Byeonghu Na (나병후)
Ph.D. Candidate
Applied Artificial Intelligence Laboratory (AAILAB) (Prof. Il-Chul Moon)
Department of Industrial and Systems Engineering
KAIST (Korea Advanced Institute of Science and Technology)
E-mail: byeonghu.na [at] kaist [dot] ac [dot] kr (or) gwp03052 [at] gmail [dot] com
[Github] [LinkedIn] [Google Scholar]
Research Interest
My research interests lie in the development and application of generative models, particularly in addressing challenges related to incomplete data. My current focus is on evaluating data incompleteness during the training and inference of diffusion-based generative models, with the aim of developing principled solutions to improve the robustness of these models. I am also interested in identifying and addressing problems of incomplete data in various multimodal systems.
Selected Publication
[3] Diffusion Rejection Sampling (DiffRS)
Byeonghu Na, Yeongmin Kim, Minsang Park, Donghyeok Shin, Wanmo Kang, and Il-Chul Moon
ICML 2024 (The Forty-first International Conference on Machine Learning)
[2] Label-Noise Robust Diffusion Models (TDSM)
Byeonghu Na, Yeongmin Kim, HeeSun Bae, Jung Hyun Lee, Se Jung Kwon, Wanmo Kang, and Il-Chul Moon
ICLR 2024 (The Twelfth International Conference on Learning Representations)
[1] Deep Generative Positive-Unlabeled Learning under Selection Bias (VAE-PU)