Xingyue Zhao
Xingyue Zhao is a researcher affiliated with Alibaba DAMO Academy, with a background spanning Peking Union Medical College Hospital and Xi'an Jiaotong University. Zhao is known for research at the intersection of medical image analysis, 3D vision, and robotics, including work on high-fidelity simulated data generation for zero-shot robotic manipulation using Gaussian Splatting and affordance-aware robotic dexterous grasping. Zhao has published at major AI conferences such as AAAI, ICCV, and BMVC.
“The paper introduces a paradigm where an operator's hand-pose stream drives a generative world model to synthesize egocentric robot videos, while the same gesture stream is retargeted into robot actions — producing complete (observation, action) trajectory pairs without ever moving a physical robot.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.