Gwangho Kim

I am a second-year MS student in Computer Science at Hanyang University, advised by Sungyoon Lee. My research interests center on the generalization of generative models, with a particular focus on diffusion models. I am broadly curious about how diffusion models manage to generalize beyond their finite training data — and what this implies for phenomena such as memorization. I am also interested in various aspects of diffusion models, including reward alignment, guidance, distillation, and data attribution.

I am actively looking for PhD positions for both 2027 Spring and Fall.

Email: 25ghkim@gmail.com / ggh1999@hanyang.ac.kr

Education

• M.S. in Computer Science, Hanyang University, March 2025 – Present
• B.S. in Computer Science & Mathematics, Hanyang University, 2018 – September 2024

Publications

• Parallel Tempering Initial Sampling in Inference-Time Reward Alignment
  Myeongjun Oh, Gwangho Kim, Sungyoon Lee ICML 2026 Workshop on Structured Probabilistic Inference & Generative Modeling (SPIGM)

• Localizing Memorized Regions in Diffusion Models via Coordinate-Wise Curvature Differences
  Gwangho Kim, Sungyoon Lee
  International Conference on Machine Learning (ICML 2026)