EdgeFNF: Toward Real-time Fake News Detection on Mobile Edge Computing

Abstract

Fake news (FN) spreads faster than ever due to social networks' ease of access, increasing reach, and lower cost. Twitter and Facebook are the most used platforms, allowing users to express news in short, simple lines that can be fake using their smartphones. Hence, real-time prediction and fast response time are vital in spotting FN and opposing its negative impact. However, smartphones have limited computational capabilities besides unreliable network connections. Relying on the amalgamation of the edge, fog, and cloud computing can relieve the previous bottleneck where computation offloads from edge devices to higher network layers on demand. In this paper, we proposed EdgeFNF, an edge fake news finder approach toward a fully Edge-to-Cloud mobile architecture. EdgeFNF collects data from social media platforms, e.g., tweets and posts, preprocess them on the mobile edge node, and uploads the metadata into a cloud server where multiple data processing techniques for text, such as Natural Language Processing (NLP), take place. Henceforth, detect fake news using NLTK and BERT algorithms. We provide the methodology, system architecture, and merits for achieving real-time, accurate detection of fake news.