Guanzhi Wang
Papers
7
Total Citations
427
H-Index
6
About
Guanzhi Wang is a leading researcher at the intersection of robotics, simulation, and foundation models, whose work is shaping how robots learn to interact with the physical world. He is best known for developing **iGibson**, a high-fidelity simulation environment for interactive tasks in large, realistic scenes. With over 177 combined citations, iGibson provides 15 fully interactive home-sized scenes populated with rigid and articulated objects, enabling reproducible research in embodied AI. Wang’s contributions extend to large-scale robotic learning through the **Open X-Embodiment** collaboration (119 citations), which produced the RT-X models—a landmark effort in consolidating diverse robotic datasets for general-purpose manipulation. He also pioneered **VIMA** (65 citations), introducing multimodal prompts for general robot manipulation, and **Eureka** (48 citations), a method that uses coding LLMs to automatically design reward functions, achieving human-level dexterity in tasks like pen spinning. His recent work on **GR00T N1** targets open foundation models for generalist humanoid robots. Wang’s research consistently bridges simulation and real-world deployment, advancing the frontier of generalizable, data-driven robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3VIMA: General Robot Manipulation with Multimodal Prompts65 citations · 2022
- 4Eureka: Human-Level Reward Design via Coding Large Language Models48 citations · 2023
- 5
- 6SECANT: Self-Expert Cloning for Zero-Shot Generalization of Visual Policies14 citations · 2021
- 7GR00T N1: An Open Foundation Model for Generalist Humanoid Robots4 citations · 2025