EdgeOL: Efficient in-situ Online Learning on Edge Devices. (arXiv:2401.16694v1 [cs.LG])

Emerging applications, such as robot-assisted eldercare and object
recognition, generally employ deep learning neural networks (DNNs) models and
naturally require: i) handling streaming-in inference requests and ii) adapting
to possible deployment scenario changes. Online model fine-tuning is widely
adopted to satisfy these needs. However, fine-tuning involves significant
energy consumption, making it challenging to deploy on edge devices. In this
paper, we propose EdgeOL, an edge online learning framework that optimizes
inference accuracy, fine-tuning execution time, and energy efficiency through
both …

applications arxiv consumption cs.lg deep learning deep learning neural networks deployment devices edge edge devices energy fine-tuning inference making model fine-tuning networks neural networks online learning recognition robot streaming