Upcoming Event: Oden Institute Seminar

Data Driven Modeling for Scientific Discovery and Digital Twins

Dongbin Xiu, Department of Mathematics, The Ohio State University

Abstract

We present a data-driven modeling framework for scientific discovery, termed Flow Map Learning (FML). This framework enables the construction of accurate predictive models for complex systems that are not amenable to traditional modeling approaches. By leveraging measurement data and the expressiveness of deep neural networks (DNNs), FML facilitates long-term system modeling and prediction even whengoverning equations are unavailable.

FML is particularly powerful in the context of Digital Twins, an emerging concept in digital transformation. Withsufficient offline learning, FML enables the construction of simulation models for key quantities of interest (QoIs) in complex Digital Twins, even when direct mathematical modeling of the QoI is infeasible. During the online execution of a Digital Twin, the learned FML model can simulate and control the QoI without reverting to the computationally intensive Digital Twin itself.

As a result, FML serves as an enabling methodology for real-time control and optimization of the physical twin, significantly enhancing the efficiency and practicality of Digital Twin applications.

Biography

Dongbin Xiu received his Ph.D. in Applied Mathematics from Brown University in 2004. He joined theDepartment of Mathematics at Purdue University in 2005 and moved to the University of Utah in 2013 as a Professor in the Department of Mathematics and the Scientific Computing and Imaging (SCI) Institute. In 2016, he joined The Ohio State University as a Professor of Mathematics and an Ohio Eminent Scholar.

Dr. Xiu received the NSF CAREER Award in 2007 and was elected a SIAM Fellow in 2023. He has served on the editorial boards of numerous journals and is currently the Editor-in-Chief of the Journal of Computational Physics. Additionally, he is the founding Associate Editor-in-Chief of the International Journalfor Uncertainty Quantification (IJUQ) and the founding Editor-in-Chief of the Journal of Machine Learning for Modeling and Computing (JMLMC).

His research focuses on developing efficient numerical algorithms for data-driven modeling, scientific machine learning, Digital Twins, and uncertainty quantification.