INFORMS Open Forum

Abstract: Surrogate modeling has become a fundamental tool for the design and analysis of computer experiments, the construction of adaptive digital twins, and physics-informed machine learning. This webinar introduces a new class of surrogate models for non-stationary systems, namely Jump Gaussian Process (Jump GP) surrogates, which are designed to model piecewise continuous response functions that remain smooth within regions while allowing discontinuities across the design space. Such discontinuities naturally arise in many engineering and scientific applications, including advanced manufacturing, autonomous materials discovery, and cyber-physical systems. The talk will present our research on piecewise continuous surrogate modeling, spanning the development of the original Jump GP framework, its extensions to deep Gaussian processes for high-dimensional problems, and active learning strategies for efficient sequential data collection. The proposed methods are supported by both theoretical analysis and extensive empirical studies on synthetic benchmarks, simulation-based examples, and real-world applications in smart manufacturing and autonomous materials design. 

Biographical Sketch: Dr. Chiwoo Park is a Professor in the Department of Industrial and Systems Engineering at the University of Washington. He received his B.S. in Industrial Engineering from Seoul National University in 2001 and his Ph.D. in Industrial Engineering from Texas A&M University in 2011. Before joining the University of Washington in 2024, he was a Professor in the Department of Industrial and Manufacturing Engineering at Florida State University. Dr. Park's research lies at the intersection of machine learning, statistics, and industrial engineering, with a focus on developing data-driven methodologies for advanced manufacturing, digital twins, and scientific machine learning. His contributions span surrogate modeling, Bayesian sequential experimental design, shape data analysis, and digital twin technologies for cyber-physical systems. His research has been recognized through several prestigious honors, including the Ralph E. Powe Junior Faculty Enhancement Award and the Brainpool Fellowship.