Document Type : Original Article

Abstract

The spread of fake news on social networks has become a serious challenge in the fields of information and cybersecurity, particularly in the realm of passive defense. Early detection of such news can play a crucial role in improving cybersecurity and controlling the dissemination of misinformation. This paper presents a novel approach that uses the correlation between headlines and news content to identify fake news. Using deep neural networks, the headline and body of news articles are analyzed as two independent components, and their correlation is measured. We fine-tuned two BERT language models on the headline and body text as the two constituent parts of the news to determine whether there is a correlation between the news headline and body text. The results showed that this approach to fake news can enhance model accuracy compared to similar models.

Keywords

• Fake news
• Cybersecurity
• Stance analysis
• Deep neural networks
• Cyber warfare

Main Subjects

• IT and cyber warfare (encryption and data security, penetration and disruption in information networks, radars, dealing with hackers, etc.)