A Decision-Based Heterogenous Graph Attention Network for Multi-Class Fake News Detection

2025-01-06

AI Summary

Existing fake news detection models are predominantly limited to binary classification, rendering them inadequate for scenarios involving information scarcity and fine-grained, multi-class identification. To address this, we propose a semi-supervised heterogeneous graph learning framework tailored for six-way fake news classification. Our approach constructs a heterogeneous graph comprising multiple node types—including news articles, entities, and claims—and introduces a decision-network-driven dynamic neighborhood selection mechanism that operates layer-wise, overcoming the limitations of conventional GNNs with static neighborhoods and binary decision boundaries. The framework integrates learnable attention mechanisms, dual-network collaboration—where the decision network guides the representation network—and semi-supervised graph learning. Evaluated on the LIAR benchmark, our model achieves approximately 4% higher accuracy than state-of-the-art methods and demonstrates strong robustness under extreme label scarcity.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

A promising tool for addressing fake news detection is Graph Neural Networks (GNNs). However, most existing GNN-based methods rely on binary classification, categorizing news as either real or fake. Additionally, traditional GNN models use a static neighborhood for each node, making them susceptible to issues like over-squashing. In this paper, we introduce a novel model named Decision-based Heterogeneous Graph Attention Network (DHGAT) for fake news detection in a semi-supervised setting. DHGAT effectively addresses the limitations of traditional GNNs by dynamically optimizing and selecting the neighborhood type for each node in every layer. It represents news data as a heterogeneous graph where nodes (news items) are connected by various types of edges. The architecture of DHGAT consists of a decision network that determines the optimal neighborhood type and a representation network that updates node embeddings based on this selection. As a result, each node learns an optimal and task-specific computational graph, enhancing both the accuracy and efficiency of the fake news detection process. We evaluate DHGAT on the LIAR dataset, a large and challenging dataset for multi-class fake news detection, which includes news items categorized into six classes. Our results demonstrate that DHGAT outperforms existing methods, improving accuracy by approximately 4% and showing robustness with limited labeled data.

Problem

• Fake News Detection
• Model Limitations
• Multi-class Classification

Innovation

• Dynamic Handling GAT
• Graph Neural Networks
• Fake News Detection