CAFE: Fact Checking in Knowledge Graphs using Neighborhood-Aware Features

Tracking #: 2449-3663

This paper is currently under review
Agustin Borrego
Daniel Ayala
inma hernandez
Carlos R. Rivero
David Ruiz

Responsible editor: 
Philippe Cudre-Mauroux

Submission type: 
Full Paper
Knowledge Graphs (KGs) currently contain a vast amount of structured information in the form of entities and relations. Because KGs are often constructed automatically by means of information extraction processes, they may miss information that was either not present in the original source or not successfully extracted. As a result, KGs might potentially lack useful and valuable information. Current approaches that aim to complete missing information in KGs either have a dependence on embedded representations, which hinders their scalability and applicability to different KGs; or are based on long random paths that may not cover relevant information by mere chance, since exhaustively analyzing all possible paths of a large length between entities is very time-consuming. In this paper, we present an approach to completing KGs based on evaluating candidate triples using a novel set of features, which exploits the highly relational nature of KGs by analyzing the entities and relations surrounding any given pair of entities. Our results show that our proposal is able to identify correct triples with a higher effectiveness than other state-of-the-art approaches (up to 60% higher precision or 20% higher recall in some datasets).
Full PDF Version: 
Under Review