Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LOSS-GAT: Label Propagation and One-Class Semi-Supervised Graph Attention Network for Fake News Detection

Batool Lakzaei, Mostafa Haghir Chehreghani, Alireza Bagheri

TL;DR

LOSS-GAT addresses fake news detection under limited labeled data by integrating One-Class Learning with a two-step label propagation framework and Graph Attention Networks. It builds a content-based similarity graph, uses Katz propagation to bootstrap pseudo-labels, then refines labels with a GATv2 classifier, further enhanced by Adamic-Adar based structural augmentation and randomized neighborhood aggregation. Across five real-world datasets, LOSS-GAT consistently outperforms both OCL and binary-labeled baselines, with improvements of up to 3 percentage points in macro-F1 and accuracy, demonstrating strong performance with minimal labeling. The approach highlights the practical value of combining graph-based propagation, structural cues, and stochastic aggregation for robust fake news detection in diverse languages and data regimes.

Abstract

In the era of widespread social networks, the rapid dissemination of fake news has emerged as a significant threat, inflicting detrimental consequences across various dimensions of people's lives. Machine learning and deep learning approaches have been extensively employed for identifying fake news. However, a significant challenge in identifying fake news is the limited availability of labeled news datasets. Therefore, the One-Class Learning (OCL) approach, utilizing only a small set of labeled data from the interest class, can be a suitable approach to address this challenge. On the other hand, representing data as a graph enables access to diverse content and structural information, and label propagation methods on graphs can be effective in predicting node labels. In this paper, we adopt a graph-based model for data representation and introduce a semi-supervised and one-class approach for fake news detection, called LOSS-GAT. Initially, we employ a two-step label propagation algorithm, utilizing Graph Neural Networks (GNNs) as an initial classifier to categorize news into two groups: interest (fake) and non-interest (real). Subsequently, we enhance the graph structure using structural augmentation techniques. Ultimately, we predict the final labels for all unlabeled data using a GNN that induces randomness within the local neighborhood of nodes through the aggregation function. We evaluate our proposed method on five common datasets and compare the results against a set of baseline models, including both OCL and binary labeled models. The results demonstrate that LOSS-GAT achieves a notable improvement, surpassing 10%, with the advantage of utilizing only a limited set of labeled fake news. Noteworthy, LOSS-GAT even outperforms binary labeled models.

LOSS-GAT: Label Propagation and One-Class Semi-Supervised Graph Attention Network for Fake News Detection

TL;DR

LOSS-GAT addresses fake news detection under limited labeled data by integrating One-Class Learning with a two-step label propagation framework and Graph Attention Networks. It builds a content-based similarity graph, uses Katz propagation to bootstrap pseudo-labels, then refines labels with a GATv2 classifier, further enhanced by Adamic-Adar based structural augmentation and randomized neighborhood aggregation. Across five real-world datasets, LOSS-GAT consistently outperforms both OCL and binary-labeled baselines, with improvements of up to 3 percentage points in macro-F1 and accuracy, demonstrating strong performance with minimal labeling. The approach highlights the practical value of combining graph-based propagation, structural cues, and stochastic aggregation for robust fake news detection in diverse languages and data regimes.

Abstract

In the era of widespread social networks, the rapid dissemination of fake news has emerged as a significant threat, inflicting detrimental consequences across various dimensions of people's lives. Machine learning and deep learning approaches have been extensively employed for identifying fake news. However, a significant challenge in identifying fake news is the limited availability of labeled news datasets. Therefore, the One-Class Learning (OCL) approach, utilizing only a small set of labeled data from the interest class, can be a suitable approach to address this challenge. On the other hand, representing data as a graph enables access to diverse content and structural information, and label propagation methods on graphs can be effective in predicting node labels. In this paper, we adopt a graph-based model for data representation and introduce a semi-supervised and one-class approach for fake news detection, called LOSS-GAT. Initially, we employ a two-step label propagation algorithm, utilizing Graph Neural Networks (GNNs) as an initial classifier to categorize news into two groups: interest (fake) and non-interest (real). Subsequently, we enhance the graph structure using structural augmentation techniques. Ultimately, we predict the final labels for all unlabeled data using a GNN that induces randomness within the local neighborhood of nodes through the aggregation function. We evaluate our proposed method on five common datasets and compare the results against a set of baseline models, including both OCL and binary labeled models. The results demonstrate that LOSS-GAT achieves a notable improvement, surpassing 10%, with the advantage of utilizing only a limited set of labeled fake news. Noteworthy, LOSS-GAT even outperforms binary labeled models.
Paper Structure (24 sections, 13 equations, 4 figures, 10 tables, 3 algorithms)

This paper contains 24 sections, 13 equations, 4 figures, 10 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Problem Formulation
  4. Proposed Model
  5. Feature extraction
  6. Graph construction
  7. Two-step label propagation
  8. Step 1: label propagation using Katz index
  9. Step 2: Label propagation using an initial classifier
  10. Classification
  11. Structural augmentation
  12. Inducing randomness in local neighborhood aggregation
  13. Prediction
  14. Experiments
  15. Datasets
  16. ...and 9 more sections

Figures (4)

  • Figure 1: LOSS-GAT: our proposed approach for fake news detection
  • Figure 2: Structural similarity: the red and green nodes are not situated within the $2$-order neighborhood of each other. Therefore, if a two-layer GNN is employed, these nodes will not contribute to their embedding vectors. However, based on the graph structure, these nodes exhibit a high degree of structural similarity: in an isomorphism from the graph to itself (automorphism), these two nodes can be mapped to each other.
  • Figure 3: Macro-F1 (a) and Accuracy (b) comparision of two-step label propagation and one-step label propagation. 0.1, 0.2, and 0.3 represent the amount of labeled data used to train the algorithms.
  • Figure 4: Macro-F1 and accuracy comparision, for different amounts of labeled fake news. 0.1, 0.2, and 0.3 represent the ratio of labeled data used to train the algorithms. (a): FNN, (b): Fake.Br, (c): Fact-checked News, (d): Fake News Detection, (e): Fake News Data.