Date: Tuesday January 16, 2024 at 4:00 PM Eastern Time
Speaker: Ahmed Elmokadem, Ph.D., Senior Scientist II, Adjunct Professor
Affiliation: Metrum Research Group, University of Connecticut & Indiana University
Title: Rise of the Machines: Will Machine Learning Take Over the World of Pharmacometrics?
Abstract: The great advancement in machine learning in recent decades gave rise to the question: How can we bridge machine learning and mathematical modeling of disease and drug dynamics, that is pharmacometrics, to help in drug decision making? Many attempts have been made to build that bridge but still seems limited. This could be the result of the complexity of these applications that might render them impractical or the limited data available for the data-hungry machine learning techniques. In this talk, I will discuss two promising applications of machine learning: Hierarchical Deep Compartment Modeling (HDCM), which focuses on learning the functional relationship between covariates and compartmental model parameters, and Bayesian Deep Quantitative Systems Pharmacology (BaDQSP), which tries to learn missing dynamics within a QSP system of interest. These approaches alleviate the need of large data because of their focused targets and they are implemented as simple and intuitive algorithms within the Julia programming language.