Jeongsol Kim
Jeongsol Kim
Embedded Files
Korea Advanced Institute of Science and Technology (KAIST)
Postdoctoral Researcher @ BISPL
I study how to turn sampling-based generative models into solvers that produce feasible solutions under real-world constraints.
News
News
FlowAlign has been accepted to ICLR 2026
I have been selected as one of the Top reviewer of NeurIPS 2025
Our two papers are accepted to NeurIPS 2025 (one spotlight, one poster)
Our paper, FlowDPS has been accepted to ICCV 2025
I will join SNAP Research - Computational Imaging team (@NYC) as Research Intern from June 2025
Our three papers are accepted to ICLR 2025 (one spotlight, two posters)
Our two papers are accepted to ECCV 2024
Recent Works
Recent Works
Reward Score Matching: Unifying Reward-based Fine-tuning of Flow and Diffusion Models
Reward Score Matching: Unifying Reward-based Fine-tuning of Flow and Diffusion Models
Jeongjae Lee*, Jinho Jang*, Jeongsol Kim†, Jong Chul Ye†
arxiv preprint
FlowAlign: Trajectory-Regularized, Inversion-Free Flow-based Image Editing
FlowAlign: Trajectory-Regularized, Inversion-Free Flow-based Image Editing
Jeongsol Kim*, Yeobin Hong*, Jonghyun Park, Jong Chul Ye
ICLR 2026
FlowDPS: Flow-driven posterior sampling for inverse problems
FlowDPS: Flow-driven posterior sampling for inverse problems
Jeongsol Kim*, Bryan Sangwoo Kim*, Jong Chul Ye
ICCV 2025
Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment
Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment
Bryan Sangwoo Kim, Jeongsol Kim, Jong Chul Ye
NeurIPS 2025 (Spotlight)
Contact Information
Email : jeongsol@kaist.ac.kr / jeongsol.kim@gmail.com
Address: 108, Taebong-ro, Seocho-gu, Seoul, South Korea
Page updated
Google Sites
Report abuse