Xinyu Zhan

Papers

Total Citations

H-Index

About

Xinyu Zhan is a rising researcher in robotics and embodied AI, whose work centers on advancing robot imitation learning through a deeper understanding of object dynamics. Her most-cited paper, "Motion Before Action: Diffusing Object Motion as Manipulation Condition" (2025), introduces a paradigm-shifting approach that prioritizes reasoning about object motion from visual observations before generating action sequences. This method, dubbed MBA, significantly enhances robotic manipulation performance by enabling robots to infer and predict object movements, leading to more intuitive and effective task execution. With 3 citations in its early publication year, this work is already gaining traction for its novel integration of motion diffusion models into manipulation pipelines. Zhan’s research sits at the intersection of computer vision, reinforcement learning, and robotics, aiming to bridge the gap between perception and action. Her contributions promise to make robots more adaptable and capable in unstructured environments, offering a fresh perspective on how machines can learn from human demonstrations. As a young scholar, Zhan is poised to make lasting impacts on the future of autonomous systems and human-robot collaboration.