Bootstrapping Multi-view Representations for Fake News Detection

TL;DR

This paper introduces a novel multi-view representation approach for fake news detection that leverages text and image features, using a bootstrap method to improve detection accuracy and understand cross-modal consistency.

Contribution

It proposes a new Bootstrapping Multi-view Representations (BMR) scheme with improved multi-gate mixture-of-experts networks for effective feature fusion and fake news detection.

Findings

BMR outperforms state-of-the-art methods on benchmark datasets.

Multi-view representations improve detection accuracy.

Cross-modal consistency prediction enhances fake news identification.

Abstract

Previous researches on multimedia fake news detection include a series of complex feature extraction and fusion networks to gather useful information from the news. However, how cross-modal consistency relates to the fidelity of news and how features from different modalities affect the decision-making are still open questions. This paper presents a novel scheme of Bootstrapping Multi-view Representations (BMR) for fake news detection. Given a multi-modal news, we extract representations respectively from the views of the text, the image pattern and the image semantics. Improved Multi-gate Mixture-of-Expert networks (iMMoE) are proposed for feature refinement and fusion. Representations from each view are separately used to coarsely predict the fidelity of the whole news, and the multimodal representations are able to predict the cross-modal consistency. With the prediction scores, we reweigh each view of the representations and bootstrap them for fake news detection. Extensive experiments conducted on typical fake news detection datasets prove that the proposed BMR outperforms state-of-the-art schemes.