Peiyan Li
Papers
2
Total Citations
10
H-Index
2
About
Peiyan Li is a rising researcher at the forefront of embodied AI and robot manipulation, with a focus on building generalist policies that bridge vision, language, and action. Their work addresses a critical bottleneck in robotics: the scarcity of fully annotated training data. In their highly cited paper "GR-MG: Leveraging Partially-Annotated Data via Multi-Modal Goal-Conditioned Policy" (2025, 8 citations), Li introduced a novel framework that enables robots to learn from partially labeled trajectories, significantly reducing the need for costly full annotations while maintaining high performance. This contribution directly tackles the time-intensive nature of data collection in robotics. Li further advanced the field with "Towards Generalist Robot Policies: What Matters in Building Vision-Language-Action Models" (2025, 2 citations), which systematically explores the key design choices for creating scalable, generalizable robot policies capable of following flexible natural language instructions. By addressing both data efficiency and model architecture, Li is shaping the path toward truly versatile robotic systems that can operate in unstructured human environments. Their work is already influencing how researchers approach the data scarcity problem in robot learning.
Research Focus
Key Achievements
Top Papers
- 1
- 2