Memory-Guided Multi-View Multi-Domain Fake News Detection

The wide spread of fake news is increasingly threatening both individuals and society. Great efforts have been made for automatic fake news detection on a single ___domain (e.g., politics). However, correlations exist commonly across multiple news domains, and thus it is promising to simultaneously detect fake news of multiple domains. Based on our analysis, we pose two challenges in multi-___domain fake news detection: 1) ___domain shift, caused by the discrepancy among domains in terms of words, emotions, styles, etc. 2) ___domain labeling incompleteness, stemming from the real-world categorization that only outputs one single ___domain label, regardless of topic diversity of a news piece. In this paper, we propose a Memory-guided Multi-view Multi-___domain Fake News Detection Framework (M^3FEND) to address these two challenges. We model news pieces from a multi-view perspective, including semantics, emotion, and style. Specifically, we propose a Domain Memory Bank to enrich ___domain information which could discover potential ___domain labels based on seen news pieces and model ___domain characteristics. Then, with enriched ___domain information as input, a Domain Adapter could adaptively aggregate discriminative information from multiple views for news in various domains. Extensive offline experiments on English and Chinese datasets demonstrate the effectiveness of M^3FEND, and online tests verify its superiority in practice. Our code is available at https://github.com/ICTMCG/M3FEND.