Cewu Lu
Cewu Lu is a professor in the Department of Computer Science and Engineering at Shanghai Jiao Tong University, where he leads the Machine Vision and Intelligence Group. He is known for his research in embodied intelligence and computer vision, with work published in venues including Nature, Nature Machine Intelligence, and top robotics and AI conferences. He has contributed to open-source projects such as AlphaPose, HAKE, and GraspNet, and has received recognitions including the MIT TR35 China award and the Scientific Exploration Award. He previously served as a research fellow at Stanford University under Prof. Fei-Fei Li and Prof. Leonidas J. Guibas.
“FA-RDP: A Frequency-Adaptive Reactive Diffusion Policy for Contact-Rich Manipulation”
Source→“standard diffusion policies use a fixed inference frequency and sampling steps throughout the episode, forcing a fundamental compromise: low-frequency, multi-step sampling better preserves pre-contact multimodality but responds slowly to force feedback, whereas high-frequency sampling improves reactivity but tends to collapse distinct pre-contact modes.”
Source→“A company listed as an affiliation for researchers Jun Lv and Cewu Lu.”
Source→“Cewu Lu — Shanghai Jiao Tong University / Shanghai Innovation Institute / Noematrix Ltd. Co-advisor. A highly prolific researcher in the Physical AI space whose lab has produced an extraordinary density of recent force-aware manipulation work (FoAR, ForceMimic, RDP, ImplicitRDP, Force Policy, FTP-1, VTAM).”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.