Research Intern, Disney Research Zurich
Worked on the diffusion team, advised by Vinicius C. Azevedo, contributing to diffusion model training and preprocessing tools.
KAIST AI / BISPL Lab
Jaa-Yeon Lee
I am a PhD student at KAIST AI in the BISPL Lab, advised by Prof. Jong Chul Ye.
My recent work focuses on diffusion RL for text-to-image alignment in diffusion models. Currently, I'm working on long-term alignment of AR video diffusion models. I also bring a medical imaging background from MRI reconstruction, and medical image analysis.
Feel free to reach out if you would like to connect or discuss research!
News
Experience
Research
Alignment-Guided Score Matching for Text-to-Image Alignment in Diffusion Models
ICML 2026 Spotlight (top 2.2%)
A reward-free diffusion RL method that improves text-to-image alignment without relying on external reward models.
SoftREPA: Aligning Text to Image in Diffusion Models is Easier Than You Think
NeurIPS 2025
A lightweight contrastive fine-tuning strategy that improves text-to-image alignment with learnable soft tokens.
UNICORN: Ultrasound Nakagami Imaging via Score Matching and Adaptation for Assessing Hepatic Steatosis
arXiv, 2026
Score matching and adaptation for ultrasound Nakagami imaging and hepatic steatosis assessment.
Earlier Publications
MR Reconstruction in k-space using Vision Transformer boosted with Masked Image Modeling, ISMRM 2023.
Metal Artifact Correction MRI Using Multi-contrast Deep Neural Networks for Diagnosis of Degenerative Spinal Diseases, MICCAI Workshop 2022.
An unsupervised two-step training framework for LDCT denoising, Medical Physics 2023.
Unsupervised Domain Adaptation for Low-dose Computed Tomography Denoising, IEEE Access 2022.
Wavelet Subband-specific Learning for Low-dose Computed Tomography Denoising, PLOS ONE 2022.
Education
PhD, Artificial Intelligence, KAIST
BISPL Lab, advised by Prof. Jong Chul Ye.
MS, Bio and Brain Engineering, KAIST
MRI Lab, advised by Prof. Sung-Hong Park.
BS, Mechanical and Biomedical Engineering, Ewha Womans University
Dean's List, five semesters. Research intern in the Medical AI & CV Lab, advised by Prof. Jang-Hwan Choi.