Wonjin Lee

Wonjin Lee

I am an integrated M.S.-Ph.D. student at the Machine Learning and Vision Group (MLV Group), advised by Prof. Kwang In Kim. I am currently enrolled in the Graduate School of Artificial Intelligence at POSTECH.

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

CV | Google Scholar | LinkedIn

Publications

HInT: Hypergraph Infusion at the Structural Layers Improves Table Understanding

Wonjin Lee, Soomi Jeong, Kwang In Kim

ICML, 2026

project page | paper | poster

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

project page | paper | poster

Subspace-based Approximate Hessian Method for Zeroth-Order Optimization

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

arXiv preprint, 2025

paper

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

paper

Mask-guided metal artifact reduction with contrastive learning

Juhyeong Ki, Wonjin Lee, Bitbyeol Kim, Dukju Kim, Seongmoon Jung, Jimin Lee

IEEE Access, 2025

paper

Experience

Machine Learning and Vision Group (MLV Group) | Research Student Jan. 2024 - Present

  • Conducted research on table understanding, structured data understanding, and vision-language reasoning
  • Developed a divide-and-conquer method for selecting question-relevant subtables in table question answering
  • Developed a hypergraph-based structural infusion method for improving table reasoning in language models

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

  • Worked on unsupervised prompt optimization for realistic zero-shot image generation

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

  • Developed a video lip-sync detection model to identify interview cheating with pre-recorded audio
  • Built a speaker classification pipeline to detect voice inconsistency in recorded speech
  • Developed a multimodal interview cheating classification model using eye movement and voice signals

UNIST Radiation and Medical Intelligence Lab | Research Intern Oct. 2021 - Apr. 2023

  • Developed deep learning models for medical image analysis using CT and infrared imaging data
  • Worked on metal artifact reduction for head and neck CT images to improve image quality for diagnosis
  • Developed a sinusitis classification model using infrared images from a medical imaging device

NCSoft Inc., Speech AI Lab | Research Intern, Singing Voice Team Jul. 2021 - Feb. 2022

  • Developed a wav-to-MIDI transcription model for automatic singing melody transcription

Education

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.

Academic Services

Conference Reviewer

Journal Reviewer