cv
Basics
| Name | Kyeongmin Yeo |
| Label | M.S. Student |
| aaaaa@kaist.ac.kr | |
| Url | https://32V.github.io/ |
| Summary | M.S. student at KAIST Visual AI Lab, conducting research on generative models and computer graphics. Interested in generative modeling and visual content generation, including images, videos, and textures. |
Interests
| Academic Interests | |
| Generative AI | |
| Diffusion & Flow-Based Generative Models | |
| Computer Vision | |
| Visual Content Creation |
Work
-
2026.02 - Present Daejeon, South Korea
M.S. Student
KAIST Visual AI Group
Conducting research on generative models and computer graphics under the supervision of Prof. Minhyuk Sung.
-
2023.12 - 2026.02 Daejeon, South Korea
Undergraduate Intern
KAIST Visual AI Group
Conducted research on generative models and computer graphics under the supervision of Prof. Minhyuk Sung.
Education
Publications
-
2026 PairFlow: Closed-Form Source-Target Coupling for Few-Step Generation in Discrete Flow Models
ICLR 2026
A closed-form source-target coupling method for few-step generation in discrete flow models.
-
2026 MatLat: Material Latent Space for PBR Texture Generation
CVPR 2026 (Highlight)
A material latent space for physically based rendering (PBR) texture generation.
-
2025 StochSync: Stochastic Diffusion Synchronization for Image Generation in Arbitrary Spaces
ICLR 2025
A stochastic diffusion synchronization framework for image generation in arbitrary spaces.
-
2025 Neural Green's Functions
NeurIPS 2025
A neural solution operator for linear partial differential equations inspired by Green's functions.
-
2025 ORIGEN: Zero-Shot 3D Orientation Grounding in Text-to-Image Generation
NeurIPS 2025
A method for zero-shot 3D orientation grounding in text-to-image generation.
-
2025 Ψ-Sampler: Initial Particle Sampling for SMC-Based Inference-Time Reward Alignment in Score Models
NeurIPS 2025 (Spotlight)
A novel initial particle sampling method for SMC-based inference-time reward alignment in score models.
-
2024 SyncTweedies: A General Generative Framework Based on Synchronized Diffusions
NeurIPS 2024
A general framework for generating diverse visual content by synchronizing multiple diffusion processes.
-
2024 Neural Pose Representation Learning for Generating and Transferring Non-Rigid Object Poses
NeurIPS 2024
A neural pose representation for generating and transferring non-rigid object poses.
Awards
- 2026
- 2025
Next-Generation Engineer Grand Prize, 2nd Cohort
Institute for Promotion of Engineering and Science in Korea (IPESK)
- 2025
- 2025
Outstanding Award (2nd place), Undergraduate Research Program (URP)
College of Engineering, KAIST
Winter/Spring 2025
- 2025
- 2024
- 2022–2023
- 2023
- 2022
National Excellence Scholarship (Natural Sciences and Engineering)
Korea Student Aid Foundation
Volunteer
-
2026 - Present