Ryan Urbanowicz, MSE, PhD

Dr. Urbanowicz’s research focuses on the development, evaluation, and application of machine learning (ML) and artificial intelligence (AI) methods for the analysis of biomedical and clinical data. This work is motivated by challenges such as the need for interpretable models, tackling large-scale data analyses, detecting and characterizing complex patterns of association (including epistasis, genetic heterogeneity, and rare variation), noisy and incomplete data with uncertainty, adjusting for covariates, and accounting for bias/fairness. Recent work has focused on adopting best practices in data science to automated machine learning (AutoML) via the Simple, Transparent, End-to-end Machine Learning Analysis Pipeline (STREAMLINE) towards accessible, rigorous, reproducible, and transparent analytics. Additionally, his lab developed and is expanding (1) FIBERS, an evolutionary algorithm approach to feature learning and risk stratification, (2) ExSTraCS, an interpretable rule-based machine learning algorithm for detecting and characterizing epistatic and heterogeneous associations, (3) ReBATE, a suite of feature selection algorithms able to account for epistatic interactions, and (4) GAMETES, software for simple and complex genetic dataset simulation. Application areas include kidney transplantation, cancer, congenital heart disease, obstructive sleep apnea, and Alzheimer’s disease.

Dr. Urbanowicz is one of the few international experts on rule-based ML algorithms and co-authored a book introducing Learning Classifier Systems in 2017. He has published over 60 peer-reviewed papers, reviews, book chapters, and editorials, mentored over 90 high school, undergraduate, and graduate students on research projects, directs the Cedars-Sinai AI Campus program, and creates educational ML and AI content posted to his lab’s YouTube channel (The URBS-Lab).