cognonto / kbpedia Goto Github PK

KBPedia Knowledge Graph & Knowledge Ontology (KKO)

knowledge-graph semantic-web rdf owl ontology

kbpedia's Introduction

Introduction

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.

Knowledge Graph

Repository Content

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.n3
wikidata.n3
schema.org.n3
dbpedia-ontology.n3
geonames.n3
opencyc.n3
umbel.n3
bibo.n3
cc.n3
dc.n3
doap.n3
event.n3
foaf.n3
frbr.n3
geo.n3
mo.n3
oo.n3
org.n3
po.n3
rss.n3
same-as.n3
sioc.n3
time.n3
transit.n3

Finally, under the typologies folder we added one file per typology (see below) with all the relationships it contains.

Typologies

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.

Here are the 30 or so core typologies organized in the KKO graph, with some upper typologies that cluster them.

Explore

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:

KKO

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.

kbpedia's People

Contributors

Stargazers

Watchers

kbpedia's Issues

404 links on http://kbpedia.org/resources/downloads/

check http://kbpedia.org/resources/downloads/
I have tried first 5 links and got 404:-(

Issues using KBpedia demo page

Hi, I am trying to figure out how to use KBpedia

I went through the demo page: https://kbpedia.org/demo

It says:

KBpedia exploits large-scale knowledge bases and semantic technologies for effective machine learning and data interoperability. Here are some example ways that KBpedia may power knowledge management-oriented Web services or APIs. Try it for yourself:
Analyze Web Page
Enter a full URL below to see some example uses of KBpedia:

It is unclear as what kind of URL should be put there, and what KBpedia will produce from this URL. Will it try to match it to your KKO and returns metadata about this concept? Will it try to parse the content of the provided URL for concepts (json-ld, RDFa, etc)?

I tried to put URLs of concepts from DBpedia and Wikidata (e.g. https://www.wikidata.org/wiki/Q90), but the demo seems to fail due to loading mixed content (loading HTTP webpages from HTTPS), even if I am providing a HTTPS URL to the Wikidata page

Analyzing a web page leads to this page:

I tried on Google Chrome and Firefox (on Ubuntu 18.04) and get the same mixed content error. I also tried to access KBpedia through HTTP: http://kbpedia.org/demo which should prevent mixed content errors, but the website is automatically redirecting to HTTPS

Is there anything I can do to make it work on my side? And what kind of URL can be provided to the KBpedia demo endpoint?

Syntactically wrong altLabels in knowledge graph?

I noticed that in https://github.com/Cognonto/kbpedia/blob/master/versions/1.60/kbpedia_reference_concepts_linkage_inferrence_extended.zip there are some altLabels stored in one literal, each seperated by ||, e.g.:

skos:altLabel "Your Paintings collection identifier||BBC Your Paintings collection identifier||Art UK collection identifier"@en ;

Whereas others are written in statements with multiple objects (literals):

skos:altLabel "apartments"@en , "flat"@en , "flats"@en ;

As far as I know, only the second statement is valid N3, right? At least Apache Jena, into which I am reading the file, will only parse the second statement correctly (as 3 independent triples) but not the first.

This may also be some issue with Apache Jena... if so I am sorry to have bothered you.

IOT Ontology integration

Hi. Do you have ontologies for IOT domain (sensor devices) such as SenML, SSN, or SAREF ontology integrated in KBpedia? Do you plan to add that use case as well?

Question about actual contents

I've downloaded the snapshot and loaded the contents of kbpedia_reference_concepts_linkage_inferrence_extended.zip in a triplestore.

The following query returns 207984 as number of entities, 626131 as the number of objects, and 1503796 as number of triples.

SELECT (count(distinct ?s) as ?nent) (count(distinct ?o) as ?nobj) (count(*) as ?nt)
FROM <kbpedia>
WHERE {
  ?s ?p ?o.
}

question: is this correct? What is the reference in the readme to 30M entities?

Thanks

Missing entities and linkages

On the KBpedia home page it is stated that-

"KBpedia, written primarily in OWL 2, includes more than 58,000 reference concepts, mapped linkages to about 40 million entities (most from Wikidata), and 5,000 relations and properties."

I am familiar with the resources made available on your Github page. However, I am unable to find the linkages to the stated "40 million" entities from the other ontologies. The linkages that are provided on Github seem to cover only a very minor portion of this. For example, the wikidata linkages file contains approximately only 46000 relations.

I would appreciate it if you could let me know if there's something I am missing, or the entire linkage data to the 40 million entities is indeed not made available for public use.

Copy edit contributions

Mike and Fred, congrats on the milestone. I started using the web interface and my second search on "climate" returned an entry with climate misspelled ("climage"). Is there some way I can submit suggestions for content edits?
I am looking forward to looking further into your work here.

Local imports

You have an imports to a local file:

<http://kbpedia.org/kbpedia/rc#> a owl:Ontology ;
	owl:versionIRI :v250 ;
	owl:imports <https://www.w3.org/2009/08/skos-reference/skos-owl1-dl.rdf> , <file:///kbpedia/2.50/owl/kko.n3> ;
	dc:date "2020-02-24T20:00:00Z"@en ;