ByungHee Cha

I am an Machine Learning Masters student from the Kim Jaechul Graduate School of AI at KAIST. I am currently advised by Professor Jong Chul Ye. I have a deep interest in pursuing a PhD in AI/ML studies, specializing in Diffusion models. Currently, my primary focus is on efficient training of Generative models by injecting structural inductive bias and enhancing cross-modal alignment on multi-modal models.

Experiences

KAIST

Seoul, Korea (Aug 2025 -)

Masters
Kim Jaechul Graduate School of AI

Dalpha

Seoul, Korea (Aug 2023 -)
AI Engineer

CNAI

Seoul, Korea (Jul 2021 - May 2023)
ML Researcher

Seoul National University

Seoul, Korea (Mar 2017 - )

Undergraduate

Computer Science and Engineering

Publications

2025

Align Your Tangent: Training Better Consistency Models via Manifold-Aligned Tangents
Beomsu Kim*, ByungHee Cha*, Jong Chul Ye
Under review on ICLR 2026.

Aligning Text to Image in Diffusion Models is Easier Than You Think
Jaa-Yeon Lee*, ByungHee Cha*, Jeongsol Kim*, Jong Chul Ye
Neurips 2025.

2024

CAS: A Probability-Based Approach for Universal Condition Alignment Score
Paper Project Page
Chunsan Hong*, ByungHee Cha*, Tae-Hyun Oh
Spotlight (Top 5%), ICLR, 2024.

ConfBALD: Deep Bayesian Active Learning against Confusing Classes
Chunsan Hong*, Dogyun Kim*, ByungHee Cha*, Bohyung Kim, Junsik Kim, Tae-Hyun Oh
Under review.

2023

Enhancing Classification Accuracy on Limited Data via Unconditional GAN
Paper
Chunsan Hong*, ByungHee Cha*, Bohyung Kim, Tae-Hyun Oh

ICCV workshop, 2023.

Page updated

Google Sites

Report abuse