Yian Wang
Papers
4
Total Citations
73
H-Index
3
About
Yian Wang is a rising star in robotics and embodied AI, whose research focuses on enabling robots to perceive and manipulate 3D articulated objects—like cabinets, doors, and faucets—in unstructured human environments. His major contributions center on learning manipulation affordances and action trajectories for these complex objects, which are notoriously difficult due to their diverse geometries and joint mechanisms. Wang’s work on AdaAfford (36 citations) pioneered a few-shot interaction framework that allows robots to adapt their understanding of how to grasp and operate an object after just a handful of physical interactions, dramatically improving generalization to unseen instances. His earlier VAT-Mart paper (25 citations) introduced visual action trajectory proposals, effectively teaching robots to plan and execute the precise motions needed to open or close articulated objects. Most recently, Wang co-authored RoboGen (10 citations), a generative simulation framework that leverages foundation models to automatically synthesize infinite training data for diverse robotic skills, addressing the data scarcity bottleneck in robot learning. With his work bridging perception, interaction, and scalable simulation, Wang is helping to build the foundational capabilities for future home-assistant robots.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4