Weichen Xu
Papers
2
Total Citations
7
H-Index
2
About
Weichen Xu is a rising researcher at the forefront of embodied intelligence and multimodal 3D scene understanding. His work bridges the gap between vision-language models and real-world robotic applications, with a particular focus on how machines can learn from physical interactions in home environments. Xu’s most-cited paper, “A survey of language-grounded multimodal 3D scene understanding” (2025, 5 citations), provides a comprehensive taxonomy of methods that align linguistic concepts with 3D spatial data—a critical step toward generalist robots that can follow natural language commands. In his notable work “SweepMM: A High-Quality Multimodal Dataset for Sweeping Robots in Home Scenarios for Vision-Language Model” (2024, 2 citations), Xu identified a key bottleneck: existing vision-language models lack domain-specific knowledge about household cleaning tasks. By curating a specialized dataset of sweeping robot interactions, he enables models to understand context like “avoid the rug” or “clean under the table.” Though early in his career, Xu’s contributions are already shaping how researchers approach embodied AI, proving that high-quality, task-specific data is essential for moving beyond generalized models toward truly intelligent home assistants.
Research Focus
Key Achievements
Top Papers
- 1A survey of language-grounded multimodal 3D scene understanding5 citations · 2025
- 2