Yinpei Dai

University of Michigan–Ann Arbor, Universitas Dharmas Indonesia

Papers

Total Citations

H-Index

About

Yinpei Dai is a leading researcher at the intersection of embodied AI, robot learning, and human-robot interaction. Their work focuses on enabling robots to operate intelligently in open-world environments through advanced language-guided navigation and manipulation. Dai’s landmark paper, “Think, Act, and Ask: Open-World Interactive Personalized Robot Navigation” (2024, 19 citations), introduces a paradigm for zero-shot object navigation where agents not only follow instructions but actively interact and ask questions to locate novel objects in unknown spaces—moving beyond rigid, pre-defined object classes. More recently, Dai’s “RACER: Rich Language-Guided Failure Recovery Policies for Imitation Learning” (2025, 3 citations) tackles a critical bottleneck in robotic manipulation: the lack of robust self-recovery mechanisms. By proposing a scalable data generation pipeline, RACER enables visuomotor policies to correct their own mistakes using rich language feedback, dramatically improving real-world task success. These contributions are reshaping how robots learn from human guidance and recover from errors, making Dai a rising authority in creating more adaptive, interactive, and resilient autonomous systems for complex, unstructured environments.