GiHoon Kim
I am a Ph.D. student in Data Science at Seoul National University, advised by Prof. Taesup Kim in the Learning Adaptation Algorithm Lab. My research focuses on generative modeling and controllable generation, especially how diffusion and flow-based models can be theoretically grounded, reliably controlled, and applied to diverse problem settings.
Research Focus
Generative modeling Understanding diffusion and flow-based models through their learning dynamics, inference behavior, and distributional properties.
Controllable generation Developing guidance, conditioning, and adaptation methods that steer generation while preserving model priors.
Adaptive applications Extending these principles to domains where generative models can be adapted, including personalization, multimodal systems, inverse problems, and LLM-based applications.
Work Experience
- NAVER CLOVA (Cloud) Research Intern, Image/Vision, Sep. 2022 - Mar. 2023 Developed 3D-aware hairstyle transfer and 3D-consistent generative pipelines.
- Teaching Assistant Programming, autonomous driving, and real-time learning courses, 2019 - 2024
Awards / Grants
- NRF Research Encouragement Grant Principal Investigator, National Research Foundation of Korea, 2025 - 2026
- Undergraduate Paper Award ICROS (Institute of Control, Robotics and Systems), 2020
Education
- Seoul National University Ph.D. in Data Science, 2024 - Present
- KAIST M.S. in Culture Technology, 2021 - 2023
- Handong Global University B.S. in Computer Science & Electronic Engineering, 2015 - 2021
I am always happy to discuss research ideas or potential collaboration. Please feel free to reach out by email.
News
- ICLR 2026: Preserve and Personalize: Personalized Text-to-Image Diffusion Models without Distributional Drift
- NeRFFaceSpeech_Code reached 65 GitHub stars.
- Served as a reviewer for NeurIPS 2026 and ICLR 2027.