Yucheng Mao

I am a first-year PhD student supervisied by Prof. Zhuowen Tu at UC San Diego. Previously, I did my undergraduate study at University of Science and Technology, Beijing. I also spent one year as a master student at UC San Diego.

My research interest lies in computer vision, generative models and multimodel learning.

Email  /  CV  /  Scholar  /  Github

profile photo

Internships

Shanghai AI Laboratory logo Shanghai AI Laboratory
Research Intern
Feb 2025 - Sep 2025

Mentors: Xudong Xu and Linning Xu.

Adobe logo Adobe Research
Research Intern
Jun 2026 - Sep 2026

Mentors: Hanwen Jiang, Hao Tan, and Chongjian Ge.

Research

* Denote equal contribution

Know, then Draw: DPG-Bench versus training compute and text-to-image samples
Know, then Draw: A New Learning Paradigm for Visual Generation
Yucheng Mao, Chongjian Ge, Zhuowen Tu, Hao Tan, Hanwen Jiang
arXiv, 2026
arXiv

Hierarchical representations in semantic space for visual generation.

QuadTok adaptive token allocation, image generation, and layout control
QuadTok: Quadtree Visual Tokenizer for Autoregressive Image Generation
Yucheng Mao *, Zeyuan Chen *, Xiaojun Shan, Xiang Zhang, Divyansh Srivastava, Bingnan Li, Zhuowen Tu
arXiv, 2026
Code / arXiv

Hierarchical representations for autoregressive image generation.

RoboVIP: Multi-View Video Generation with Visual Identity Prompting Augments Robot Manipulation
Boyang Wang *, Haoran Zhang *, Shujie Zhang *, Jinkun Hao, Mingda Jia, Qi Lv, Yucheng Mao Zhaoyang Lyu, Jia Zeng, Xudong Xu, Jiangmiao Pang,
arxiv, 2026
Project page / arXiv

Video Generation for Embodied AI

AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Views
Yucheng Mao *, Lihan Jiang *, Linning Xu, Tao Lu, Kerui Ren, Yichen Jin, Xudong Xu, Mulin Yu, Jiangmiao Pang, Feng Zhao, Dahua Lin, Bo Dai,
SIGGRAPH Asia 2025, ACM TOG
Project page / arXiv

3D Gaussian Splating within a minute

Sparse Image Sets Restoration with Multi-View Diffusion Model
Yucheng Mao *, Boyang Wang *, Nilesh Kulkarni, Jeong Joon Park
CVPR, 2025
Project page / arXiv

Reconstruct a 3D scene with degraded observation.

PreSight: Enhancing Autonomous Vehicle Perception with City-Scale NeRF Priors
Tianyuan Yuan, Yucheng Mao, Jiawei Yang, Yicheng Liu,
Yue Wang, Hang Zhao,
ECCV, 2024
Code / arXiv

A novel framework leverages past traversals to construct static prior, enhancing online perception in later navigations.

Occ3D: A Large-Scale 3D Occupancy Prediction Benchmark for Autonomous Driving
Xiaoyu Tian*, Tao Jiang*, Yunfei Long, Yucheng Mao, Huitong Yang, Yue Wang, Yilun Wang, Hang Zhao,
NeurIPS, 2023
Project page / arXiv

Open Source Contribution

DETR3D-MMDetection3D Framework
Third-Party Implementation Of UltraLiDAR
Third-Party Implementation Of GS-LRM

Website inspired by Jon Barron.