Deep learning¶

Deep learning approaches using neural networks for fake news and misinformation detection.

Key papers and articles¶

The Evolution of Distributed Systems for Graph Neural Networks and their Origin in Graph Processing and Deep Learning: A Survey — Survey of distributed systems for training graph neural networks at scale; covers partitioning, sampling, communication, synchronization strategies applicable to deep learning on graphs.
We used Neural Networks to Detect Clickbaits: You won't believe what happened Next! — Bidirectional LSTM with word and character embeddings for clickbait detection; demonstrates effectiveness of combining embedding types for headline classification without hand-crafted features.
DeepCas: an End-to-end Predictor of Information Cascades — End-to-end deep learning for information cascade prediction; represents cascade graphs as random walk paths processed through bidirectional GRU with attention mechanisms.
Rana et al. (2022) — Deepfake Detection: A Systematic Literature Review: Comprehensive SLR showing deep learning methods dominate deepfake detection (77% of 112 studies); CNNs most prevalent architecture; deep learning achieves 89.7% mean accuracy vs. 85% for traditional ML
Ruchansky, Seo, & Liu (2017) — CSI: A Hybrid Deep Model for Fake News Detection: Uses LSTM for temporal pattern capture and fully connected layers for user behavior modeling.
A Benchmark Study of Machine Learning Models for Online Fake News Detection — Evaluates 6 deep learning models (CNN, LSTM, Bi-LSTM, C-LSTM, HAN, Convolutional HAN) on fake news detection; finds deep learning outperforms traditional ML on large datasets but shows high overfitting on small datasets compared to pre-trained transformer models.
FakeBERT: Fake News Detection in Social Media with a BERT-based Deep Learning Approach — Combines BERT with parallel 1D CNNs using varying kernel sizes, achieving 98.90% accuracy and demonstrating the effectiveness of bidirectional transformer embeddings for fake news classification.
A Heuristic-driven Uncertainty based Ensemble Framework for Fake News Detection in Tweets and News Articles — Ensemble of seven pre-trained transformer models for fake news detection with Statistical Feature Fusion Network; demonstrates effectiveness of combining multiple deep learning architectures and uncertainty quantification via Monte Carlo Dropout.