I am a Professor in the Department of Industrial and Systems Engineering at University of Washington. I obtained B.S. degree in Industrial Engineering from Seoul National University in 2001. I received Ph.D. degree in Industrial Engineering from Texas A&M University, College Station, TX in 2011 under the guidance of Yu Ding in Industrial and Systems Engineering and Jianhua Huang in Statistics at Texas A&M University. In Fall 2011, I joined the Department of Industrial and Manufacturing Engineering at Florida State University as an Assistant Professor. I became an Associate Professor with tenure in 2017 and was promoted to a Professor in 2023. I moved to University of Washington in 2024, establishing Digital Transformation Lab.

My primary research interests lie in machine learning and digital twins, with applications in advanced manufacturing and physical sciences. A central goal of my research is to develop digital twin models that remain both computationally efficient and closely synchronized with their physical counterparts. To achieve this, my work focuses on surrogate modeling, online model calibration, and data-driven synchronization of physical and computational systems. 

One major research direction is surrogate modeling for physical and computational experiments. Surrogate models are used to replace computationally expensive simulations and to characterize discrepancies between simulation outputs and real-world observations, enabling more accurate and adaptive digital twins. Gaussian Processes (GPs) have been a standard framework for surrogate modeling, but conventional GPs often struggle with nonstationarity, regime changes, discontinuities, and categorical inputs that commonly arise in digital twin applications. To address these limitations, my group develops alternative surrogate modeling frameworks, including new GP constructions, deep learning approaches, and deep learning–GP hybrids. Representative examples include the Jump Gaussian Process and Deep Jump Guassian Process models. We also study active learning and optimal experimental design methods for efficiently constructing high-fidelity surrogates, such as Active Learning of Jump Gaussian Process Surrogates. 

Another major focus is the synchronization of digital twins with evolving physical systems. Traditional Bayesian model calibration provides a principled framework for aligning computational models with physical observations, but it is typically formulated as an offline procedure under stationary assumptions. Data assimilation methods support online model updates, yet they are often less effective under abrupt system changes and do not explicitly account for systematic model bias. Our group develops methodologies for continuously synchronizing digital twin models under both gradual system drift and abrupt regime changes while explicitly estimating and correcting model bias. This direction is summarized in our recent work, Online Bayesian Calibration under Gradual and Abrupt System Changes.  

I am also focused on the modeling and analysis of object data, which involves the statistical study of complex structures and their visual features—such as images, shapes, motion, functions, and directions. Unlike traditional numerical data, object data are typically non-Euclidean, meaning that conventional statistical tools developed for Euclidean (or vector) spaces are not directly applicable. Research in this area involves defining appropriate probability spaces for object data and developing corresponding statistical inference algorithms. Recently, our efforts have centered on modeling and analyzing human performance (and potentially robotic performance, robot/human interactions). Using a shape-theoretic approach, we model human performance as a time series of body postures. Our initial work in this direction is presented in the paper, Data Science for Motion and Time Analysis in Operations Research, with more studies currently underway. Many of my methodological studies in this direction are also motivated by the problem of understanding processing-structure-property relations in manufacturing and physical sciences. In 2021, I authored the book Data Science for Nano Image Analysis, which summarizes many of my works in these areas. 

1. Xu, Y., and Park, C. (2026). Online Bayesian Calibration under Gradual and Abrupt System Changes. (Pre-Print).

2. Chen, Y., Park, C., and Srivastava, A. (2025) Statistical Emulations of Human Operational Motions in Industrial Environments. (Pre-Print).

3. Xu, Y., and Park, C. (2025). Deep Jump Gaussian Processes for Surrogate Modeling of High-Dimensional Piecewise Continuous Functions. (Pre-Print).

4. Park, C., Waelder, R., Kang, B., Maruyama, B., Hong, S., and Gramacy, R. (2026). Active Learning of Piecewise Gaussian Process Surrogates. Technometrics. 68(1):186-201. (Paper).

5. Park, C. (2022) Jump Gaussian Process Model for Estimating Piecewise Continuous Regression Functions. Journal of Machine Learning Research. 23(278):1−37 (Paper) 

6. Park, C., Noh, S., & Srivastava, A. (2022). Data Science for Motion and Time Analysis with Modern Motion Sensor Data. Operations Research. 70(6):3217-3233 (Paper) 

7. Park, C. and Ding, Y. (2021) Data Science for Nano Image Analysis. Springer Nature. ISBN 978-3-030-72821-2 (Paper)

8. Park, C., & Apley, D. (2018) Patchwork Kriging for Large-scale Gaussian Process Regression. Journal of Machine Learning Research. 19(7): 1-43 (Paper) 

9. Park, C., Woehl, T. J., Evans, J. E., & Browning, N. D. (2015). Minimum Cost Multi-way Data Association for Optimizing Multitarget Tracking of Interacting Objects. IEEE Transactions on Pattern Analysis and Machine Intelligence, 37(3), 611-624 (Paper)

10. Park, C. (2014). Estimating Multiple Pathways of Object Growth using Nonlongitudinal Image Data. Technometrics, 56(2), 186-199 (Paper) 

11. Park, C., Huang, J. Z., Ji, J., & Ding, Y. (2013). Segmenting, Inference and Classification of Partially Overlapping Nanoparticles. IEEE Transactions on Pattern Analysis and Machine Intelligence, 35 (3), 669-681 (Paper) 

12. Park, C., Huang, J. Z., & Ding, Y. (2010). A Computable Plug-in Estimator of Minimum Volume Sets for Novelty Detection. Operations Research, 58(5), 1469-1480 (Paper)