Contrastive Token-level Explanations for Graph-based Rumour Detection

TL;DR

This paper introduces a new explainability framework called CT-LRP for GNN-based rumour detection, providing detailed token-level explanations to improve interpretability and trustworthiness of models detecting harmful social media rumours.

Contribution

The paper presents CT-LRP, a novel method that enhances graph explainability by offering token-level relevance explanations for GNN-based rumour detection models.

Findings

CT-LRP produces high-fidelity explanations across multiple GNN models.

It improves interpretability of rumour detection systems.

Demonstrates effectiveness on three public datasets.

Abstract

The widespread use of social media has accelerated the dissemination of information, but it has also facilitated the spread of harmful rumours, which can disrupt economies, influence political outcomes, and exacerbate public health crises, such as the COVID-19 pandemic. While Graph Neural Network (GNN)-based approaches have shown significant promise in automated rumour detection, they often lack transparency, making their predictions difficult to interpret. Existing graph explainability techniques fall short in addressing the unique challenges posed by the dependencies among feature dimensions in high-dimensional text embeddings used in GNN-based models. In this paper, we introduce Contrastive Token Layerwise Relevance Propagation (CT-LRP), a novel framework designed to enhance the explainability of GNN-based rumour detection. CT-LRP extends current graph explainability methods by providing token-level explanations that offer greater granularity and interpretability. We evaluate the effectiveness of CT-LRP across multiple GNN models trained on three publicly available rumour detection datasets, demonstrating that it consistently produces high-fidelity, meaningful explanations, paving the way for more robust and trustworthy rumour detection systems.