About Me
I am Songchun Zhang (张菘淳), a first-year Ph.D. student at MMLAB@HKUST, supervised by Prof. Anyi Rao and Prof Huamin Qu. I obtained my master’s degree at Zhejiang University, supervised by Prof. Chunhui Zhao, where I also collaborated with Prof. Wei-Wei Xu and Prof. Changqing Zou at the State Key Lab of CAD&CG. Previously, I received my Bachelor’s degree from Hunan University.
My research focuses on multimodal real-time interactive world models for embodied intelligence and game prototyping:
- Embodied Intelligence: Building world models that enable agents to perceive, reason, and interact with physical and virtual environments in real-time.
- Game Prototyping: Developing generative systems for rapid creation and iteration of interactive game content and mechanics.
🔥 News
- 2025.06: 🎉🎉 One paper accepted by ICCV 2025.
- 2024.12: 🎉🎉 One paper accepted by AAAI 2025.
- 2024.02: 🎉🎉 One paper accepted by CVPR 2024.
- 2023.12: 🎉🎉 One paper accepted by TCSVT.
📝 Publications
[Arxiv] Astrolabe: Steering Forward-Process Reinforcement Learning for Distilled Autoregressive Video Models
Songchun Zhang, Zeyue Xue, Siming Fu, Jie Huang, Xianghao Kong, Yue Ma, Haoyang Huang, Nan Duan, Anyi Rao.
[Project page]
[Paper]
[Code]
- This paper presents an efficient online RL framework for aligning distilled autoregressive video models with human visual preferences.
[CVPR 2026 Highlight] Composing Concepts from Images and Videos via Concept-prompt Binding
Xianghao Kong, Zeyu Zhang, Yuwei Guo, Zhuoran Zhao, Songchun Zhang, Anyi Rao.
[Project page]
[Paper]
- This paper presents a one-shot method for flexible visual concept composition by binding visual concepts with prompt tokens.
[ICCV 2025] SpatialCrafter: Unleashing the Imagination of Video Diffusion Models for Scene Reconstruction from Limited Observations
Songchun Zhang, Huiyao Xu, Sitong Guo, Zhongwei Xie, Pengwei Liu, Hujun Bao, Weiwei Xu, Changqing Zou.
[Project page]
[paper]
- This paper presents a 3D scene reconstruction method from sparse inputs.
[AAAI 2025] Pragmatist: Multiview Conditional Diffusion Models for High-Fidelity 3D Reconstruction from Unposed Sparse Views
Songchun Zhang, Chunhui Zhao.
[Paper]
[Code]
- This paper presents a 3D object reconstruction method from sparse and unposed inputs.
[CVPR 2024] 3D-SceneDreamer: Text-Driven 3D-Consistent Scene Generation
Songchun Zhang, Yibo Zhang, Quan Zheng, Rui Ma, Wei Hua, Hujun Bao, Weiwei Xu, Changqing Zou.
[Project page]
[Paper]
[Code]
- This paper presents a novel text-driven 3D scene generation method that improves visual quality and 3D consistency.
[TCSVT 2023] Cross-Video Contextual Knowledge Exploration and Exploitation for Ambiguity Reduction in Weakly Supervised Temporal Action Localization
Songchun Zhang, Chunhui Zhao.
[Paper]
[Code]
- This paper presents a weakly-supervised action localization framework leveraging cross-video information.
🖥️ Experience
- March 2024 - Sept. 2024
Research Intern - Anti-Entropy Research Group, miHoYo
Advisor: Cheng Lin
Research included: 3D Scene Generation, Video World Model - Sept. 2023 - Feb. 2024
Research Intern - Taobao and Tmall Group, Alibaba
Research included: Sparse View Object Reconstruction - April 2023 - Dec. 2023
Research Assistant - State Key Lab of CAD&CG, Zhejiang University
Advisor: Prof. Changqing Zou and Prof. Weiwei Xu
Research included: Text-Guided 3D Generation - Sept. 2021 - April 2022
Research Intern - OpenDriveLab of Shanghai AI Laboratory
Advisor: Prof. Hongyang Li and Xiangwei Geng
Research included: Self-supervised Depth Estimation
🎓 Academic Service
- Reviewing
- Conferences: CVPR, ICRA, NeurIPS, AAAI, Siggraph Asia
- Journals: TCSVT, TMM, KBS
🎖 Honors and Awards
- 2024.12 Outstanding Graduate of Zhejiang University
- 2021.12 National Scholarship (Top 1% among all undergraduates)
- 2020.12 First Prize in China Undergraduate Mathematical Contest in Model (Top 0.1% among all undergraduates)
- 2020.12 National Scholarship (Top 1% among all undergraduates)
- 2019.12 National Scholarship (Top 1% among all undergraduates)