Haeun Moon

About

I am an assistant professor in the School of Transdisciplinary Innovations and the Department of Statistics at Seoul National University.

Before joining SNU, I was a postdoctoral researcher in the Department of Statistics and Data Science at Carnegie Mellon University, mentored by Dr. Kathryn Roeder and Dr. Jing Lei. I was also involved in several genetic projects mentored by Dr. Bernie Devlin. I received my Ph.D. in 2022 from the Department of Statistics at the University of Pittsburgh, advised by Dr. Kehui Chen. Before joining a Ph.D. program, I worked as an economic statistician at the Bank of Korea for 2.5 years. I obtained a M.A. in Economics from Seoul National University and a B.S. in Mathematical Science from KAIST.

Research Interest 

My research centers on statistical inference under minimal assumptions, with a broad interest in association studies for modern data. A recurring theme of my work is how we should measure and reason about such relationships when we are unwilling to commit to a parametric model.

One line of research develops nonparametric measures of dependence, variable importance, and related quantities that go beyond classical correlation-based perspectives. A particular focus is on characterizing conditional relationships—how association varies across rich covariate structures—without discretizing the conditioning space or imposing restrictive parametric forms. This work connects naturally to independence testing, feature screening across heterogeneous data types, variable selection, multiple testing, and related problems in high-dimensional inference. Much of it is motivated by modern biological and genomic data, where signals are often weak, variables are mixed in type, and dependencies take forms that classical methods overlook.

A second line of research studies how the predictive power of machine learning can be combined with rigorous statistical guarantees. While black-box learning algorithms often achieve excellent predictive performance, they typically provide limited inferential validity or uncertainty quantification. I am interested in procedures that calibrate such methods and deliver valid inferential statements, so that flexible learning tools can support reliable statistical conclusions. This includes conformal and distribution-free approaches to inference, as well as methods for analyzing incomplete, noisy, or heterogeneous biomedical data.

Here is my CV (update: 2024.9.1).

Academic Positions

• Assistant Professor, Seoul National University (2024–present)
  Joint appointment in the School of Transdisciplinary Innovations and the Department of Statistics

• Postdoctoral Researcher, Department of Statistics and Data Science, Carnegie Mellon University (2022–2024)