Fact validation is needed to ensure that the triples within a knowledge graph are true. To this end, several algorithms have been developed that try to validate a given triple (i.e., a single fact) with respect to some reference knowledge.
The goal of this project group is the usage of a set of open source approaches and combine them to a single fact validation approach. The combination should be done using a meta classifier that learns which of the systems is more reliable for a particular fact. More information (including papers and links to projects) can be found in the slides.
L.079.07005 Project Group: Fact Validation Ensemble Learning (in English)