Human-aware
machines.
Yiteng Xu is a ShanghaiTech Ph.D. researcher creating visual intelligence systems that can understand people, motion, and real-world interaction.
Research as a sequence of systems, datasets, and human questions.
ReMoGen
Responsive reaction motion generation across human-human and human-scene interaction.
SymBridge
Real-to-sim human-robot learning with an adaptive cyber-physical interactive system.
UniPVU-Human
A unified and efficient framework for human-centric point cloud video understanding.
HuCenLife
Human-centric scene understanding for 3D large-scale scenarios.
From 3D perception to interactive learning.
My research sits at the intersection of 3D Computer Vision, Robotics, and Embodied AI. I build sensing and learning systems that connect visual understanding with meaningful physical interaction.
Beyond papers, this site also contains practical resources, including LinuxDO information digests, a curated Shanghai internship page, and travel projects.