Wonjin Lee

My research interests include multimodal large language models, structured data understanding such as graphs and tables, and document understanding.

Publications

HInT: Hypergraph Infusion at the Structural Layers Improves Table Understanding

Wonjin Lee, Soomi Jeong, Kwang In Kim

ICML, 2026

Piece of Table: A Divide-and-Conquer Approach for Selecting Sub-Tables in Table Question Answering

Wonjin Lee, Kyumin Kim, Sungjae Lee, Jihun Lee, Kwang In Kim

ACL, 2026

Subspace-based Approximate Hessian Method for Zeroth-Order Optimization

Dongyoon Kim, Sungjae Lee, Wonjin Lee, Kwang In Kim

arXiv preprint, 2025

Dual-encoder architecture for metal artifact reduction for kV-cone-beam CT images in head and neck cancer radiotherapy

Juhyeong Ki, Jung Mok Lee, Wonjin Lee, Jin Ho Kim, Hyeongmin Jin, Seongmoon Jung, Jimin Lee

Scientific Reports, 2024

Mask-guided metal artifact reduction with contrastive learning

Wonjin Lee et al.

IEEE Access, 2025

Research on machine learning, vision, table understanding, and optimization.

KAIST Algorithmic Intelligence Lab | Research Intern May 2023 - Oct. 2023

Worked on machine learning research in the Algorithmic Intelligence Lab.

GenesisLab Inc. | AI Researcher Nov. 2022 - Apr. 2023

Developed video lip-sync detection, speaker classification, and interview video cheating classification models.

Worked on sinusitis classification from infrared images and metal artifact reduction for head and neck CT.

Developed a wav-to-MIDI automatic singing melody transcription model for automating manual transcription tasks.

Pohang University of Science and Technology | Pohang, South Korea Mar. 2024 - Present

Integrated M.S.-Ph.D. in Artificial Intelligence, advised by Prof. Kwang In Kim.

Korea University | Seoul, South Korea Mar. 2016 - Feb. 2023

B.S. in Environmental Science & Ecological Engineering.

Conference Reviewer

Journal Reviewer