Sangwon Jang

Contact: sangwon.jang [at] kaist.ac.kr

About

I am a PhD student at KAIST (MLAI Lab), working under the supervision of Sung Ju Hwang.

My research focuses on computer vision and deep learning, particularly generative tasks using multimodal foundation models like text-to-video models. Currently, I’m interested video generative models and the applications of generative foundation models toward complex capabilities like world modeling as physical simulator and world action models (WAM).

Last updated: May. 1, 2026

Publications (*: equal contribution / ^: equal advising)

Self-Refining Video Sampling
[project] [paper] [code]
Sangwon Jang*, Taekyung Ki*, Jaehyeong Jo*, Saining Xie, Jaehong Yoon^, Sung Ju Hwang^
International Conference on Machine Learning (ICML), 2026

Avatar Forcing: Real-Time Interactive Head Avatar Generation for Natural Conversation
[project] [paper] [code]
Taekyung Ki*, Sangwon Jang*, Jaehyeong Jo, Jaehong Yoon, Sung Ju Hwang
Conference on Computer Vision and Pattern Recognition (CVPR), 2026

Frame Guidance: Training-Free Guidance for Frame-Level Control in Video Diffusion Model
[project] [paper] [code]
Sangwon Jang*, Taekyung Ki*, Jaehyeong Jo, Jaehong Yoon, Soo Ye Kim, Zhe Lin, Sung Ju Hwang
International Conference on Learning Representations (ICLR), 2026

Silent Branding Attack: Trigger-free Data Poisoning Attack on Text-to-Image Diffusion Models
[project] [paper] [code]
Sangwon Jang, June Suk Choi, Jaehyeong Jo, Kimin Lee^, Sung Ju Hwang^
Conference on Computer Vision and Pattern Recognition (CVPR), 2025

Identity Decoupling for Multi-Subject Personalization of Text-to-Image Models
[project] [paper] [code]
Sangwon Jang*, Jaehyeong Jo*, Kimin Lee^, Sung Ju Hwang^
Conference on Neural Information Processing Systems (NeurIPS), 2024

Text-Conditioned Sampling Framework for Text-to-Image Generation with Masked Generative Models
[paper]
Jaewoong Lee*, Sangwon Jang*, Jaehyeong Jo, Jaehong Yoon, Yunji Kim, Jin-Hwa Kim, Jung-Woo Ha, Sung Ju Hwang
International Conference on Computer Vision (ICCV), 2023

Preprints

Enhancing Variational Autoencoders with Smooth Robust Latent Encoding
[paper]
Hyomin Lee*, Minseon Kim*, Sangwon Jang, Jongheon Jeong, Sung Ju Hwang
Preprint, 2025