Ganghyun Kim

Ganghyun Kim

M.S. Student in AI

LAIT Lab, UNIST

About

Master's student at UNIST AI Graduate School, working in the Lab of Advanced Imaging Technology. My research focuses on deep learning, specifically detecting AI-generated media and improving position encoding in vision transformers.

Research

AI-Generated Video Detection

Robust methods to detect synthetic media from generative models

Position Encoding in Vision Transformers

Rotary position embeddings and their extensions for 2D visual data

Publications

Interpreting Stress Detection Models Using SHAP and Attention for MuSe-Stress 2022
Ho-min Park, Ganghyun Kim, Jinsung Oh, Arnout Van Messem, Wesley De Neve
IEEE Transactions on Affective Computing, 2025
DOI
Evaluating Visual Explanations of Attention Maps for Transformer-Based Medical Imaging
Minjae Chung, Jong Bum Won, Ganghyun Kim, Yujin Kim, Utku Ozbulak
MICCAI Workshops, 2024
DOIarXivCode
RAMM: A Residual Attention Multimodal Model for Humor Detection
Ankit Kumar Singh, Ganghyun Kim, Jaeheon Kim, Ho-min Park, Bong Jun Choi, Wesley De Neve
IHCI, 2024
DOI
Exploring Multimodal Approaches and Fusion Methods for CEO Social Attribute Prediction
Ankit Kumar Singh, Ganghyun Kim, Jaeheon Kim, Ho-min Park, Bong Jun Choi, Wesley De Neve
MuSe @ ACM Multimedia, 2024
DOI
MuSe-Personalization 2023: Feature Engineering, Hyperparameter Optimization, and Transformer-Encoder Re-discovery
Ho-min Park, Ganghyun Kim, Arnout Van Messem, Wesley De Neve
MuSe @ ACM Multimedia, 2023

Education

2025 - Present

M.S. in Artificial Intelligence

UNIST, South Korea
Advisor: Prof. Jaejun Yoo
2018 - 2024

B.S. in Molecular Biotechnology

Ghent University Global Campus
Cum Laude