Yuliang Xu
About me
I am an Assistant Professor in the Department of Applied Mathematics and Statistics at Johns Hopkins University, with a secondary appointment in the Department of Biostatistics.
My recent research focuses on the distributional evaluation of generative models for scientific data generation, especially synthetic microbiome data generation and evaluation.
Previously, I was a postdoctoral researcher advised by Prof. Li Ma in the Department of Statistics and Data Science Institute at the University of Chicago (July 2025–June 2026), and in the Department of Statistical Science at Duke University (July 2024–June 2025). I obtained my Ph.D. in Biostatistics from the University of Michigan, advised by Prof. Jian Kang.
During my time at Michigan, I also collaborated with Prof. Florian Gunsilius. I received my B.S. in Mathematics and Applied Mathematics from South China University of Technology in 2017 and my M.S. in Statistics from the University of Waterloo in 2019.
Email: yxu296@jh.edu (work)
yuliangx@umich.edu (permanent)
Scientific validity for synthetic data
Core research question
When can synthetic data generated from scientific datasets be trusted to support the same scientific conclusions as the real data?
I develop statistical tools for assessing whether generative models preserve the features of scientific data that matter for downstream analysis and inference. A central goal is to understand when synthetic data can be safely used to answer scientific questions—and how much real data is needed before such guarantees become possible.
More broadly, I am interested in defining what it means for a generative model to be scientifically valid, rather than merely statistically or perceptually realistic.
Have a scientific dataset or workflow where synthetic data could make a difference? I welcome conversations with applied researchers about evaluating generative models in real scientific settings.
Research areas
Theory and methods
- Generative models
- Nonparametric Bayesian statistics
- Optimal transport
- Causal inference
Applications and data
- Brain imaging (volumetric fMRI)
- Microbiome data (counts and relative abundances)
- DNA sequencing data
Education
- Ph.D. in Biostatistics, University of Michigan, Ann Arbor, MI, 2024
- M.S. in Statistics, University of Waterloo, Waterloo, ON, 2019
- B.S. in Mathematics and Applied Mathematics, South China University of Technology, Guangzhou, China, 2017
Employment
- Assistant Professor, Department of Applied Mathematics and Statistics; secondary appointment in Biostatistics, Johns Hopkins University, starting July 2026
- Postdoctoral Researcher, Department of Statistics and Data Science Institute, University of Chicago, July 2025–June 2026
- Postdoctoral Researcher, Department of Statistical Science, Duke University, July 2024–June 2025