Jiyao Zhang

Jiyao Zhang is a PhD student at Peking University's Center on Frontiers of Computing Studies, advised by Prof. Hao Dong, where he focuses on Embodied AI, robotics, and 3D vision. He is best known for his work on category-level object pose estimation using diffusion models (GenPose, NeurIPS 2023), depth simulation and restoration for transparent objects (DREDS, ECCV 2022), and general robotic manipulation via object-centric interaction primitives (OmniManip, CVPR 2025). He has also contributed to universal 6D object pose estimation and tracking (Omni6dpose) and dexterous grasping in cluttered environments (ClutterDexGrasp).