Hyunsoo Lee

Undergraduate student, Seoul National University

I study generative modeling and representation learning. My research develops controllable generative systems across images, human motion, and 3D geometry. More recently, I have become interested in how intelligent systems learn and adapt representations. See my Full Profile for details.

Email CV Scholar GitHub LinkedIn X

Short Bio

Undergraduate student at Seoul National University, studying Electrical and Computer Engineering and Mathematical Sciences.
Author of papers published in venues including NeurIPS, CVPR, ECCV, ICML, and WACV.
Conducting research at the SNU Computer Vision Lab with Prof. Bohyung Han, and at the SNU Machine Perception and Reasoning Lab with Prof. Jonghyun Choi.
Previously worked with Prof. Stephan Mandt at UC Irvine and Prof. Young Min Kim at SNU.

News

2026.09	I received the Youlchon AI Young Researcher Scholarship (USD 5,300).
2026.09	I’m attending ECCV 2026 to present the poster “Accelerated Likelihood Maximization for Diffusion-based Versatile Content Generation”.
2026.08	My research was featured by the College of Engineering at Seoul National University. As an undergraduate student, I’m grateful for the opportunity to have my work and research highlighted!
2026.07	I was selected as an Outstanding Reviewer at ECCV 2026.

Old entries

2026.07	I’m attending ICML 2026, presenting the poster “Calibrated Test-Time Guidance for Bayesian Inference”.
2026.06	One paper accepted to ECCV 2026: “Accelerated Likelihood Maximization for Diffusion-based Versatile Content Generation”.
2026.06	I joined SNU Machine Perception and Reasoning Lab as a research intern.
2026.06	I’m attending CVPR 2026, presenting two papers: Geometric Stylization and ScaleEdit.

Research Overview

My research focuses on generative visual computing, where generative models are used to control visual content. I have studied how generative models can incorporate user intent and be adapted to diverse tasks. You can find the my publications here.

Recently, I have been exploring representation learning mechanisms that could make intelligent systems more interpretable. This direction builds on my previous work by moving from controlling the knowledge contained in models toward understanding how useful representations are learned.

Research overview. Click to zoom.

Explore More

Full Profile

Detailed research vision, full publication list, education, research experience, awards, and services.

Full Profile

Collaboration

I am grateful to all collaborators from both ongoing and previous projects. I'm also open to new collaborations.

Collaboration