Shiyan Du

Sun Yat-sen University

Research Interests

Image/Video Generation

Multimodal Understanding & MLLMs

Reinforcement Learning & Post-Training

About Me

Hi👋. I am a third-year master student at the CSE, Sun Yat-sen University (SYSU), advised by Prof. Dongyu Zhang. I am a member of HCP-LAB. Prior to this, I received my B.Eng. degree in Aeronautics and Astronautics Engineering from SYSU in 2024.

My research interests include generative models, image and video generation, multimodal understanding with MLLMs, and reinforcement learning for generative models.

I am currently a research intern at Tencent (PCG, Tencent Video), working on video foundation models.

Selected Publications

Temporally Grounded Compositional Camera Motion Understanding via Geometric Knowledge Distillation

Temporally Grounded Compositional Camera Motion Understanding via Geometric Knowledge Distillation

Dazhao Du, Shiyan Du, Jian Liu, Yongjian Yu, Bohai Gu, Tao Han, Hualuo Liu, Eric Liu, Yujia Zhang, Xi Chen, Song Guo

arXiv Preprint

DEIG: Detail-Enhanced Instance Generation with Fine-Grained Semantic Control

DEIG: Detail-Enhanced Instance Generation with Fine-Grained Semantic Control

Shiyan Du, Conghan Yue, Xinyu Cheng, Dongyu Zhang

AAAI 2026, Oral

Improving Fine-Grained Control via Aggregation of Multiple Diffusion Models

Improving Fine-Grained Control via Aggregation of Multiple Diffusion Models

Improving Fine-Grained Control via Aggregation of Multiple Diffusion Models

Conghan Yue, Zhengwei Peng, Shiyan Du, Zhi Ji, Chuangjian Cai, Le Wan, Dongyu Zhang

arXiv Preprint

Image Restoration Through Generalized Ornstein-Uhlenbeck Bridge

Image Restoration Through Generalized Ornstein-Uhlenbeck Bridge

Conghan Yue, Zhengwei Peng, Junlong Ma, Shiyan Du, Pengxu Wei, Dongyu Zhang

ICML 2024

News

2026-08

One paper has been released on arXiv, the code and datasets are now available!

2025-11

One paper has been accepted by AAAI 2026 as an Oral presentation 🎉

2024-10

One paper (AMDM) has been released on arXiv, the code is now available!

2024-05

One paper (GOUB) has been accepted by ICML 2024 🎉