Automatic verification of the support for statements in referenced documents

What is it about?

Knowledge graphs represent information as triples of the form (subject,predicate,object) and are widely used in many modern applications. Ensuring the verifiability of the data in a knowledge graph through documented provenance is a task that is crucial to guarantee its quality and promote trustworthiness. To work at scale such provenance verification task needs to be actively supported by automated and semi-automated tools to help data curators and editors cope with the sheer volume of information. However, as of this paper's publication, no such tools had been deployed to large knowledge graphs and only a very small family of approaches tackle this task following a rigorous scientific methodology. This paper presents a computational process through which the support for claims of knowledge graph triples can be automatically checked against the documents provided as reference.

Featured Image

Photo by Hartono Creative Studio on Unsplash

Photo by Hartono Creative Studio on Unsplash

Why is it important?

Knowledge graphs underpin the information used by many services, such as Wikipedia infoboxes, search engines, and some voice-activated assistants, amongst others. Ensuring the quality of this information is not easily done manually. The process presented in this paper provides a framework in which automatic verification of knowledge graph triples is possible.