JoungBin Lee

Ph.D. student · CVLAB, KAIST AI · Seoul, Korea

Generative modeling for world models, with a focus on controllable and geometry-aware image and video generation.

I am advised by Prof. Seungryong Kim. I am interested in diffusion-based image and video generation that preserves 3D structure, temporal consistency, and user-specified controls such as camera trajectories, using 3D/4D representations, and in generative modeling as data augmentation for downstream visual tasks.

CV Google Scholar GitHub LinkedIn joungbinlee@kaist.ac.kr

Education

KAISTSeoul, Korea

Mar 2025 – Present

M.S./Ph.D. Integrated Course in Artificial Intelligence

Advisor: Seungryong Kim

Korea UniversitySeoul, Korea

Mar 2023 – Feb 2025

B.S. in Data Science

GPA 3.9 / 4.5

University of SeoulSeoul, Korea

Mar 2021 – Feb 2023

B.S. in Artificial Intelligence

GPA 4.2 / 4.5

Experience

Sony AITokyo, Japan

Nov 2026 – Feb 2027, Expected

Research Intern

Advisor: Yuki Mitsufuji
Mentors: Takuya Narihira, Kazumi Fukuda

Sony AITokyo, Japan

2025 – Present

Remote Research Collaboration

Advisor: Yuki Mitsufuji
Mentors: Takuya Narihira, Kazumi Fukuda
Papers: 3DScenePrompt (ICLR 2026), Vid-CamEdit (AAAI 2026), MVTrack4Gen (arXiv 2026)

Publications

* equal contribution · † corresponding author

Honors & Service

Awards

2023Silver Award, iNTton DatathonKorea University SW Convergence Foundation
2021Excellence AwardGraduate School of Chemical Safety Management, University of Seoul

Leadership

2021Vice President, AI Department Student CouncilUniversity of Seoul
2021Vice Representative, AI Department Student CouncilUniversity of Seoul