Audrey Huang

Papers

Total Citations

H-Index

About

Audrey Huang is a rising researcher at the intersection of robotics, computer vision, and imitation learning, with a focus on enabling machines to learn complex behaviors from visual demonstrations. Her most-cited work, "Graph-Structured Visual Imitation" (2019, 14 citations), introduces a novel framework that reimagines imitation learning as a visual correspondence problem. By rewarding a robotic agent when its actions align the relative spatial configurations of visual entities—detected in both its workspace and a teacher’s demonstration—Huang bridges perception and control in a structured, graph-based manner. This approach leverages advances in computer vision to allow robots to generalize from single demonstrations, reducing the need for extensive training data. Though early in her career, Huang’s contributions are notable for their conceptual clarity and practical potential, offering a pathway toward more adaptable and sample-efficient robotic learning. Her work has been recognized within the imitation learning community for its innovative use of spatial reasoning, and it continues to inspire further research in visual correspondence and robot skill acquisition.