Abstract

This thesis, titled “Data-Driven Control and Dynamic Learning of Population Systems using Moment Representations”, presents innovative machine learning techniques for controlling and learning complex biological and computational systems characterized by inherent diversity and dynamics. The work explores a novel reinforcement-based ensemble control strategy for both labeled and unlabeled systems (Chapter 2), and advances moment-based methods for learning both static and dynamic data (Chapter 3), including a new framework for multivariate time series data. A significant portion of the thesis (Chapter 4) is dedicated to the application of these techniques to real-world medical datasets, specifically targeting Monoclonal Gammopathy of Undetermined Significance (MGUS) and Multiple Myeloma (MM). Overall, the thesis offers advanced data handling methods through moment representations and validates their practical utility in not only the control of complex systems, but also the medical domain for machine-learning-assisted medical decision-making.

Committee Chair

Jr-Shin Li

Degree

Doctor of Philosophy (PhD)

Author's Department

Electrical & Systems Engineering

Author's School

McKelvey School of Engineering

Document Type

Dissertation

Date of Award

Summer 9-14-2023

Language

English (en)

Available for download on Thursday, September 07, 2028

Share

COinS