Xinzhu Liu

Papers

Total Citations

H-Index

About

Xinzhu Liu is a rising researcher at the forefront of embodied AI and multi-agent robotics, with a focus on enabling intelligent agents to perceive, navigate, and collaborate in complex, real-world environments. Her work tackles fundamental challenges in visual semantic navigation, where she has pioneered methods that allow robots to locate target objects using only egocentric vision and scene prior knowledge, moving beyond the limitations of single-agent systems. Liu’s most cited paper (33 citations) introduces a novel framework for multi-agent embodied visual semantic navigation, a critical step toward deploying robot teams in dynamic settings. She further advances human-robot collaboration by addressing the critical problem of ambiguous instruction interpretation, developing planning systems that allow robots to infer implicit human intent. Her research also explores cross-modal learning, using self-supervised techniques to align visual and audio data for sound source separation, a key capability for scene understanding. With a growing body of work that bridges perception, planning, and multi-agent coordination, Liu is shaping the future of how robots interact with both their environment and the humans they serve.