Changhun Kim

I am a first-year Ph.D. student in Computer Science at the University of Wisconsin–Madison, advised by Prof. Grigorios Chrysos. Before that, I was a researcher at AITRICS advised by Prof. Eunho Yang, where I also worked with Prof. Evan Shelhamer at UBC and Prof. Juho Lee at KAIST. I received my M.S. in Artificial Intelligence, also under Prof. Eunho Yang, and my B.S. in Computer Science and Mathematics, both from KAIST.

My research focuses on the algorithmic and theoretical foundations of deep generative models. I am particularly interested in understanding why diffusion models generalize and when they hallucinate, and in turning these insights into more principled algorithms that exploit structured inductive biases. I am also excited about extending generative models to world (action) models for embodied agents, and more broadly about building trustworthy machine learning systems.

I am always happy to connect and collaborate with researchers whose interests align with mine.

📌 I am actively seeking an industry research intern position for Summer 2027. If our interests align, please feel free to reach out.