Reproduce “On-Device Training Empowered Transfer Learning For Human Activity Recognition” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Pixi Kang, Julian Moosmann, Sizhen Bian, Michele Magno
The paper presents on-device transfer learning to personalize human-activity recognition for a new user. It uses a 1D-CNN with a fixed backbone and updates only its final classifier with labeled user samples across three HAR settings on STM32F756ZG and GAP9 devices. In QVAR gesture recognition, accuracy rises from 67.12% to 84.50% under the evaluated protocol; Ultra’s seventh fold declines.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.