November 14, 2025
Li Ma, AB’06, SM’06, is a professor of Statistics and Data Science at the University of Chicago. His research focuses on generative models, Bayesian modeling and computation, nonparametric inference, tree-based methods, and their applications in biomedical science. Prior to joining UChicago, he was a faculty member in the Department of Statistical Science at Duke University from 2011 to 2025. He is a Fellow of the International Society for Bayesian Analysis and a Fellow of the American Statistical Association.
What inspired you to pursue research in your specific field?
I love learning statistics, and so I want to keep learning more statistics my whole life. During my senior year in college here at UChicago, when I was fantasizing about a few possible “glorious” career paths, or more precisely a bit clueless about what I should be doing next, Professor Hugo Sonnenschein guided me toward understanding that I shouldn’t do something only because others aspire to do it. This made profound impacts on me: first, I realized how much I enjoyed statistics out of everything I learned in college, and second, Professor Sonnenschein's example made me really want to be an educator just like him, one who not only does world-class research but truly cares about the student. This leads naturally to my choice as an academic statistician.
What overarching questions are you trying to answer with your research?
Over the past century, statistics has profoundly influenced nearly every scientific discipline that involves the collection and analysis of data. The greatest statisticians that I admire have always drawn motivation from the pressing data challenges of their time. Today, with the rise of generative AI and emerging data technologies, new challenges continue to arise—and this is an exciting time to be a statistician. I hope my research will help people better understand how to model, compute, and analyze these new forms of data, quantify their uncertainty, and recognize their limitations. Ultimately, I hope my work will help broaden the applicability of modern generative models and other statistical models to a wider range of scientific investigations and societal decision-making contexts, where sound data analysis directly impacts the well-being of individuals and communities.
What are you working on right now?
The recent explosion of computer-generated data driven by advances in generative AI presents both new challenges and opportunities. One of my recent research areas is design-aware assessment of generative models. After all, a synthetic dataset should be evaluated within the context of the study it is intended to emulate.
For example, in a microbiome study, which is an interdisciplinary domain that has inspired a lot of my statistical thinking, a biomedical researcher who employs a generative model to synthesize microbial compositions must assess that model in the context of its biological and study-specific constraints, rather than simply comparing synthetic and real data using generic distance metrics. Proper evaluation requires accounting for key structural and distributional characteristics relevant to the application.
In the coming years, I hope my research will contribute new ideas on how to build and assess such models in a design and study-aware fashion, advancing both the theoretical foundations and the practical effectiveness of generative models in this rapidly evolving landscape.