HybridFC: A Hybrid Fact-Checking Approach for Knowledge Graphs

TL;DR

HybridFC is a novel ensemble-based fact-checking method for knowledge graphs that combines multiple approaches to significantly improve prediction accuracy over existing methods.

Contribution

It introduces a hybrid ensemble approach that leverages diverse fact-checking techniques to enhance accuracy in knowledge graph assertions.

Findings

Outperforms state-of-the-art by 0.14 to 0.27 AUC on FactBench

Effectively combines multiple fact-checking categories

Open-source implementation available

Abstract

We consider fact-checking approaches that aim to predict the veracity of assertions in knowledge graphs. Five main categories of fact-checking approaches for knowledge graphs have been proposed in the recent literature, of which each is subject to partially overlapping limitations. In particular, current text-based approaches are limited by manual feature engineering. Path-based and rule-based approaches are limited by their exclusive use of knowledge graphs as background knowledge, and embedding-based approaches suffer from low accuracy scores on current fact-checking tasks. We propose a hybrid approach -- dubbed HybridFC -- that exploits the diversity of existing categories of fact-checking approaches within an ensemble learning setting to achieve a significantly better prediction performance. In particular, our approach outperforms the state of the art by 0.14 to 0.27 in terms of Area Under the Receiver Operating Characteristic curve on the FactBench dataset. Our code is open-source and can be found at https://github.com/dice-group/HybridFC.