Bryan Wong | KAIST

About Me

I am currently a Ph.D. candidate in Data Science at KAIST, advised by Prof. Mun Yong Yi, and have been fortunate to collaborate with and be mentored by Dr. Huazhu Fu at A*STAR. Before beginning my Ph.D., I received my M.S. in Data Science from KAIST in 2023 and completed my B.S. in Computer Science at NTUST in 2021, graduating one year early.

My current research focuses on multimodal AI for gigapixel medical images, advancing how vision–language models (VLMs) and multimodal LLMs understand large, complex visual data. The first study, HiVE-MIL (NeurIPS 2025), integrates VLMs into a multiple instance learning framework to explicitly model interactions across different magnifications, enabling hierarchical and multimodal understanding of gigapixel images. The second study moves beyond cancer diagnosis and static, black-box pipelines, focusing on multimodal agentic reasoning that dynamically explores informative regions of gigapixel images in a clinician-like manner.

News

[Jan. 2026] Selected as an Outstanding Award Recipient (Best Runner-up) by the Graduate School of Data Science (GSDS), KAIST, based on the 2025 Annual Research Performance Evaluation.
[Nov. 2025] Awarded the Daewoong Foundation Global Scholarship Program (Nov. 2025–Feb. 2026), supporting outstanding international students in Korea [Certificate].
[Nov. 2025] I'll be attending and presenting our paper (HiVE-MIL) at NeurIPS 2025, held in San Diego, California, US, from Dec 2–7, 2025. Looking forward to connecting!
[Sep. 2025] One paper on a hierarchical vision–language MIL framework for gigapixel medical images (HiVE-MIL) was accepted to NeurIPS 2025!
[Aug. 2025] One paper (LOKT) was accepted at EMNLP (Findings) 2025!
[Aug. 2025] One paper (MF2Vec) was accepted at CIKM 2025!
[Jul. 2025] Selected as a student volunteer at MICCAI 2025, held in Daejeon, South Korea, from Sept 23–27, 2025!
[Jun. 2025] One paper (MicroMIL) was accepted at MICCAI 2025!
[May. 2025] New preprint: Few-Shot Learning from Gigapixel Images via Hierarchical Vision-Language Alignment and Modeling [📑 preprint].
[Jan. 2025] One paper (RMTS) was accepted at NAACL (Findings) 2025!
[Jan. 2025] Two papers (Rethinking MIL FE, Time Series MIL) were accepted at IEEE ISBI 2025!
[Jan. 2024] One paper (Predicted Loss AL) was accepted at BMC Medical Imaging 2024!
[Aug. 2023] Started Ph.D. program at the Graduate School of Data Science (GSDS), KAIST!
[Jun. 2023] Successfully defended my Master’s thesis!
[Jun. 2022] One paper (Gender Pay Gap) was accepted at IC2S2 2022 and KCC 2022!
[Aug. 2021] Started M.S. program at the Graduate School of Data Science (GSDS), KAIST!
[Jun. 2021] Graduated one year early with a B.S. in Computer Science, NTUST!

Publications

NeurIPS

Few-Shot Learning from Gigapixel Images via Hierarchical Vision-Language Alignment and Modeling

Bryan Wong*, Jongwoo Kim*, Huazhu Fu, and Mun Yong Yi

The Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS), 2025.

(Acceptance rate: 5,290 / 21,575 = 24.52%)

PDF Code OpenReview, Poster, Slide & Video BibTex

EMNLP

Not All Options Are Created Equal: Textual Option Weighting for Token-Efficient LLM-Based Knowledge Tracing

Jongwoo Kim*, SeongYeub Chu*, Bryan Wong, and Mun Yong Yi

Conference on Empirical Methods in Natural Language Processing (EMNLP Findings), 2025.

(Acceptance rate: 1,812 / 8,174 = 22.16% for Main; 1,420 / 8,174 = 17.35% for Findings)

PDF Code Poster BibTex

CIKM

Leveraging Multi-facet Paths for Heterogeneous Graph Representation Learning

Jongwoo Kim, SeongYeub Chu, Hyeongmin Park, Bryan Wong, Keejun Han, and Mun Yong Yi

34^th ACM International Conference on Information and Knowledge Management (CIKM), 2025.

(Acceptance rate: 443 / 1,627 = 27%)

PDF Code BibTex

MICCAI

MicroMIL: Graph-based Contextual Multiple Instance Learning for Patient Diagnosis Using Microscopy Images

Jongwoo Kim*, Bryan Wong*, Huazhu Fu, Willmer Rafell Quiñones Robles, Young Sin Ko, and Mun Yong Yi

28^th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), 2025.

(Acceptance rate: 1,014 / 3,447 = 29%)

PDF Code Poster BibTex

NAACL

Rationale Behind Essay Scores: Enhancing S-LLM’s Multi-Trait Essay Scoring with Rationale Generated by LLMs

SeongYeub Chu*, Jongwoo Kim*, Bryan Wong, and Mun Yong Yi

Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL Findings), 2025.

PDF Code Poster BibTex

ISBI

Rethinking Pre-Trained Feature Extractor Selection in Multiple Instance Learning for Whole Slide Image Classification

Bryan Wong, Sungrae Hong, and Mun Yong Yi

IEEE International Symposium on Biomedical Imaging (ISBI), 2025.

PDF Code Poster BibTex

ISBI

Towards Classifying Histopathological Microscope Images as Time Series Data

Sungrae Hong, Hyeongmin Park, Youngsin Ko, Sol Lee, Bryan Wong, and Mun Yong Yi

IEEE International Symposium on Biomedical Imaging (ISBI), 2025.

PDF Poster BibTex

BMC

A predicted-loss based active learning approach for robust cancer pathology image analysis in the workplace

Mujin Kim, Willmer Rafell Quiñones Robles, Young Sin Ko, Bryan Wong, Sol Lee, and Mun Yong Yi

BMC Medical Imaging, 2024.

PDF BibTex
IC2S2 & KCC

Analyzing Gender Pay Gap in STEM Fields by Life Trajectory

Wenchao Dong, Bryan Wong, Hasnain Irshad Bhatti, Lanu Kim, and Meeyoung Cha

8^th International Conference on Computational Social Science (IC2S2) and Korea Computer Congress (KCC), 2022.
PDF Poster BibTex