The Semantic Web Landscape

1. The Semantic Web Landscape Dean Allemang Working Ontologist, LLC

2. Semantic Web – what it is and isn’t Lot’s of cool technologies that could be seen as “Semantic”:

3. What does “Semantic” mean? Definition of SEMANTIC 1:of or relating to meaning in language What is the meaning of meaning?

4. What is the “Web”? A way to share documents on a global scale. Anyone can read any document on the web

5. What is the “Semantic Web”? A way to share data on a global scale. Anyone can queryanydata on the web all

6. WANDERING IN THE WILDERNESS

7. Living in the data wilderness

8. Life in the Data Cathedral

9. Growing a business is an uncertain endeavor – enterprise workers are finding their way through a wilderness of valuable information. • As a group, they try to make order of the chaos. Some embark on very formal data management projects. These projects develop a sort of ‘priesthood’ (the churches) • Some rogues set out on their own, building their own resources (shrines) • As a group, they dream of something that can unite them all – a way to have someone manage all their data (MDM – “Single version of the truth”) • Single Version of the Truth remains elusive

10. Moral to the Story? Whether you like it or not, a lot of your life will be spent in the wilderness. And a question - how do we help people cope in the wilderness?

11. Problems you can solve on the web • Search (help me find what I am looking for) • Comparison (help me compare two things) • Route planning (get me from here to there) • Diagnosis (what’s wrong with me?) • Measurement (how much? How far?) • Broadcast (let everyone know) • … Semantic Web doesn’t do any of these – it allows data to be shared, to facilitated these functions and more

12. Example: search An index to guide the search Stuff to search through (information, products, services, …

13. Example: search • Text analytics? • Link analytics? • Crowdsourcing? • Self-annotation? • Expert curation? Where does that index come from?

14. Sharing data on the Semantic WeB

15. Semantic Web description of Mathematics (abridged) Semantic Web – emphasis is on the “WEB” about Arithmetic Fermat’s Theorem Algebra has broader term about Sphere Packing Mathematics Geometry solved Probabilities has broader term works in solved Game Theory Wiles Hales about works in about Siegel studied at algebra.com studied at miseregames.com Cambridge

16. Semantic Web description of Mathematics (abridged) Semantic Web – emphasis is on the “WEB” about Arithmetic Fermat’s Theorem about Sphere Packing Geometry solved wikipedia Wiles studied at Cambridge

17. Semantic Web description of Mathematics (abridged) Semantic Web – emphasis is on the “WEB” Arithmetic Algebra has broader term Mathematics Geometry Probabilities has broader term Game Theory Library of Congress

18. Semantic Web description of Mathematics (abridged) Semantic Web – emphasis is on the “WEB” Sphere Packing Geometry works in solved Hales studied at https://sites.google.com/site/thalespitt/bio Cambridge

19. Semantic Web description of Mathematics (abridged) Semantic Web – emphasis is on the “WEB” Game Theory miseregames.com works in about Siegel miseregames.com

20. Semantic Web description of Mathematics (abridged) Semantic Web – emphasis is on the “WEB” Algebra about algebra.com algebra.com

21. On the Semantic Web, metadata is as easy to manage as data

22. Example: Structured Search “Find me someone who has ‘Semantic Web’ as part of their job title” What kind of index would we need to answer this sort of question?

23. Interlingua name job title work phone email photo Suppose we all agreed on how to represent information about a person. Suppose that we all used that to mark up our pages. Now we could search all the information about all the people in a uniform manner. This is the basis of microformats and schema.org

24. Interlingua name job title work phone email photo full name position office phone email image What if there are competing formats? Easy – get the other guy to change!! … or have a way to talk about how they relate.

25. What is an “Ontology”? • Person • Name • Title • First, middle, last • Position • Organization • Title • Address • Street address • City • State • Country • Zip Code • Phone number • Work • Home • mobile • Employee • Name • Employee number • First, middle, last • Job • Division • Job Description • Address • Street address • City • County • Country • Post code • Phone number • Work • Home • mobile • Sharable, modular piece of metadata. • Sharable = referenceable from other places • Modular means self-contained • Metadata describes other data

26. Web of metadata Ontologies refer to one another just as web pages do • Person • Name • Title • First, middle, last • Position • Organization • Title • Address • Street address • City • State • Country • Zip Code • Phone number • Work • Home • mobile • Person • Name • Title • First, middle, last • Position • Organization • Title • Address • Street address • City • State • Country • Zip Code • Phone number • Work • Home • mobile • Person • Name • Title • First, middle, last • Position • Organization • Title • Address • Street address • City • State • Country • Zip Code • Phone number • Work • Home • mobile • Person • Name • Title • First, middle, last • Position • Organization • Title • Address • Street address • City • State • Country • Zip Code • Phone number • Work • Home • mobile • Person • Name • Title • First, middle, last • Position • Organization • Title • Address • Street address • City • State • Country • Zip Code • Phone number • Work • Home • mobile

27. Semantic Agents on the web • Agents provide services for people • Agents read data from pages • Agents merge data from multiple pages • Examples: • Mashups • Auto-fill forms • Facebook “like” or “share” buttons • Person • Name • Title • First, middle, last • Position • Organization • Title • Address • Street address • City • State • Country • Zip Code • Phone number • Work • Home • mobile • Person • Name • Title • First, middle, last • Position • Organization • Title • Address • Street address • City • State • Country • Zip Code • Phone number • Work • Home • mobile

28. Two Strategies for How The Web was Won • First, get everything into the same form, get everyone to agree, get them all into the Cathedral. Then, data exchange is easy. • Learn to live in the wilderness, making use of the valuable things there, while dealing with the dangerous ones.

29. RDF and Corporate Data(i.e., spreadsheets and relational databases)

30. RDF distributes tables

31. Distribute column-by-column? Which row is which?

32. Distribute row-by-row? Which column is which?

33. Distribute cell-by-cell? mat219 eth125 Which column is which? bio100 Which row is which?

34. Triples in RDF Predicate mat219 Object Subject <mat219> <Course Title> "Introduction to College Algebra" . “triple” Introduction To College Algebra course title mat219

35. RDF distributes tables Cultural Diversity eth125 Sociology Sociology soc120 Algebra 1a Mathematics mat116 Introduction To College Algebra mat219 Science Finite Math mat205 Statistics mat230 Geology course title 5 Introduction To Life Science With Lab sci245 area 3 bio100 hours Environmental Science 7 sci275

36. RDF Combines Information Science Sociology Mathematics has broader term

37. RDF distributes tables Cultural Diversity eth125 Science Sociology soc120 Algebra 1a Sociology mat116 Introduction To College Algebra Mathematics mat219 Finite Math mat205 Statistics mat230 Geology course title 5 Introduction To Life Science With Lab sci245 area 3 bio100 hours Environmental Science has broader term 7 sci275

38. Semantic WEB AND Metadata

39. Project Failure Recipe Step 1. Build a model that encompasses all data that will be meaningful in your business 1 Step 2. Map all data sources to this universal model 2 Step 3. Start reaping benefits from your integrated data3 1 GET IT RIGHT this time! 2 This is easy, since you GOT IT RIGHT (see 1) 3 Enough benefits to pay off the initial investment

40. What is “Metadata”? • Bibliographic data – date created, author, Intellectual Property rights, provenance, … • Cataloging, indexing, classification, tagging, SEO – what kind of thing is this? • Schema information – what are the sorts of things this data contains?

41. 3 Schema Metadata

42. Schema Metadata Excerpt from UDEF – types of “Person” a dataset might mention

43. UDEF overview UDEF Entities UDEF Relationships a.a.a.j.5 Domain Entities Domain Relationships Procurement Officer Domain Entities Domain Relationships FAR officer

44. Is UDEF “Getting it right”? Schema.org UDEF Entities microformat microformat microformat Domain Entities Domain Entities Domain Entities Domain Entities

45. 2 Catalog Metadata

46. Classification Systems/Controlled Vocabularies • Library of Congress Subject Headings (etc.) • Dewey Decimal • AGROVOC • National Agriculture Library • Association for Computing Machinery subject headings • STW Economics vocabulary • American Psychological Society • Amazon subject headings

47. How to use a classification system Arithmetic Algebra has broader term Mathematics Geometry Probabilities has broader term Game Theory John H. Conway, On numbers and games, second ed., A. K. Peters, Ltd., Natick, MA, 2001. Library of Congress ?

48. Catalog Metadata Excerpt from AGROVOC – types of cattle a document might talk about

49. Facebook and Controlled Vocabularies Controlled Vocabulary : People known to Facebook (filtered by friends first) Controlled Vocabulary : Places known to Facebook (filtered by location)

50. Content Management Challenges in the Enterprise Too much backlog Knowledge Workers are too busy to tag documents New documents show up too fast Knowledge Workers are lazy We need a better vocabulary They don’t understand the value of tagging We need software to help them tag