Preserving Data Privacy via Federated Learning: Challenges and Solutions

Zengpeng Li, Vishal Sharma*, Saraju P. Mohanty

*Corresponding author for this work

Research output: Contribution to specialist publicationArticle

11 Citations (Scopus)

Abstract

Data have always been a major priority for businesses of all sizes. Businesses tend to enhance their ability in contextualizing data and draw new insights from it as the data itself proliferates with the advancement of technologies. Federated learning acts as a special form of privacy-preserving machine learning technique and can contextualize the data. It is a decentralized training approach for privately collecting and training the data provided by mobile devices, which are located at different geographical locations. Furthermore, users can benefit from obtaining a well-trained machine learning model without sending their privacy-sensitive personal data to the cloud. This article focuses on the most significant challenges associated with the preservation of data privacy via federated learning. Valuable attack mechanisms are discussed, and associated solutions are highlighted to the corresponding attack. Several research aspects along with promising future directions and applications via federated learning are additionally discussed.

Original languageEnglish
Pages8-16
Number of pages9
Volume9
No.3
Specialist publicationIEEE Consumer Electronics Magazine
DOIs
Publication statusPublished - 01 May 2020

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Hardware and Architecture
  • Computer Science Applications
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Preserving Data Privacy via Federated Learning: Challenges and Solutions'. Together they form a unique fingerprint.

Cite this