Technology will host Ming Zhong for a lecture titled “Scientific Machine Learning for Knowledge Discovery” on Wednesday, October 5, from 3:15–4:30 p.m. in Room 108 of Perlstein Hall.

Abstract

Identifying the driving force for dynamical motion (i.e., planetary motion) or the leading cause for infectious disease (John Snow’s Grant Experiment in the 1800s) from data has been a crucial part of the scientific development of human knowledge. As observation/sensing techniques has boomed in the recent years, how to make scientific discoveries from large data set has become a great challenge. We propose several scientific machine learning (SciML) techniques as a modern mathematical tool for knowledge discovery. We discuss two major applications of SciML, one is to use knowledge-based statistical learning to discover dynamical models for modeling self organization (clustering, flocking, swarming, etc.), and the other is to use physics informed machine learning to solve nonlinear stiff and hyperbolic PDEs (Burgers, Allen Can, Euler Equations, etc.) related to forward and backward problems. We will discuss both theoretical advancements and computational challenges in these applications.

Biography

Ming Zhong is an assistant professor in the Department of Applied Mathematics at Illinois Institute of Technology. He was an assistant research scientist in the  Institute of Data Science at Texas A&M University before he moved to Illinois Tech. He obtained his Ph.D. in applied mathematics under the guidance of Eitan Tadmor at the University of Maryland. His research interests include scientific machine learning, inverse problems, and numerical ODE/PDE. In particular, he focuses on developing efficient and effective physics-informed machine learning algorithms to derive physical models from observation data.

Event Contact

Getting to Campus