1 / 37

Grande Challenges for Ontology Design (or is it Vente?)

Grande Challenges for Ontology Design (or is it Vente?). Tom Gruber tomgruber.org. Questions for Today. Why make ontologies? What are they for? How can we guide ontology development? What are important applications for ontology development?. ontologies methods applications.

Download Presentation

Grande Challenges for Ontology Design (or is it Vente?)

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. Grande Challenges for Ontology Design (or is it Vente?) Tom Grubertomgruber.org

  2. Questions for Today • Why make ontologies? What are they for? • How can we guide ontology development? • What are important applications for ontology development? (c) 2007 Thomas Gruber ontologiesmethods applications

  3. Why make ontologies? • Truth? • Beauty? • Fame? • Fortune? • Why make software? (c) 2007 Thomas Gruber ontologiesmethods applications

  4. What makes a Good Ontology? • Truth? • Beauty? • Popularity? • Commercial Success? (c) 2007 Thomas Gruber ontologiesmethods applications

  5. What are Ontologies* For? • Enable data and information exchange • (for example, the Semantic Web) • Provide a conceptual and representational foundation on which to build systems. • Thus, Ontologies are Enabling Technology for Applications that Matter. *Which Ontologies? The ones we are talking about here  (c) 2007 Thomas Gruber ontologiesmethods applications

  6. What makes a Good Ontology. • Claim: Ontologies should be designed and evaluated with respect to how well they achieve their purposes. • Observation: Ontologies are agreements, made in a social context, to accomplish shared objectives. • Question: Which objectives? • Approach: Follow the process of collaborative engineering design. (c) 2007 Thomas Gruber ontologiesmethods applications

  7. Engineering Design Process • Requirements: Identify needs, use cases, constraints, desired functionality • Review existing solutions, technologies, tools, and operational environments • Design solution • Implement and Test solution • Deploy and Maintain solution (In modern practice, the process is iterative.) (c) 2007 Thomas Gruber ontologiesmethods applications

  8. Example: Tag Ontology • TagCommons group is working on agreements to enable the sharing of tagging data across the Web. • To guide the collaborative process, we are • Identifying use cases and functions • Derive ontology requirements • Survey existing ontologies and applications • Design/adapt/extend/minimize an ontology • Map it to formats, other ontologies, data sources, applications http://tagcommons.org (c) 2007 Thomas Gruber ontologiesmethods applications

  9. Bookmarking across sites Browsing others’ tags across sites Social search (collab filtering using tags) Multimedia cross reference resources Indexing documents and code in source repositories Tag Metasearch and Metamonitoring Social Science research Connecting the social and semantic webs Use Cases for Tag Ontology http://tagcommons.org/2007/02/28/functional-requirements-for-sharing-tag-data/ (c) 2007 Thomas Gruber ontologiesmethods applications

  10. Resulting Requirements • Core concepts: tagger, tagged, tag label, tag source/venue • Auxiliary metadata: dates, polarity, language • Identity and matching on core concepts • Namespaces for core concepts • Mappings among sources with different identity schemes • Bridges to other ontologies and standards (c) 2007 Thomas Gruber ontologiesmethods applications

  11. Tag Ontology Design Issues are framed and guided by use cases. • How to represent taggers (people)? • Don’t want to solve the whole problem of identity on the web – just matching of taggers • How to handle missing data and extensions? • Don’t need hard core nonmonotonic logics – just polymorphic relations with defaults (c) 2007 Thomas Gruber ontologiesmethods applications

  12. General Ontology Design Principles • clarity - context-independent, unambiguous, precise definitions • coherence – internally consistent • extendibility – anticipate the uses of the vocabulary, allow monotonic extension • minimal encoding bias – avoid representational choice for benefit of implementation • minimal ontological commitment – define only necessary terms, omit domain theory http://tomgruber.org/writing/onto-design.htm (c) 2007 Thomas Gruber ontologiesmethods applications

  13. How to stay grounded in applications? • Practical, application development stakeholders on the working group • They need an agreement on tag data to make their work feasible, not as the goal of their work. • Bridge to Wild Wild Web culture of microformats, REST APIs, etc. • Semantic Web GRRDL (c) 2007 Thomas Gruber ontologiesmethods applications

  14. Applying this to the Larger Ontology Community • What are the killer apps for ontologies? • What could be done with ontologies that couldn’t be done more cheaply, easily, or quickly without them? • What problems are important enough to do things “the right way”? (c) 2007 Thomas Gruber ontologiesmethods applications

  15. Semantic Web, meet the Social Web • Social Web: • architecture of participation – user data • emergent, bottom-up value creation • vital ecosystem of software and data reuse • Semantic Web: • architecture of computation – structured data • value from integration • ecosystem of service composition • The Killer Apps of Social + Semantic Web: • Collective Knowledge Systems (c) 2007 Thomas Gruber ontologiesmethods applications

  16. But what is “collective intelligence” in the social web sense? • intelligent collection? • collaborative bookmarking, searching • “database of intentions” • clicking, rating, tagging, buying • what we all know but hadn’t got around to saying in public before • blogs, wikis, discussion lists “database of intentions” – Tim O’Reilly (c) 2007 Thomas Gruber ontologiesmethods applications

  17. the wisdom of clouds? http://flickr.com/photos/tags/ (c) 2007 Thomas Gruber ontologiesmethods applications

  18. “Collective Knowledge” Systems • The capacity to provide useful information • based on human contributions • which gets better as more people participate. • typically • mix of structured, machine-readable data and unstructured data from human input http://tomgruber.org/writing/social-meets-semantic-web.htm (c) 2007 Thomas Gruber ontologiesmethods applications

  19. Collective Knowledge is Real • FAQ-o-Sphere - self service Q&A forums • Citizen Journalism – “We the Media” • Product reviews for gadgets and hotels • Collaborative filtering for books and music • Amateur Academia (c) 2007 Thomas Gruber ontologiesmethods applications

  20. What about Ontologies and the Semantic Web? (c) 2007 Thomas Gruber ontologiesmethods applications

  21. Roles for Technology • capturing everything • storing everything • distributing everything • enabling many-to-many communication • creating value from the data • Your ontology here  (c) 2007 Thomas Gruber ontologiesmethods applications

  22. Potential Roles for Semantic Net Technology: Two examples • Composing and integratinguser-contributed data across applications • example: tagging data • Creating aggregate valuefrom a mix of structured and unstructured data • example: blogging data (c) 2007 Thomas Gruber ontologiesmethods applications

  23. Role 2: Creating aggregate value from structured data • Problem: In a collective knowledge system, the value of the aggregate content must be more than sum of parts • Approach: Create aggregate value by integrating user contributions of unstructured content with structured data. (c) 2007 Thomas Gruber ontologiesmethods applications

  24. Example: Collective Knowledge about Travel • RealTravel attracts people to write about their travels, sharing stories, photos, etc. • Travel researchers get the value of all experiences relevant to their target destinations. http://tomgruber.org/technology/realtravel.htm (c) 2007 Thomas Gruber ontologiesmethods applications

  25. (c) 2007 Thomas Gruber

  26. Pivot Browsing – surfing unstructured content along structured lines • Structured data provides dimensions of a hypercube • location • author • type • date • quality rating • Travel researchers browse along any dimension. • The key structured data is the destination hierarchy • Contributors place their content into the destination hierarchy, and the other dimensions are automatic. (c) 2007 Thomas Gruber ontologiesmethods applications

  27. Destination data is the backbone • Group stories together by destination • Aggregate cities to states to countries, etc • Inherit locations down to photos • From destinations infer geocoordinates, which drive dynamic route maps • Destinations must map to external content sources (travel guides) • Destinations must map to targeted advertising (c) 2007 Thomas Gruber ontologiesmethods applications

  28. Contextual Tagging • Tags are bottom up labels, words without context. • A structured data framework provides context. • Combining context and tags creates insightful slices through the aggregate content. (c) 2007 Thomas Gruber ontologiesmethods applications

  29. (c) 2007 Thomas Gruber

  30. (c) 2007 Thomas Gruber

  31. Travel Recommendation Engine • Interview users about travel interests. • Match them to trips that people have written about. • Recommend places to go and things to do. (c) 2007 Thomas Gruber ontologiesmethods applications

  32. Recommendation Engine Results (c) 2007 Thomas Gruber

  33. Problems that Semantic Web could have helped • No standard source of structured destination data for the world • or way to map among alternative hierarchies • Integrating with other destination-based sites is expensive • e.g. travel guides • No standard collection of travel tags • or way to share RealTravel’s folksonomy • Integrating with other tagging sites is ad hoc • need a matching / translation service (c) 2007 Thomas Gruber ontologiesmethods applications

  34. Resources That Did Help • Open source software or free services • powerful databases • fancy UI libraries • search engines • usage analytics • Open APIs from Google (maps) and Flickr (photos) • Commercially available geocoordinate data and services (c) 2007 Thomas Gruber ontologiesmethods applications

  35. Grande Challenges • Distributing and adding structured data to systems like Del.icio.us, Wikipedia, and RealTravel • Tag spaces and tag data sharing • World destination hierarchy and other geospatial databases • Portable user identity and reputation • Site-independent rating and filtering • Semantic search and spam filtering (c) 2007 Thomas Gruber ontologiesmethods applications

  36. Vente Challenges • How to get knowledge from all those intelligent people on the Internet • How to give everyone the benefit of everyone else’s experience • How to leverage and contribute to the ecosystem that has created today’s web. (c) 2007 Thomas Gruber ontologiesmethods applications

  37. Social + Semantic Web What will the future look like? Social Web stock images from istockphoto.com; cover image by neilsethlevine.com (c) 2007 Thomas Gruber

More Related