End-to-end Deep Networks with Hierarchical Attention and Capsule Capabilities for Misinformation Detection on Microblogging Platforms

Widespread dissemination of misinformation over online social media has brought about negative consequences that disrupt lives on so many levels, from personal to an entire society. Inspired by ongoing occurrences, this study aims to explore an efficient method that can detect misinformation accurately since it is a key to help prevent or at least mitigate the chaos that arises due to misleading or false claims. The paper proposes a deep neural architecture that leverages the capability of hierarchical attention networks together with capsule networks to learn effective representation for the misinformation detection task. Our finding suggests that the hierarchical structure of each event as well as capsule networks are contributing factors that lead to overall performance gain. Results from extensive experiments conducted on two real-world datasets indicate that the proposed approach can accurately detect events that carry misinformation, outweighing a range of competitive baselines.