About Me
I'm a Ph.D. student in Statistics at Seoul National University, advised by Prof. Gunwoong Park . I'm interested in topics related to directed acyclic graph (DAG) models, particularly in structure learning algorithms. I focus on developing theories and algorithms that relax traditional assumptions—such as identifiability, faithfulness, and causal sufficiency—to enable DAG learning in a wider variety of data scenarios. Additionally, I explore how DAG perspectives can enhance transparency and reliability in complex machine learning models through representation theory, causal AI, and explainable AI (XAI).
Education
-
Ph.D. in Statistics, Seoul National University
2024 – Present -
M.S. in Statistics, Seoul National University
2022 – 2024 -
B.S. in Statistics, Seoul National University
2018 – 2022 -
Visiting Research Affiliate, University of California, Santa Barbara
(Department of Statistics & Applied Probability)
2025.07.15 – 2026.02.28
Publications
Authors: Sunmin Oh, Seungsu Han, Gunwoong Park
Abstract: Much of science involves discovering and modeling causal relationships in nature. Significant progress has been made in developing statistical methods for representing and identifying causal knowledge from data using Linear Non-Gaussian Acyclic Models (LiNGAMs). Despite successes in learning LiNGAMs across various sample settings, the optimal sample complexity for high-dimensional LiNGAMs remains unexplored. This study establishes the optimal sample complexity for learning the structure of LiNGAMs under a sub-Gaussianity assumption. Specifically, it introduces a structure recovery algorithm using distance covariance that achieves the optimal sample complexity, without assuming faithfulness or a known indegree. The theoretical findings and superiority of the proposed algorithm compared to existing algorithms are validated through numerical experiments and real data analysis.
View PaperAuthors: Seyong Hwang, Kyoungjae Lee, Sunmin Oh, Gunwoong Park
Abstract: Causal discovery in high-dimensional linear Bayesian networks is challenging, even with partial structural knowledge. Such information is often edge-specific and noisy, and naively enforcing uniform shrinkage or hard constraints can induce structural ambiguity. We propose Probabilistic Edge Modulation (PEM), a principled probabilistic framework that replaces hard structural constraints with soft, edge-specific modulation via a spike-and-slab formulation. PEM integrates heterogeneous priors into ordering recovery and parent selection through a unified MAP formulation that remains computationally tractable in polynomial time. We establish high-dimensional consistency under both sub-Gaussian and heavy-tailed errors with bounded moments, and demonstrate robustness to prior misspecification. Experiments on synthetic and real retail data demonstrate improved structural stability and graph recovery in sparse and data-limited regimes; in a real e-commerce dataset, probabilistic transfer from a data-rich group stabilizes smaller groups and yields non-degenerate graph estimates.
View PaperProjects
- Review and consult statistical tests.
- Review and advise machine learning methodology.
- Predict the copyright royalty of each K-pop music for the next 30 years.
- Propose a system for establishing equitable pricing in copyright acquisition.
- Assess the impacts of performance marketing through data analytics.
- Analyze review comments of products.
- Develop tailored audio-book recommendation systems, utilizing user data and app usage records.
- Content-based filtering, Deep sequential neural network approaches are employed.
- Develop tailored plastic surgery recommendation systems, utilizing user data and app usage records.
- Rule-based, Content-based filtering, Deep learning approaches are employed.
Talks & Posters
| Type | Details | Action |
|---|---|---|
| Poster | 2023.06 - The Korean Statistical Society Summer Conference | |
| Presentation | 2024.05 - The joint international seminar in collaboration with Shiga University and RIKEN | Link |
| Presentation | 2024.06 - The joint international seminar in collaboration with Kyushu University | Link |
| Presentation | 2024.11 - The BRL workshop in Jeon-ju | |
| Poster | 2025.05 - Airtificial Intelligence and Statistics 2025 | |
| Presentation | 2025.06 - The Korean Statistical Society Summer Conference | |
| Presentation | 2026.04 - The joint international seminar in collaboration with Kyushu University | Link |
| Poster | 2026.08 - The Conference on Uncertainty in Artificial Intelligence 2026 |