Neural Approaches to Fake News Detection¶
Methods using neural networks (RNNs, CNNs, transformers, graph neural networks) to automatically learn representations of claims, articles, or propagation patterns for credibility assessment and misinformation detection. These approaches eliminate the need for hand-crafted linguistic features and enable end-to-end learning from data.
Key papers¶
- DeClareE: Debunking Fake News and False Claims using Evidence-Aware Deep Learning — bidirectional LSTM with attention mechanisms for evidence-aware credibility assessment; demonstrates end-to-end neural learning outperforms feature engineering.
- CSI: A Hybrid Deep Model for Fake News Detection — hybrid deep model combining text LSTM, temporal engagement, and user group behavior for fake news detection.
- Rumor Detection on Twitter with Tree-structured Recursive Neural Networks — recursive neural networks on propagation trees for rumor verification.