BriefGPT.xyz
Jul, 2023
不要冻结：微调编码器以获得更好的自监督HAR
Don't freeze: Finetune encoders for better Self-Supervised HAR
HTML
PDF
Vitor Fortes Rey, Dominique Nshimyimana, Paul Lukowicz
TL;DR
本文探讨了自监督学习在人类活动识别领域解决标记数据不足的问题，通过学习预测任务，可以得到有用的表示用于分类，并发现不冻结表示可以在预测任务中实现显著性能提升并且效果与标记数据量成反比。
Abstract
Recently
self-supervised learning
has been proposed in the field of
human activity recognition
as a solution to the
labelled data
availabi
→