Yishun Lu
Yishun Lu is a researcher in the Department of Engineering Science at the University of Oxford, affiliated with the Oxford e-Research Centre (OERC) and St Anne's College. He completed his DPhil at Oxford in 2024 and his research spans deep learning optimization, multimodal machine learning, and vision-language models. He is known for work on second-order optimization methods for large-scale model training and is a co-author on the VLA-AD paper 'Offline Semantic Guidance for Efficient Vision-Language-Action Policy Distillation,' which compresses large vision-language-action policies for robotics applications.
“Paper: 'Offline Semantic Guidance for Efficient Vision-Language-Action Policy Distillation' Authors: Jin Shi, Brady Zhang, Yishun Lu (UCL / Oxford)”
Source→“compressing a 7-billion-parameter VLA policy down to 158 million parameters while maintaining essentially identical task performance...matches the teacher with only a 0.27% average relative gap”
Source→“Yishun Lu — Department of Engineering Science, University of Oxford. Third author. Oxford's engineering science department has been active in robotic manipulation research.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.