Deep Learning Models to Detect Online False Information: a Systematic Literature Review

Abstract

The amount of disseminated information from online content volume is increasing rapidly including trusted and untrusted information that are published by different sources. To counter this problem, we need a comprehensive knowledge of existing methods and techniques emerging in the area of False News Detection (FND).This research survey provides a comprehensive review of most effective Deep Learning (DL) models that are used to detect false news and information. We are focusing in DL models and techniques, which use the textual published content and perform FND based on content features. We have considered the research papers in the last five years starting from 2017 onward. In this research paper, the published articles about proposing and developing FND based DL models are included whether the dataset are collected from social platforms or extracted from other news sources. In addition, this research study helps the researchers to have a complete view of the developed DL models that have been proposed in the field of false information detection, the DL models gaps in FND and how they can be improved.