About

Fei Xia is a pioneering researcher at the intersection of embodied AI, robotics, and large language models, whose work has fundamentally shaped how intelligent agents perceive and interact with the physical world. Best known for bridging simulation and real-world robotics, Xia co-developed the Gibson Environment (714 citations), a landmark platform enabling embodied agents to learn visual perception and sensorimotor control in realistic virtual settings. This laid the groundwork for iGibson 1.0, a richly interactive simulation suite that advanced benchmark standards for household robotic tasks. Xia's more recent contributions have been transformative in applying large language models to robotic control. His work on "Do As I Can" (516 citations), "Code as Policies" (561 citations), and "Inner Monologue" (206 citations) demonstrated how LLMs can be grounded in physical affordances and used to generate executable robot behaviors from natural language. His contributions to RT-1 (512 citations), RT-2 (267 citations), and PaLM-E (350 citations) further established scalable, transformer-based architectures as the foundation of modern generalist robots. With thousands of cumulative citations, Xia's research has made him one of the most influential voices defining the future of language-guided, real-world robotic intelligence.

Research Focus

Key Achievements

29
H-Index
61
Papers
4,988
Total Citations
82
Avg Citations/Paper
🏆 Most Cited Paper
Gibson Env: Real-World Perception for Embodied Agents
714 citations · 2018
📈 Most Prolific Year: 2023 (19 Papers)
🤝 Key Collaborators: 355
🏛 Institutions: Stanford University, Google (United States), Google DeepMind (United Kingdom), Henan Polytechnic University, Georgia Institute of Technology

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 33 days ago