Yunchao Yao
Papers
2
Total Citations
12
H-Index
2
About
Yunchao Yao is a researcher advancing the frontier of robotic manipulation through innovative work in 3D visual representation learning and deformable object control. His research primarily spans visual reinforcement learning (visual RL) and deformable object manipulation, with a focus on bridging the gap between 2D and 3D representations for more robust robotic perception. In his highly cited work "On the Efficacy of 3D Point Cloud Reinforcement Learning" (2023, 6 citations), Yao systematically compared 3D point cloud representations with traditional 2D methods across diverse tasks, revealing critical insights into when and why 3D representations outperform their 2D counterparts—a foundational contribution to the visual RL community. His second major contribution, "DeformGS: Scene Flow in Highly Deformable Scenes for Deformable Object Manipulation" (2023, 6 citations), tackles the notoriously difficult problem of manipulating non-rigid objects like cloth. By developing a scene flow approach that handles frequent occlusions and infinite-dimensional state spaces, Yao has opened new pathways for automating tasks such as folding and draping. His work stands out for its rigorous experimental methodology and practical relevance to real-world robotics, earning recognition as a rising voice in embodied AI and robotic manipulation.
Research Focus
Key Achievements
Top Papers
- 1On the Efficacy of 3D Point Cloud Reinforcement Learning6 citations · 2023
- 2