KBpedia is a comprehensive knowledge structure for promoting data interoperability and knowledge-based artificial intelligence, or KBAI. The KBpedia knowledge structure combines seven (7) public knowledge bases — Wikipedia, Wikidata, schema.org, DBpedia, GeoNames, OpenCyc, and standard UNSPSC products and services — into an integrated whole. KBpedia’s upper structure, or knowledge graph, is the KBpedia Knowledge Ontology. We base KKO on the universal categories and knowledge representation theories of the great 19th century American logician, polymath and scientist, Charles Sanders Peirce.
KBpedia, written primarily in OWL 2, includes more than 58,000 reference concepts, mapped linkages to about 30 million entities (mostly from Wikidata), and 5,000 relations and properties, all organized according to about 75 modular typologies that can be readily substituted or expanded. We subject items added to KBpedia to a rigorous suite of logic and consistency tests — and best practices — before acceptance. The result is a flexible and computable knowledge graph that can be sliced-and-diced and configured for all sorts of machine learning tasks, including supervised, unsupervised and deep learning.
KBpedia, KKO and its mapped information can drive multiple use cases include providing a computable framework over Wikipedia and Wikidata, creating word embedding models, fine-grained entity recognition and tagging, relation and sentiment extractors, and categorization. Knowledge-based AI models may be set up and refined with unprecedented speed and accuracy by leveraging the integrated KBpedia structure. KBpedia is also a powerful nucleus for setting up your own coherent domain ontology or knowledge graph.
To learn more, try out the KBpedia demo or explore the KBpedia knowledge graph.
The open source KBpedia, its voluminous mappings, and its typologies are available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.
Each KBpedia version exists in the versions folder. Each sub-folder is a version folder such as 1.60. In each of the version folder, you will have the following files:
| file name | description |
|---|---|
kbpedia_reference_concepts.zip | This is the code KBpedia reference concepts structure with all the 54k concepts |
kbpedia_reference_concepts_linkage.zip | This is the same structure as above where we added all the linkages to other ontologies (see below) |
kbpedia_reference_concepts_linkage_inferrence_extended.zip | This is the same structure that includes the linkages, but we added all inferred relationships between the concepts and their links to other ontologies |
Then we added a sub-folder called linkages where each link to other ontologies concepts have been listed. There is one file per linked ontology. The ontologies currently linked to KBpedia are:
wikipedia.n3wikidata.n3schema.org.n3dbpedia-ontology.n3geonames.n3opencyc.n3umbel.n3bibo.n3cc.n3dc.n3doap.n3event.n3foaf.n3frbr.n3geo.n3mo.n3oo.n3org.n3po.n3rss.n3same-as.n3sioc.n3time.n3transit.n3
Finally, under the typologies folder we added one file per typology (see below) with all the relationships it contains.
The KKO knowledge graph has a relatively thin upper layer, informed by the trichotomous logic and categories of Charles Sanders Peirce, that sits astride (mostly) typologies of entity classes organized according to shared attributes.
Most of the 30 or so core typologies in KBpedia do not overlap with one another, what is known as disjoint. Disjointness enables powerful reasoning and subset selection (filtering) to be performed on the KKO graph. There are upper typologies useful for further organizing the core ontologies, plus providing homes for shared concepts. Living Things, for example, can capture concepts shared by all plants and animals, by all life, which then enables better segregation of those life forms. These natural segregations are applied across the KKO structure.
To explore KBpedia, simply use the KBpedia Knowledge Graph explorer. Possible matching concepts are presented as you type. Once you enter the knowledge graph, you can explore and navigate in many different ways. Alternatively, try one of these KBpedia concepts as a way to get started:
Below is a complete representation of the KBpedia Knowledge Ontology (KKO), the upper portions of the knowledge graph. Note that the specific entries you may search and find within the knowledge graph reside under the Generals branch of the KKO.
