Edward Yao Li
Papers
1
Total Citations
7
H-Index
1
About
Edward Yao Li is at the forefront of robotic manipulation, with a particular focus on enabling machines to intelligently use tools in unstructured environments. His research lies at the intersection of computer vision, point cloud processing, and imitation learning, tackling the fundamental challenge of translating 3D geometric understanding into dexterous physical actions. Li’s major contribution is the development of ToolFlowNet, a pioneering framework that learns to predict tool flow directly from point clouds, allowing robots to grasp and manipulate objects with precision. This work, published in 2022, has already garnered 7 citations, signaling its growing influence in the robotics community. By addressing the long-standing difficulty of policy learning from point cloud data—a modality rich in geometric detail but notoriously hard to leverage for action generation—Li has opened new pathways for more adaptable and capable robotic systems. His approach stands out for its ability to bypass traditional segmentation and classification steps, directly mapping raw 3D inputs to manipulation policies. For students and researchers, Li’s work exemplifies how deep learning and geometric reasoning can converge to push the boundaries of autonomous tool use, a critical step toward robots that can assist in complex real-world tasks.
Research Focus
Key Achievements
Top Papers
- 1