*Presentation will be recorded.
Abstract:
Technological advancements in diagnostic imaging, smart sensing, and health information systems have resulted in a big data environment for Precise Medicine and Population Health. My research lies in the intersection of statistical modeling, machine learning, and healthcare data science in the patient care cycle from screening, diagnosis, prognosis, treatment, monitoring, and care, to system-level decision-making. In this talk, I will present three topics that tackle data science challenges in phenotype discovery, privacy-preserving telemedicine, and preventative care, respectively. The first topic introduces a novel Multi-modal Mixed-type Factor Mixture Model for phenotype discovery from multi-faceted medical data, coupled with an efficient Gauss-Hermite Expectation-Majorization-Maximization (GH-EMM) algorithm that uses the Gauss-Hermite Quadrature to provide a tractable solution and integrates an efficient Majorization Maximization algorithm for optimization. This work has been employed in both neurological and cardiometabolic diseases to enable refined disease classification and phenotype-optimized treatment and intervention. The second topic presents a novel federated Gradient Boosting algorithm with the Least Squares Approximation (fed-GB-LSA) that leverages functional data for automated diagnosis and telemedicine of obstructive sleep apnea with privacy preservation. The fed-GB-LSA provides a one-shot federated learning approach that is proven to enjoy communicational and statistical efficiency. The last topic presents a novel spatially-constrained multi-task Linear Mixed Model (spatial-multi-LME) that identifies system-wide facilitators and barriers for multi-task care activities. This work facilitates the development of cost-effective interventions that eventually promote the uptake of preventative screenings and improve the quality of life. Theoretically, the spatial-multi-LME estimator enjoys the oracle property. I will conclude this talk by briefly going over my other research efforts and plans.
Bio:
Bing Si is an Assistant Professor in Systems Science and Industrial Engineering at the State University of New York (SUNY) at Binghamton. She received her Ph.D. in Industrial Engineering from Arizona State University in 2018. She received her B.S. in Mathematics from the University of Science and Technology of China and an M.S. in Industrial Engineering from ASU in 2012 and 2014, respectively. Dr. Si’s research focuses on data fusion and statistical machine learning for complex-structured heterogeneous data in healthcare and medicine, in collaboration with Mayo Clinic, Brigham and Women’s Hospital, and Colleges of Nursing at the University of Rhode Island and SUNY Binghamton. Her research is sponsored by both industry and federal agencies including NIH and AHRQ with R03, R21, and R01. She is a recipient of multiple awards such as Early-Stage Distinguished Research Award (2023), Million Dollar Award (2023), ISE Featured Article (2017, 2023), Grace Hopper Faculty Scholarship (2020), Dean’s Dissertation Award from ASU (2017), Outstanding Emerging Fulton Student Organization Leader (2017), and IISE Healthcare Student Best Paper Finalist (2017). She is a member of IISE, INFORMS, IEEE, and SRS. Dr. Si is a Guest Editor for Mathematics and an Associate Editor for IISE Transactions on Healthcare Systems Engineering.