Yanzhe Lyu

I'm a first-year PhD student in Computer Science at Stanford University, currently rotating with Prof. Jiajun Wu. Previously, I was an undergraduate student in the School of the Gifted Young at the University of Science and Technology of China, where I was advised by Prof. Xuejin Chen. In summer 2025, I visited Stanford through the UGVR program, working with Profs. Jiajun Wu and Shangzhe Wu.

My research interests broadly lie in generative AI and intelligent systems, particularly in developing AI systems that can understand, generate, reason about, and interact with the physical world. I am especially interested in 3D/4D vision, embodied intelligence, and agentic systems.

Email  /  GitHub

Photo of Yanzhe Lyu

Research

DyCraft animated teaser
DyCraft: An Agentic System for Long-Horizon 4D Scene Generation
Jeff Tan*, Hang Dai*, Zeqian Long*, Yanzhe Lyu, R. Kenny Jones, Shangzhe Wu, Chen Geng†, Jiajun Wu†
arXiv 2026
[Arxiv]   [Website]   [Code]

tl;dr: We generate long-horizon 3D animations from text using coding agents that plan scenes, verify 3D storyboards, and author motion between key poses.

Choreographing a World of Dynamic Objects
Yanzhe Lyu*, Chen Geng*, Karthik Dharmarajan, Yunzhi Zhang, Hadi Alzayer, Shangzhe Wu, Jiajun Wu
CVPR 2026
[Arxiv]   [Website]   [Code]

tl;dr: We propose a framework that generates motion for multi-object scenes conditioned on a text prompt, modeling both object deformation and inter-object interactions.

ResGS teaser
ResGS: Residual Densification of 3D Gaussian for Efficient Detail Recovery
Yanzhe Lyu, Kai Cheng, Xin Kang, Xuejin Chen
ICCV 2025
[Arxiv]   [Website]   [Code]

tl;dr: We identify a key limitation in 3D-GS densification and introduce a new densification strategy together with a coarse-to-fine training pipeline.

Selected Awards

[2021] National Olympiad in Informatics (NOI), Silver medal

Last updated: Nov. 2026
Template from Jon Barron