TEGRA: Text Encoding With Graph and Retrieval Augmentation for Misinformation Detection

TL;DR

This paper introduces TEGRA, a novel framework that combines graph-based document encoding with retrieval-augmented knowledge integration to improve misinformation detection accuracy.

Contribution

The paper presents TEGRA, a new hybrid approach that encodes documents as graphs and incorporates external knowledge, outperforming language models alone in misinformation detection.

Findings

TEGRA improves classification accuracy over baseline models.

Graph-based encoding captures structured information effectively.

Knowledge integration enhances detection performance in domain-specific cases.

Abstract

Misinformation detection is a critical task that can benefit significantly from the integration of external knowledge, much like manual fact-checking. In this work, we propose a novel method for representing textual documents that facilitates the incorporation of information from a knowledge base. Our approach, Text Encoding with Graph (TEG), processes documents by extracting structured information in the form of a graph and encoding both the text and the graph for classification purposes. Through extensive experiments, we demonstrate that this hybrid representation enhances misinformation detection performance compared to using language models alone. Furthermore, we introduce TEGRA, an extension of our framework that integrates domain-specific knowledge, further enhancing classification accuracy in most cases.