Crowd-Sourced Knowledge Graph Extension: A Belief Revision Based Approach.

Abstract

Knowledge graphs are gaining popularity as key ingredients of many advanced applications. Some are created by experts covering specific fields. However, for many applications there is a need of having the common sense knowledge that is not domain specific, and, therefore, can be provided by nonexperts. In this paper we introduce a novel crowd-sourcing approach that allows the user to provide their update in a simplistic intuitive form without having the information about the knowledge already contained in the graph. The approach roots in belief revision theory and is capable of analyzing the user input, identifying the compliance with the existing structure and singling out new suggestions. When providing the update and upon submission the users obtain intuitive colorcoded feedback on their input w.r.t. to consistency and discrepancies with the existing knowledge. This feedback enables the educational aspect of the approach. The approach guarantees the consistency of the crowd-sourced knowledge when it is being integrated into the knowledge graph.