Yeseul Jeon

Employment 

• Assistant Professor, Department of Mathematics and Statistics, University of North Carolina at Charlotte, August 2026 -

• Research Fellow, Department of Statistics, Texas A&M University and  Department of Epidemiology & Biostatistics, University of California, San Francisco,  September 2024 - July 2026.

• AI Researcher, Laboratory for Artificial Intelligence, Vive Company, February 2016 - February 2021

Statistical Modeling Meets Deep Learning

My research is at the intersection of statistical modeling, Bayesian inference and deep learning. I build learning architectures that directly embed statistical structure into deep neural networks for strong predictive performance, interpretability, and principled uncertainty quantification.

Rather than treating deep learning as a purely black-box predictive tool, I am interested in building models that account for the structure, dependence, and scientific context of the data.

Data That Motivate My Research

I am particularly interested in complex data that are difficult to analyze using standard statistical methods, including:

• Spatial and spatiotemporal data

• Large-scale and high-dimensional datasets

• Neuroimaging and biomedical imaging data

• Unstructured and strongly correlated data

My methodological work has been motivated by applications in neuroimaging, spatial transcriptomics, environmental science, and public health. During my Ph.D., I developed Bayesian methods for structured high-dimensional data, complex networks, and spatial biomedical data, which provided the foundation for my current research program.

Emerging Research Directions

I am currently expanding this research toward functional data analysis and supervised clustering. In particular, I am interested in developing models that can learn from complex functional or spatial observations while identifying scientifically meaningful subgroups associated with outcomes of interest.

I believe that good models begin with data. The structure of the data, the scientific questions behind them, and the practical challenges of the application should guide the development of new methodology.

I am always interested in collaborating with researchers who share this data-driven research philosophy. If you work with complex datasets and would like to develop thoughtful statistical solutions together, please feel free to reach out 😊.