The Systems Medicine Seminar welcomes Logan Rose, hosted by Dr. Luis Sordo Vieira.
Presenter: Logan Rose, Graduate Student, Teaching Assistant
Affiliation: Mathematics Department, University of Kentucky
Title: Parameter Personalization of Medical Digital Twins
Abstract: Medical digital twins (MDT) are virtual models of disease that are continuously updated using longitudinal patient data. Effective MDTs must be adaptable to specific patient needs and recoverable from longitudinal measurement data. We develop novel methods to recover and personalize parameters of models relevant to MDTs. In doing so, we decompose selected model parameters into a common and personalized component. The common component represents a commonality between patients and is estimated once, while the personalized component is patient specific and may be updated with new measurement data. Given a set of synthetic trajectories from known dynamical systems (i.e. Lotka-Volterra, Hoffman) and a library of basis functions, we recover an estimated model and set of parameters, which we take as the common components. Then, given a new set of measurement data, we estimate personalized components for each using the common component estimates and the well-known Adam algorithm. Finally, we apply our methods to AML (Acute Myeloid Leukemia) measurement data.
Date: Tuesday, January 27, 2026, at 4:00 PM Eastern Time
Venue: This speaker will be conducting the seminar virtually via Zoom. Join by Zoom or in MSB Room M-434, the LSM Conference Room.