Objective-Driven Optimal Experimental Design: Quantifying the Uncertainties that Matter and Reducing Them Efficiently and Effectively

Abstract: Modeling complex real-world systems involve immense uncertainties due to a wide variety of factors, including inherent stochasticity of the system, limited availability of data for modeling the system, presence of unobservable hidden variables, and measurement noise. In practice, these uncertainties may not be completely eliminated and difficult to reduce, requiring one to focus on the uncertainties that actually matter for achieving the modeling goal(s).

In this talk, we present how one may quantify the model uncertainty in an objective-driven manner and develop optimal experimental design (OED) techniques that can reduce the uncertainty that critically affects the operational goal(s) of the model. Furthermore, we discuss how one may leverage machine learning (ML) approaches to accelerate uncertainty quantification (UQ), and ultimately, OED. To demonstrate the advantages and potentials of these approaches, we will consider examples in systems biology, drug discovery, and material design.

Bio: Byung-Jun Yoon received his B.S. degree from the Seoul National University and M.S. and Ph.D. degrees from the California Institute of Technology, all in Electrical Engineering. Since 2008, he has been with the Department of Electrical and Computer Engineering, Texas A&M University, where he is currently a professor. Yoon holds a joint appointment at Brookhaven National Laboratory, where he is a Scientist in Computational Science Initiative, Applied Mathematics Group.