1 / 34

UMBEL: A Lightweight Subject Reference Structure for the Web

UMBEL is a lightweight reference structure for placing Web content and data in context with other data, comprising about 20,000 subject concepts and relationships. It is free, open source, and aims to provide coherence and context to web content. UMBEL serves as middleware, providing a standard set of subjects for integrating content and datasets. It acts as a roadmap and backbone in inter-relating content by offering a contextual graph for different domains. The ontology introduces new classes and properties, enabling aggregation and contextualization of content. UMBEL focuses on tying in web datasets, named entities, and leveraging existing ontologies like Cyc.

jbrewster
Download Presentation

UMBEL: A Lightweight Subject Reference Structure for the Web

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. UMBEL: A Lightweight Subject Reference Structure for the Web Upper Mapping and Binding Exchange Layer Michael K. Bergman Editor, UMBEL CEO, Zitgist LLC Ontolog Forum September 25, 2008

  2. UMBEL is . . . . . . a lightweight reference structure for placing Web content and data in context with other data. . . . comprised of about 20,000 subject concepts and their relationships — with one another and with external vocabularies and named entities. . . . free, non-commercial and open source under a Creative Commons Share Alike ver. 3.0 license.

  3. Early Objectives • Provide subject matter “glue” to the semantic Web  what is content about ? • Build from, leverage the success of user-generated content (e.g., Wikipedia) • Keep flat; no structure, no concept relationships But: • Only first objective withstood diligence

  4. Later Observations • Coherence essential for use as a reference framework • Coherence: • Internally, logically consistent • Adequate, testable representation of the world as we know it • Current UGC (user-generated content) lacks structural coherence • Flat was not desirable; structure brings other benefits (search, discovery, inference) Thus: • What should be the basis of this coherent structure?

  5. Candidates Investigated • Topic Maps • SUMO • BFO • PROTON • Library classification systems • Wikipedia, DMOZ, etc. • WordNet • Gellish • DOLCE • Eventual choice: Cyc

  6. Derived from OpenCyc . . . 90% reduction

  7. UMBEL is a Meta Layer for Metadata • Provides a standard set of subjects for tieing in content and datasets

  8. UMBEL is Middleware • UMBEL ontology is based on RDF • RDF Schema vocabulary of SKOS (Simple Knowledge Organization System) • OWL Full to aid class, domain and range relationships with external ontologies

  9. UMBEL is a Roadmap • A set of roadsigns to help find related content • A contextual way to get from one concept space to another • More detailed maps used for specific domains

  10. UMBEL is a Backbone • The backbone to UMBEL is the relationships amongst its subject concepts • The backbone provides the contextual graph for inter-relating content • The backbone is a reference structure for more specific domains or ontologies • The backbone is contextual, and relatively stable and constant

  11. UMBEL is an Ontology • The UMBEL ontology is based on RDFS, SKOS and OWL-Full 1 • The UMBEL ontology is a binding vocabulary and relationships for tieing Web datasets together • The UMBEL ontology introduces: • 3 new classes • 9 new properties (predicates) 1RDFS – Resource Description Framework Schema; SKOS – Simple Knowledge Organization System; OWL – Web Ontology Language

  12. Subject Concepts • 20,000 defined reference points in information space • Means to assert what a given chunk of content is about • Enable similar content to be aggregated • Place content in context with other content • Aggregation points for tying in named entities • Totally derived and a subset of OpenCyc, the open source version of the Cyc knowledge base http://www.opencyc.comhttp://www.cycorp.com

  13. Named Entities • The places, events, people, objects, and specific things of the real world • Literally millions of notable instances • Each belongs to one or more subject concept(s) • Currently, the predominate basis for Linked Data • Important sources include Wikipedia (Yago) and other public sources • Can be readily mixed-and-matched with private entities

  14. ‘Exploding the Domain’ • Each subject concept is an individual of the skos:Concept class, which are themselves OWL classes • Since subject concepts are classes, can relate to external ontology classes using rdfs:subClassOf and owl:equivalentClass • Under right circumstances, can also inherit properties • Through inferencing, silo data now can inherit open world understanding, therefore: exploding the domain

  15. A Deep Dive into the Backbone . . . . . . all 20,000 nodes

  16. A Deep Dive into the Backbone . . . . . . top 750 nodes

  17. A Deep Dive into the Backbone . . . . . . top 350 nodes

  18. A Deep Dive into the Backbone . . . . . . zeroing in on ‘automobiles’

  19. A Deep Dive into the Backbone . . . . . . zeroing in on ‘Saabs’

  20. Re-use of Existing Ontologies • Properties of the Dublin Core ontology: • dcterms:language • Properties of the SKOS ontology: • skos:prefLabel • skos:altLabel • skos:definition • skos:broader • skos:narrower • skos:related • skos:broaderTransitive • skos:narrowerTransitive • seven other minor ones • Properties of the RDFS ontology: • rdfs:subClassOf • Properties of the OWL ontology: • owl:equivalentClass

  21. New UMBEL Classes • umbel:SubjectConcept – concrete, subject-related, and non-abstract “things” with members or instances that are part of the real world • umbel:AbstractConcept – abstract or ephemeral notions often used solely for organization or categorization; used only internally for graph interconnectivity • umbel:Semset – sematically close terms or phrases synonymous or nearly so with subject concepts or named entities; may embrace jargon and contemporary tags

  22. New UMBEL Predicates • umbel:isAligned – associative class to subject concept link • umbel:isAbout – associative named entity to subject concept link • umbel:isLike – associative link between individuals where similar, but exact sameness uncertain • umbel:hasSemset – links a subject concept to its semset • umbel:linksConcept – inverse property of isAligned • umbel:linksEntity – inverse property of isAbout • umbel:superClassOf – inverse to rdfs:subClassOf • umbel:withAlignment – degree of alignment estimate for isAligned • umbel:withLikelihood – degree of alignment estimate for isAbout

  23. Sample Specification

  24. FOAF(Friend-of-a-Friend Ontology) FOAF Yandex(Friend-of-a-Friend Yandex extension Ontology) RSS 1.0(Really Simple Syndicate Ontology) SIOC(Socially Interlinked Online Communities Ontology) DOAP(Description of a Project Ontology) BIBO (Bibliographic Ontology) GEO WGS84(Geo wgs84 Ontology) GEONAMES(Geonames Ontology) CC(Creative Commons Ontology) EVENT(Event Ontology) PO(Programme Ontology) MO (Music Ontology) FRBR(Functional Requirements for Bibliographic Records) Linkage to External Ontologies

  25. Find Subject Concepts Subject Concept Report Subject Concept Detailed Report List Sub-Concepts & Sub-Classes List Super-Concepts & Super-Classes List Equivalent External Classes Discover Subject Concepts Verify Sub-Class Relationship Verify Super-Class Relationship Verify Equivalent Class Relationship Subject Concepts Explorer DBpedia Linkage Yago Ontology UMBEL’s Web Services

  26. Demos, Docs & APIs

  27. Subject Concept Explorer

  28. Conclusion and Further Information

  29. Why Important to the Semantic Web? • Context, Context, CONTEXT • For Linked Data • Context ! • For Web Search: • Infinite facets • Disambiguation • Cleaner navigation • For Legacy Data Stovepipes: • Making the connections • Closed World schemas become Open World information • Context !

  30. Ongoing Efforts • SPARQL endpoints and APIs • Now creating domain-specific ontologies  • Analogous to OBO (Open Biomedical Ontologies), except: • Central role of subject concept reference ontology (UMBEL) • Exposure as linked data • Some may be proprietary, enterprise-specific • Following same design as UMBEL: • Lightweight • Use of UMBEL vocabulary (predicates) • Structural consistency with UMBEL, Cyc • Use Cyc, other tools for consistency and coherence checks for new relationships

  31. For Further Information • Main Web site: http://www.umbel.org • UMBEL Web services: http://umbel.zitgist.com • Download ontology: http://umbel.org/ontology/umbel.n3 • Documentation: http://www.umbel.org/documentation.html • Wiki: http://wiki.umbel.org/ • Google group: http://groups.google.com/group/umbel-ontology • Slide presentations: http://www.slideshare.net/mkbergman

  32. Contributions and Thanks

  33. Contact Information Mike Bergman, editor Zitgist LLC mike@zitgist.com Fred Giasson, editor Zitgist LLC fred@zitgist.com

  34. Thanks to for hosting us !

More Related