Research Area:  Machine Learning

Abstract:

In this article, we study both communication and computation issues in edge networks by emphasizing data privacy in a smart home scenario. We design an architecture that incorporates federated edge learning to promote data privacy and a node weighting and dropping scheme to select the appropriate participating devices with quality data and therefore improve the training and reduce communication cost. We further apply Long Short-Term Memory to predict future tasks and proactively store them locally at the edge device. We adopt the computation reuse concept to satisfy incoming tasks with lessto-no computation and thus eliminating redundant computation and further decreasing the computation cost. Simulation results based on real-world dataset show the effectiveness and efficiency of the proposed architecture. The training phase is reached with few iterations, while computation and communication are reduced by up to 80 percent and 70 percent, respectively, compared with existing schemes while data privacy is promoted.

Keywords:  

Author(s) Name:  Boubakr Nour; Soumaya Cherkaoui; Zoubeir Mlika

Publisher name:  IEEE

DOI:  10.1109/TNSE.2021.3131246

Volume Information:  

Paper Link:   https://ieeexplore.ieee.org/document/9628038