Semantic Web - Applications / search engines

1,862 views
1,734 views

Published on

Published in: Technology, Education
0 Comments
5 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,862
On SlideShare
0
From Embeds
0
Number of Embeds
8
Actions
Shares
0
Downloads
0
Comments
0
Likes
5
Embeds 0
No embeds

No notes for slide

Semantic Web - Applications / search engines

  1. 1. Faculty of Science, Technology and Communication (FSTC) Bachelor en informatique (professionnel) Semantic Web Unit 10: Semantic Search Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 1
  2. 2. 10. Semantic search Semantic Web Roadmap: Controlled growth bottom up according to this architecture. Architecture was (slightly) modified in the last years. Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 2
  3. 3. 10. Semantic search 10.1. Google 10.2. Insert semantics into HTML 10.3. Multimedia information retrieval 10.4. Towards a semantic search engine 10.5. Visions and outlook 10.6. References Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 3
  4. 4. 10. Semantic search 10.1. Google Boiling point of Radium (Ra) 1140 °C google.com 1737 °C chemicalelements.com 1413 °C education.jlab.org 1500 °C chemicool.com 1536 °C environmentalchemistry.com all sites from first page of Google results Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 4
  5. 5. 10. Semantic search 10.1. Google Google keyword trends semantic web porn http://www.google.com/trends Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 5
  6. 6. 10. Semantic search 10.1. Google https://www.youtube.com/watch?v=wSF82AwSDiU Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 6
  7. 7. 10. Semantic search 10.1. Google same result with “google works does how” source: http://www.google.com/trends Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 7
  8. 8. 10. Semantic search 10.1. Google Page Rank, by Larry Page (1998) The “Page Rank” of a web page depends on the number of incoming links The PR of each web page is initialized equally, here: 0,25. Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 8
  9. 9. 10. Semantic search 10.2. Insert semantics into HTML Information retrieval process knowledge representation Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 9
  10. 10. 10. Semantic search 10.1. Google The answer is not part of the query Linked data “Best results” first  ranking Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 10
  11. 11. 10. Semantic search 10.2. Insert semantics into HTML Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 11
  12. 12. 10. Semantic search 10.2. Insert semantics into HTML Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 12
  13. 13. 10. Semantic search Microformats web based approach to semantic markup http://microformats.org/ 10.2. Insert semantics into HTML Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 13
  14. 14. Microdata 10.2. Insert semantics into HTML Supporting vocabulary that can be used by applications, i.e., search engines 10. Semantic search Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 14
  15. 15. 10. Semantic search 10.2. Insert semantics into HTML Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 15
  16. 16. 10. Semantic search 10.2. Insert semantics into HTML <?xml version="1.0"?> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:myNS="http://www.linckels.lu/myNS/" xmlns:foaf="http://xmlns.com/foaf/0.1/"> <rdf:Description rdf:about="http://www.linckels.lu/Demo"> <foaf:name>Neil Armstrong</foaf:name> <myNS:hasWalkedOn rdf:resource="http://en.wikipedia.org/wiki/Moon" /> </rdf:Description> </rdf:RDF> Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 16
  17. 17. 10. Semantic search 10.2. Insert semantics into HTML Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 17
  18. 18. 10. Semantic search 10.3. Multimedia information retrieval Information retrieval process knowledge mining Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 18
  19. 19. 10. Semantic search 10.3. Multimedia information retrieval ambiguity alert Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 19
  20. 20. 10. Semantic search 10.3. Multimedia information retrieval ambiguity alert TOP Human Thing Man Cyclist Astronaut Place Automobile Musician Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 20
  21. 21. 10. Semantic search 10.3. Multimedia information retrieval Multimedia content and metadata Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 21
  22. 22. 10. Semantic search 10.3. Multimedia information retrieval Multimedia content and metadata Document (d1) Query (q) Aim of information retrieval Document (d2) compute similarities: sim(q,d1) vs. sim(q,d2) Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 22
  23. 23. 10. Semantic search 10.3. Multimedia information retrieval Boolean model Index terms: K = {nude, woman, street} Document (d1) Document (d2) Query (q)            0 1 1  1 d            1 1 0  2 d            1 0 0  q Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 23
  24. 24. 10. Semantic search 10.3. Multimedia information retrieval Vector model Index terms: K = {nude, woman, street} Document (d1) Document (d2) Nude (y) q d2 Query (q) Woman (x) d1 Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 24
  25. 25. 10. Semantic search 10.3. Multimedia information retrieval Other models Probabilistic model Latent semantic indexing model Fuzzy set model Neural network model Baysian network ... Classical MIR vs. semantic search Classical approaches fail when it comes to more complex queries Need of better human-machine interfaces, e.g., natural language input Semantic search is not based on keyword / index term checking, but on the reasoning over the sense of the metadata Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 25
  26. 26. 10. Semantic search 10.4. Towards a semantic search engine Semantic interpretation Translation from an informal language into a formal language D1  Picture hasColor.BW isArtistic pictureOf.(Woman isNaked) isLocated.(Street isNarrow) D2  Frame isNarrow hasContent.(Woman isColor.Black) hasContent.(Man isColor.Black) Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 26
  27. 27. 10. Semantic search 10.4. Towards a semantic search engine Semantic interpretation Translation from an informal language into a formal language D1  Picture hasColor.BW isArtistic pictureOf.(Woman isNaked) isLocated.(Street isNarrow) similarity Q  Photo hasColor.BW Q  Photo Picture hasColor.BW photoOf.(Woman isNude) isOutdoors.Daylight pictureOf.(isNaked) isOutdoors.Daylight similarity D2  Frame isNarrow hasContent.(Woman isColor.Black) hasContent.(Man isColor.Black) Ontology inside image picture photo movie Photo Picture pictureOf  photoOf isNude  isNaked equivalences Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 27
  28. 28. 10. Semantic search 10.4. Towards a semantic search engine Semantic distance D1  Picture hasColor.BW isArtistic pictureOf.(Woman isNaked) isLocated.(Street isNarrow) similarity Q  Photo Picture hasColor.BW pictureOf.(Woman isNaked) isOutdoors.Daylight Miss(Q,D): requested in Q but not delivered in D Miss(Q,D1) = Q − lcs(Q,D1) Miss(Q,D1) = Photo isOutdoors.Daylight |Miss(Q,D1)| = 1 + (2 + 1) = 4 Rest(Q,D): delivered in D but not requested in Q Rest(Q,D1) = C − lcs(Q,D1) Rest(Q,D1) = isLocated.(Street isNarrow) isArtistic |Rest(Q,D1)| = 2 + (1 + 2) + 2 = 7 cover(Q,D): requested in Q and delivered in D cover(Q,D1) = Picture hasColor.BW pictureOf.(Woman isNaked) Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 28
  29. 29. 10. Semantic search 10.4. Towards a semantic search engine Semantic distance Q  Photo Picture hasColor.BW pictureOf.(Woman isNaked) isOutdoors.Daylight Miss(Q,D): requested in Q but not delivered in D Miss(Q,D2) = Q − lcs(Q,D2) Miss(Q,D2) = Q |Miss(Q,D2)| = 13 Rest(Q,D): delivered in D but not requested in Q Rest(Q,D2) = C − lcs(Q,D2) Rest(Q,D2) = D2 |Rest(Q,D1)| = 13 cover(Q,D): requested in Q and delivered in D cover(Q,D2) =Т similarity D2  Frame isNarrow hasContent.(Woman isColor.Black) hasContent.(Man isColor.Black) Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 29
  30. 30. 10. Semantic search 10.4. Towards a semantic search engine Semantic distance Query (q) Document (d1) Document (d2) |Miss(Q,D1)| = 4 |Rest(Q,D1)| = 7 |Miss(Q,D2)| = 13 |Rest(Q,D2)| = 13 Best cover object with smallest rest and miss preference is given to smallest miss Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 30
  31. 31. 10. Semantic search 10.4. Towards a semantic search engine Examples of semantic search engines E-Librarian Service – www.linckels.lu/research/elibrarian Hakia – www.hakia.com Ask.com – www.ask.com WolframAlpha – www.wolframalpha.com Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 31
  32. 32. 10. Semantic search 10.5. Visions and outlook A vision on the evolution… Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 32
  33. 33. 10. Semantic search 10.5. Visions and outlook Semantic Web tools Categories: – Triple Stores – Inference engines – Converters – Search engines – Middleware – CMS – Semantic Web browsers – Development environments – Semantic Wikis – … Some names: – Jena, AllegroGraph, Mulgara, Sesame, flickurl, … – TopBraid Suite, Virtuoso, Falcon, Drupal 7, Redland, Pellet, … – Disco, Oracle 11g, RacerPro, IODT, Ontobroker, OWLIM, Talis Platform, … – RDF Gateway, RDFLib, Open Anzo, Zitgist, Protégé, … – Thetus publisher, SemanticWorks, SWI-Prolog, RDFStore… Deployment communities Major communities pick the technology up: digital libraries, defense, eGovernment, energy sector, financial services, health care, oil and gas industry, life sciences … Semantic Web also appear in the “Web 2.0/Web 3.0” applications exchange of social data, personal “space” applications, dynamic Web site backends, multimedia asset management, etc. Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 33
  34. 34. 10. Semantic search 10.6. References E-Librarian Service User-Friendly Semantic Search in Digital Libraries Serge Linckels, Christoph Meinel Creating the Semantic Web with RDF: Professional Developer's Guide Johan Hjelm Foundations of Semantic Web Technologies Pascal Hitzler, Markus Krötzsch, Sebastian Rudolph Semantic Web ::: Serge Linckels ::: http://www.linckels.lu/ ::: serge@linckels.lu 34

×