About Me
I am a Ph.D. candidate at the Wangxuan Institute of Computer Technology, School of Intelligence, Peking University, supervised by Prof. Jiahuan Zhou. I received my B.Eng. in Software Engineering from the College of Computer Science and Technology, Zhejiang University, in 2025.
My research interests lie in 3D vision and graphics, with a focus on point cloud processing, scene understanding, and neural scene generation. I am particularly interested in developing efficient, geometry-aware models that unify perception and generation within a single framework. Ultimately, my goal is to build powerful real-world 3D perception models that seamlessly integrate understanding and generation capabilities, and can be deployed on humanoid robot platforms for intelligent interaction in complex environments.
Previously, I was a research intern at the State Key Laboratory of CAD&CG, Zhejiang University, where I worked with Prof. Zhaopeng Cui on large-scale 3D scene reconstruction. These experiences have strengthened my passion for advancing both the theoretical foundations and practical applications of 3D visual understanding and generation.
News
- 2025.06: One paper accepted by ICCV 2025.
- 2025.05: Two papers accepted by ICML 2025.
- 2024.12: One paper accepted by AAAI 2025.
- 2024.09: Joined the School of Intelligence, Peking University. 🎉🎉
Honors and Awards
- 2025.06: Outstanding Graduate of Zhejiang Province, 2025
- 2025.05: Outstanding Graduate of Zhejiang University, 2025
- 2024.10: China National Scholarship, 2024
- 2024.10: First Class Scholarship, Zhejiang University, 2024
Education
- 2025.08 - 2030.06 (Expected): PhD Candidate in Intelligent Science and Technology, School of Intelligence, Peking University.
- 2021.09 - 2025.07: Bachelor in Software Engineering, College of Computer Science and Technology, Zhejiang University.
Publications
UPP: Unified Point-Level Prompting for Robust Point Cloud Analysis
ICCV 2025
Zixiang Ai, Zhenyu Cui, Yuxin Peng, Jiahuan Zhou†
- We propose a unified point-level prompting method that reformulates point cloud denoising and completion as a prompting mechanism, enabling robust analysis in a parameter-efficient manner.
Vision Graph Prompting via Semantic Low-Rank Decomposition
ICML 2025
Zixiang Ai, Zichen Liu, Jiahuan Zhou†
- In this paper, we propose Vision Graph Prompting (VGP), a novel framework tailored for vision graph structures. Our core insight reveals that semantically connected components in the graph exhibit low-rank properties.
GAPrompt: Geometry-Aware Point Cloud Prompt for 3D Vision Model
ICML 2025
Zixiang Ai, Zichen Liu, Yuanhang Lei, Zhenyu Cui, Xu Zou, Jiahuan Zhou†
- In this paper, we propose a novel Geometry-Aware Point Cloud Prompt (GAPrompt) that leverages geometric cues to enhance the adaptability of 3D vision models.
GURecon: Learning Detailed 3D Geometric Uncertainties for Neural Surface Reconstruction
AAAI 2025 (Oral)
Zesong Yang*, Jiale Shi*, Ru Zhang, Zixiang Ai, Boming Zhao, Hujun Bao, Luwei Yang, Zhaopeng Cui†
- In this paper, we present a novel framework, i.e, GURecon, which establishes a geometric uncertainty field for the neural surface based on geometric consistency.