Dr. Bing Yao
Assistant Professor
Industrial Engineering & Management
Oklahoma State University
Friday, March 4, 2022
2:00-3:00 John Tickle Building 500
ABSTRACT
Advanced sensing provides unprecedented opportunities for data-driven modeling, monitoring, and control of complex systems. Realizing the full potential of sensing data depends greatly on novel analytical methods for system informatics and decision making. My research objective is to integrate physics-based principles with advanced machine learning for complex system modeling and optimization. In this talk, I will present two of my recent works in physics-constrained machine learning to model space-time system dynamics. First, a physics-constrained deep learning (P-DL) framework is developed for robust inverse ECG modeling. This method integrates the physics law of the cardiac electrical wave propagation with deep learning to predict the spatiotemporal electrodynamics in the heart from body-surface sensor measurements. Second, I will present a physics-constrained deep active learning framework for spatiotemporal modeling of cardiac electrodynamics based on sparse sensor measurements. This method combines both the prediction uncertainty of the P-DL and the space-filling design over the 3D complex geometry to seek informative sensor observations for the robust modeling of space-time complex systems.
About the speaker:
Dr. Bing Yao is currently an assistant professor in the School of Industrial Engineering and Management at Oklahoma State University. She received her dual-title PhD degree in Industrial Engineering and Operations Research from the Pennsylvania State University. Her research focuses on developing innovative physical-statistical models for decision optimization in complex systems. This research has broad applications in both advanced manufacturing and healthcare. Her research contributions in this area have been recognized by multiple Best Paper/Poster awards in international research conferences such as the INFORMS annual meeting and IISE annual conference.