Yiyuan Lee

Papers

Total Citations

H-Index

About

Yiyuan Lee is a roboticist whose research lies at the intersection of motion planning, manipulation, and decision-making under uncertainty. His work focuses on developing algorithms that enable robots to operate efficiently and robustly in complex, real-world environments. A central contribution is the **Generator-Critic framework**, introduced in his most-cited paper (6 citations), which adaptively guides sampling in high-degree-of-freedom motion planning—significantly improving exploration efficiency over traditional uniform methods. Lee further extended this framework to online POMDP planning with **MAGIC**, enabling robots to learn macro-actions for long-horizon tasks under sensing and actuation uncertainty. In manipulation, his work on **object reconfiguration** leverages simulation-derived feasible actions to solve physically grounded rearrangement problems, bridging the gap between simulation and reality. With over 10 citations across his key publications, Lee’s work is gaining traction for its practical impact on autonomous systems. His contributions are particularly notable for unifying learning and planning—a crucial step toward deploying robots that can adapt on the fly in unstructured settings.