Fake News Detection using Deep Learning

Abstract

Social media plays a significant role in people’s daily lives. More people read news online than in conventional newspapers. The risk of spreading false information is rising as online news outlets start to grow and social media applications gain more and more user popularity. Society is seriously harmed by fake news. Multimedia news is becoming more commonplace alongside text-based news . To properly identify false news nowadays, several modalities, including pictures, audio, and video, must be taken into account. In this article, we present a thorough analysis of early, late, and hybrid fusion-based false news detection methods. We use two publicly available sources to show a hybrid CNN-RNN technique for false news identification. Along with that, we also try the Transformers-based approach to improving the results of the FAKES dataset. We describe the further improvement efforts which are data-specific and use more complex models than the baselines themselves.