Xuweiyi Chen

University of Michigan–Ann Arbor

Papers

3

Total Citations

67

H-Index

2

About

Xuweiyi Chen is at the forefront of integrating large language models (LLMs) with 3D perception, pioneering new capabilities for embodied AI and household robotics. His primary research focuses on open-vocabulary 3D visual grounding and reducing hallucination in 3D-LLMs. Chen’s landmark work, **LLM-Grounder** (2024, 60 citations), introduces a novel agent-based framework that leverages an LLM to decompose complex language queries and orchestrate specialized 3D perception modules, enabling robots to navigate and manipulate objects without requiring extensive labeled data. This approach overcomes a critical limitation of prior methods, allowing for robust, zero-shot grounding in dynamic environments. Building on this, Chen co-created **3D-GRAND** (2025), a million-scale dataset designed to improve grounding fidelity and significantly reduce hallucination in 3D-LLMs, providing a vital benchmark for the field. His contributions directly address the core challenge of making robots truly understand and interact with the physical world through natural language. With his work already garnering significant early attention, Chen is establishing himself as a leading voice in the crucial intersection of language, 3D vision, and embodied intelligence.

Research Focus

Key Achievements

2
H-Index
3
Papers
67
Total Citations
22
Avg Citations/Paper
🏆 Most Cited Paper
LLM-Grounder: Open-Vocabulary 3D Visual Grounding with Large Language Model as an Agent
60 citations · 2024
📈 Most Prolific Year: 2024 (1 Papers)
🤝 Key Collaborators: 6
🏛 Institutions: University of Michigan–Ann Arbor

Top Papers

  1. 1
  2. 2
  3. 3

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 13 days ago