Interacting Dynamical System Modeling for Science: Construction and Generalization

Abstract: Many real-world systems such as disease transmission, molecular dynamics, and spring systems can be considered as multi-agent dynamical systems, where multiple objects interact with each other and exhibit complex behavior along the time. In this talk, I will discuss my current research on interacting dynamics system modeling for scientific problems, especially focusing on model construction and model generalization. I will begin by discussing my work on graph ODEs for efficiently capturing continuous high-order correlations. Then, I will discuss different types of distribution shifts in dynamical system modeling and how to address them to improve the generalization ability. Finally, I will introduce future research directions in the field of dynamical system modeling.
Speaker: Xiao Luo, Ph.D.
IDRE Fellow
Department of Computer Science
University of California Los Angeles Time: 11:30 AM - 12:30 PM (PST)
Date: March 28, 2024