Li Zhang
Li Zhang is a Full Professor in the School of Data Science at Fudan University, where she leads the Logos Robotics Group and holds an affiliation with the Shanghai Innovation Institute. She is best known for her research at the intersection of robot learning, computer vision, and embodied AI, with over 27,000 citations and recent publications at venues including ICLR, CVPR, and ICRA. She earned her PhD from Queen Mary University of London and completed postdoctoral work at the University of Oxford under Philip Torr and Andrew Zisserman before joining Fudan. Her recent work includes the DeFI framework, a disentangled pretraining approach for robot manipulation that achieved state-of-the-art results on the CALVIN ABC-D benchmark, presented at ICLR 2026.
“DeFI achieves an average task length of 4.51 on CALVIN ABC-D (the hardest generalization split, training on ABC, testing on unseen environment D), beating the prior SOTA VPP at 4.33 (+4.2%), Physical Intelligence's π0 at 3.84, and GR00T N1 at 4.01.”
Source→“This decoupled pretraining paradigm unleashes the potential of massive action-free videos for policy learning, while retaining robot-specific action grounding.”
Source→“Li Zhang — Corresponding author. Oversees a lab working across video understanding and embodied AI. The institutional affiliation with the Shanghai Innovation Institute signals proximity to Chinese national AI infrastructure priorities.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.