European Academic Research ISSN 2286-4822
ISSN-L 2286-4822
Impact Factor: 3.4546 (UIF)
DRJI Value : 5.9 (B+)
Article Details :
Article Name :
Federated Learning Approaches for Edge Computing and IoT Cybersecurity: An Investigation
Author Name :
Shahad Sharaf Aldeen Yahya, Dr. Essa Ibrahim Essa
Publisher :
Bridge Center
Article URL :
Abstract :
Federated learning with edge computing and the internet of things (IoT) is considered for energy-efficient privacy-enhancing cybersecurity. The upshot of the?review is that federated learning allows training of shared models without exchanging any raw data with the cloud and reduces risk of leaks and improves privacy (in particular in multi-device settings where device capabilities and network connectivity vary). The paper investigates how edge computing can support the placement of intelligence near sensors and IoT objects to?reduce response times and enable real-time security responses (e.g., threats detection, anomaly identification). It differentiates between centralized and decentralized/diffused learning methodologies, and improves understanding of the merits of decentralized approaches for eradicating centralized?single points of failure and enhancing attack resilience, and the challenges of coordinating and harmonizing models. In conclusion, the paper identifies promising research directions for future work in terms of?the more pervasive use of decentralized learning-based approaches for enhancing edge security for IoT systems.
Keywords :
Federation learning, Edge Computing, Internet of Thing, Cybersecurity, Cross Device, Decentralized Learning.
Announcements
New Launched Project
onlineresearch
Recommend & Share
Copyright © 2013: www.euacademic.org , Best Viewed in IE7+, Chrome, Mozilla Firefox.