Seungwoo Yoo

E3-5, Rm. 510, KAIST

291 Daehak-ro,

Yuseong-gu, Daejeon,

South Korea 34141

I am a second-year Ph.D. student in the KAIST Visual AI Group, headed by Prof. Minhyuk Sung. I earned my B.S. degree in Computer Science (Major) and Electrical Engineering (Minor) from KAIST in Aug. 2023 and my M.S. degree in Computer Science from KAIST in Feb. 2025.

My research interest encompasses various computational problems arising from the intersection of machine learning and science. In particular, my research area includes but is not limited to:

Generative Modeling;
AI for Science;
Geometry Processing.

For details, refer to my Curriculum Vitae (PDF).

News

Jul 11, 2026	I am selected as one of the 120 recipients of the Graduate Presidential Science Scholarship from a pool of 2,651 applicants. I am deeply grateful for the support provided by the Presidential Science Scholarship throughout my undergraduate, master’s, and Ph.D. studies.
Feb 21, 2026	Our paper on warping tokens instead of pixels for multi-view MLLM reasoning has been accepted at CVPR 2026. Stay tuned for more details.
Jan 26, 2026	Our papers, PairFlow and BézierFlow, have been accepted at ICLR 2026.
Jan 13, 2026	Our paper, DiffusionRollout, which introduces an adaptive rollout strategy for diffusion models applied to time-dependent PDEs, has been accepted at TMLR. More details coming soon.
Dec 25, 2025	Our paper introducing PairFlow, a lightweight preprocessing algorithm for training Discrete Flow Models (DFMs) to enable few-step sampling, is now available on arXiv.

Publications

2026

Token Warping Helps MLLMs Look from Nearby Viewpoints

Phillip Y. Lee*, Chanho Park*, Mingue Park, Seungwoo Yoo, Juil Koo, and Minhyuk Sung

CVPR, 2026

arXiv Bib Code Website

@inproceedings{lee2026tokenwarping,
  title = {{Token Warping Helps MLLMs Look from Nearby Viewpoints}},
  author = {Lee*, Phillip Y. and Park*, Chanho and Park, Mingue and Yoo, Seungwoo and Koo, Juil and Sung, Minhyuk},
  booktitle = {CVPR},
  year = {2026},
}

PairFlow: Closed-Form Source-Target Coupling for Few-Step Generation in Discrete Flow Models

Mingue Park*, Jisung Hwang*, Seungwoo Yoo*, Kyeongmin Yeo, and Minhyuk Sung

ICLR, 2026

arXiv Bib Code Website

@inproceedings{park2026pairflow,
  title = {{PairFlow: Closed-Form Source-Target Coupling for Few-Step Generation in Discrete Flow Models}},
  author = {Park*, Mingue and Hwang*, Jisung and Yoo*, Seungwoo and Yeo, Kyeongmin and Sung, Minhyuk},
  booktitle = {ICLR},
  year = {2026},
}

BézierFlow: Learning Bézier Stochastic Interpolant Schedulers for Few-Step Generation

Yunhong Min*, Juil Koo*, Seungwoo Yoo, and Minhyuk Sung

ICLR, 2026

arXiv Bib Code Website

@inproceedings{min2026bezierflow,
  title = {{BézierFlow: Learning Bézier Stochastic Interpolant Schedulers for Few-Step Generation}},
  author = {Min*, Yunhong and Koo*, Juil and Yoo, Seungwoo and Sung, Minhyuk},
  booktitle = {ICLR},
  year = {2026},
}

DiffusionRollout: Uncertainty-Aware Rollout Planning in Long-Horizon PDE Solving

Seungwoo Yoo*, Juil Koo*, Daehyeon Choi*, and Minhyuk Sung

TMLR, 2026

arXiv Bib

@article{yoo2026diffusionrollout,
  title = {{DiffusionRollout: Uncertainty-Aware Rollout Planning in Long-Horizon PDE Solving}},
  author = {Yoo*, Seungwoo and Koo*, Juil and Choi*, Daehyeon and Sung, Minhyuk},
  journal = {TMLR},
  year = {2026},
}

Proxy-Free Gaussian Splats Deformation with Splat-Based Surface Estimation

Jaeyeong Kim, Seungwoo Yoo, and Minhyuk Sung

3DV, 2026

arXiv Bib Code

@inproceedings{kim2026gsdeform,
  title = {{Proxy-Free Gaussian Splats Deformation with Splat-Based Surface Estimation}},
  author = {Kim, Jaeyeong and Yoo, Seungwoo and Sung, Minhyuk},
  booktitle = {3DV},
  year = {2026},
}

2025

Neural Green’s Functions

Seungwoo Yoo, Kyeongmin Yeo, Jisung Hwang, and Minhyuk Sung

NeurIPS, 2025

arXiv Bib PDF Code

2024

Neural Pose Representation Learning for Generating and Transferring Non-Rigid Object Poses

Seungwoo Yoo, Juil Koo, Kyeongmin Yeo, and Minhyuk Sung

NeurIPS, 2024

arXiv Bib PDF Code Website

@inproceedings{yoo2024neuralpose,
  title = {{Neural Pose Representation Learning for Generating and Transferring Non-Rigid Object Poses}},
  author = {Yoo, Seungwoo and Koo, Juil and Yeo, Kyeongmin and Sung, Minhyuk},
  booktitle = {NeurIPS},
  year = {2024},
}

As-Plausible-As-Possible: Plausibility-Aware Mesh Deformation Using 2D Diffusion Priors

Seungwoo Yoo*, Kunho Kim*, Vladimir G. Kim, and Minhyuk Sung

CVPR, 2024

arXiv Bib Code Website

@inproceedings{yoo2024apap,
  title = {{As-Plausible-As-Possible: Plausibility-Aware Mesh Deformation Using 2D Diffusion Priors}},
  author = {Yoo*, Seungwoo and Kim*, Kunho and Kim, Vladimir G. and Sung, Minhyuk},
  booktitle = {CVPR},
  year = {2024},
}

2023

SALAD: Part-Level Latent Diffusion for 3D Shape Generation and Manipulation

Juil Koo*, Seungwoo Yoo*, Minh Hieu Nguyen*, and Minhyuk Sung

ICCV, 2023

arXiv Bib Code Demo Website

@inproceedings{koo2023salad,
  title = {{SALAD}: Part-Level Latent Diffusion for 3D Shape Generation and Manipulation},
  author = {Koo*, Juil and Yoo*, Seungwoo and Nguyen*, Minh Hieu and Sung, Minhyuk},
  booktitle = {ICCV},
  year = {2023},
  demo = {https://huggingface.co/spaces/KAIST-Geometric-AI-Lab/salad-demo},
}