Detecting Fake News via Deep Learning Techniques

Abstract:
The problem caused by the rapid growth of news information is that it brings too much fake news. Fake news negatively impacts both economy and society, even causes public panic, and threatens public security. And the huge amount of data makes it more difficult to identify fake news manually. Hence, how to automatically detect false news from both the internet and the traditional public media while preventing the spread of fake news in time attracts research in various fields. In this paper, we revise some state-of-the-art work on fake news detection. Specifically, we first describe and compare the definition of both fake news and fake news detection. Secondly, an introduction of the metrics and datasets of fake news detection is shown. And then, we revise some deep learning-based fake news detection. These works are divided into two categories following the modality, namely, single-modal methods and multi-modal methods. Finally, we analyse some potential and important challenges of fake news detection. The review cannot provide an introduction for researchers but also a good reference for rookies.