Xujin He
Papers
1
Total Citations
4
H-Index
1
About
Xujin He is a researcher at the forefront of embodied AI and 3D perception, whose work bridges the gap between common-sense reasoning and multi-sensory robotic understanding. His most notable contribution, FusionSense, pioneers a novel framework that integrates foundation model priors with sparse visual and tactile observations, enabling robots to perform robust 3D reconstruction from remarkably limited data—a feat that mimics human intuitive spatial reasoning. This work, already garnering 4 citations since its 2025 publication, addresses a critical bottleneck in robotics: how to reconstruct complete environments from partial sensory input. He’s key research areas span 3D scene understanding, multi-modal sensor fusion, and the application of large pre-trained models to physical world tasks. By showing how common-sense knowledge can compensate for sparse sensory data, He’s research has immediate implications for manipulation in cluttered or occluded settings, where full visibility is impossible. His approach stands out for its elegance in leveraging existing foundation models rather than requiring massive task-specific training data, pointing toward more adaptable, human-like robotic perception systems.
Research Focus
Key Achievements
Top Papers
- 1