Our research focuses on the identification and characterization of some novel statistical nature of modern complex systems that would lead to better modeling, prediction, and uncertainty quantification. We have studied complex systems that range from manufacturing, healthcare, and transportation. Examplary projects include

• Quality Science for Networked Systems: we develop methodologies for modeling, monitoring, anomaly detection, diagnosis, and prognosis of complex networked systems. These networked systems are continuously monitored by a distributed structure of sensors. The statistical uncertainty and interdependency of the nodes is complex and can’t be sufficiently characterized by classic multivariate statistical models. Examples of these networked systems are abundant in manufacturing, healthcare, and any application that implementes internet of things (IoT).
• Multi-modal Data Fusion: We develop machine learning models to integrate heterogeneous datasets such as neuroimaging, genomics, proteomics, laboratory tests, demographics, and clinical variables, for facilitating scientific discoveries in biomedical research and better decision making in clinical practices.
• User Behavioral Modeling: We develop user behavioral and choice-making models (i.e., under rewards) to characterize how users interact with an APP system and make decisions. We also develop methods to detect anomalous users (the so-called “bad actors”) from their behavioral patterns.
• Personalized Machine Learning Models: We develop the collaborative learning framework that can achieve multi-level characterization of a population: it builds a canonical structure shared by all individuals and the individual models together, extending the traditional mixed effect models to a more fine-grained statistical characterization of the population heterogeneity, while also borrowing strength from modern optimization tools to generate robust and generalizable personalized models even under sparse individual data and uneven distribution of samples across individuals.

Acknowledgment for Funding Support

• National Science Foundation (CMMI-1505260, CMMI-1536398, CMMI-1824623, CCF-1715027, CIS-2114260)
• Breakthrough T1D (formerly known as the Juvenile Diabetes Research Foundation)
• NIH
• AHRQ
• DARPA WASH; DARPA D3M
• AFOSR DDDAS
• USDOT
• Byrd Alzheimer’s Institute
• Royalty Research Foundation
• Helmsley Foundation
• Amazon
• Meta