2025: Paper 'The Oversmoothing Fallacy: A Misguided Narrative in GNN Research' published on arXiv, challenging the prevailing belief about oversmoothing in deep GNNs
2025: Paper 'Influence Functions for Edge Edits in Non-Convex Graph Neural Networks' published on arXiv, proposing influence functions for GNNs
Mar 2025: Paper on personalizing LLMs with a graph-based collaborative filtering framework uploaded to arXiv
Oct 2024: Paper on gradient analysis in GNNs uploaded to arXiv
Jun 2024: Paper on label smoothing in GNNs uploaded to arXiv
Background
Ph.D. student at the POSTECH Machine Learning Lab, Graduate School of Artificial Intelligence, POSTECH
Advised by Prof. Dongwoo Kim
Research interests include various aspects of machine learning, particularly Graph Neural Networks (GNNs) and dynamical systems
Recently exploring interpretation of deep learning architectures and training methods through the lens of Ordinary Differential Equations (ODEs)
Aims to develop innovative and efficient neural network designs by integrating advanced mathematical principles with practical applications