CV
Curriculum Vitae — KyeongKook Seo
Contact Information
| Name | KyeongKook Seo (서경국) |
| Professional Title | Ph.D. Student in Artificial Intelligence |
| kyeongkookseo@unist.ac.kr | |
| Location | AIGS, UNIST, 50 UNIST-gil, Ulju-gun, Ulsan, South Korea 44919 |
Professional Summary
Ph.D. student in the Lab of Advanced Imaging Technology (LAIT), AIGS, UNIST, advised by Prof. Jaejun Yoo. Research interests include generative models, federated learning, safety AI, loss landscape geometry, and weight-space learning.
Experience
-
2022 - 2022 Ulsan, South Korea
Graduate Researcher
Lab of Advanced Imaging Technology (LAIT), UNIST
Generative Feature Matching Network — improving performance of GFMN.
-
2022 - 2022 Member
SmileGate AI Membership
Developed project originating from K-digital Hackathon.
-
2022 - 2022 Participant
K-digital Hackathon
Filtering Violence Scene in Video Data — multi-modal violence detection (image + audio features).
-
2022 - 2022 Trainee
Naver Boostcamp AI Tech
Completed bootcamp certification.
Education
-
2025 - Present Ulsan, South Korea
-
2023 - 2025 Ulsan, South Korea
-
2016 - 2022 Incheon, South Korea
Awards
-
2026 Silver Reviewer
International Conference on Machine Learning (ICML)
Recognized as a Silver Reviewer for ICML 2026.
Publications
-
2025 PRISM: Privacy-Preserving Improved Stochastic Masking for Federated Generative Models
International Conference on Learning Representations (ICLR)
K. Seo, D.-J. Han, and J. Yoo. ICLR 2025, Singapore.
-
2025 Understanding Flatness in Generative Models: Its Role and Benefits
International Conference on Computer Vision (ICCV)
T. Lee, K. Seo, J. Yoo, and S. W. Yoon. ICCV 2025.
-
2026 What Linear Probes Miss: Multi-View Probing for Weight-Space Learning
International Conference on Machine Learning (ICML)
E. Heo, K. Seo, and J. Yoo. ICML 2026, Vancouver (Poster).
Skills
Programming Languages: Python, C, C++
Frameworks: PyTorch
Interests
Research Interests: Generative Models, Safety AI, Federated Learning, Loss Landscape Geometry, Weight-Space Learning