I am a PhD candidate at POSTECH.
My research interests lie in topological deep learning on symmetric simplicial sets, a new class of higher-order domains, and its applications to hypergraphs.
For a detailed overview of my work and background, please see my [Curriculum Vitae] and [Research Statement].
Email: sjin [dot] choi1 [at] gmail [dot] com
Education
Ph.D. in Mathematics, POSTECH, March 2014 – Present (Advisor: Yong-Geun Oh)
B.S. in Mathematics, POSTECH, March 2010 – February 2014
Publications
Journal Articles
G. Kim, S. Choi, and S. Y. Yun, "A mutual information-based framework for enhancing graph neural networks on heterophily," IEEE Access, 2025, doi: 10.1109/ACCESS.2025.3624415.
S. Choi, G. Kim, and Y. G. Oh, "Hypergraph neural sheaf diffusion: A symmetric simplicial set framework for higher-order learning," IEEE Access, vol. 13, pp. 131823--131838, Jul. 2025, doi: 10.1109/ACCESS.2025.3592104.
Workshop Papers
S. Choi, G. Kim, and S. Y. Yun, "Node mutual information: Enhancing graph neural networks for heterophily," in NeurIPS 2023 Workshop: New Frontiers in Graph Learning, 2023. [Online].
Preprints
S. Choi, G. Kim, and S. Y. Yun, "Weisfeiler-Lehman go categorical," in progress.
S. Choi and J. Park, "Cellular sheaf Laplacians on the set of simplices of symmetric simplicial set induced by hypergraph," arXiv preprint arXiv:2411.08458, Nov. 2024. [Online].
Invited talks
Workshop on Mathematical Signal Processing, RWTH Aachen University, Germany, Sep 2025.
Talk: "Symmetric Simplicial Lifting for Hypergraph Learning". Slides (PDF)
Poster: "Weisfeiler-Lehman Go Categorical".
Topological Data Analysis and Industrial Mathematics: Bridging Theory and Applications, Kyushu University, Japan, Aug 2025, "Symmetric Simplicial Lifting for Hypergraph Learning". Slides (PDF)
Asia Pacific Seminar on Applied Topology and Geometry, Online, Jan 2025, "Sheaf Laplacians on hypergraphs via symmetric simplicial sets". Slides (PDF)