Your SlideShare is downloading. ×
0
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Combining Multimedia and Semantics (LACNEM2010)
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×
Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

Combining Multimedia and Semantics (LACNEM2010)

1,691

Published on

Invited talk given at LACNEM2010 (Cali, Colombia), in September 2010, on the combination of semantics and multimedia.

Invited talk given at LACNEM2010 (Cali, Colombia), in September 2010, on the combination of semantics and multimedia.

Published in: Technology
0 Comments
2 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
1,691
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
44
Comments
0
Likes
2
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. Combining Multimedia and Semantics<br />Oscar Corcho (ocorcho@fi.upm.es)<br />Universidad Politécnica de Madrid<br />http://www.oeg-upm.net/<br />LACNEM 2010, Cali, ColombiaSeptember 9th 2010<br />Credits: Adrián Siles, Mariano Rico, Víctor Méndez, Hector Andrés García-Silva, María del Carmen Suárez-Figueroa, Ghislain Atemezing, Raphaël Troncy<br />WorkdistributedunderthelicenseCreativeCommonsAttribution-Noncommercial-Share Alike3.0<br />http://www.slideshare.net/ocorcho<br />
  • 2. 2<br />Asunción Gómez Pérez<br />OntologyEngineering Group. Whomwe are<br />Director: A. Gómez-Pérez<br />Research Group (37 people)<br />2 Full Professor<br />4 AssociateProfessors<br />1 AssistantProfessor<br />3 Postdocs<br />17 PhD Students<br />8 MScStudents<br />2 Software Engineers<br /> Management (4 people)<br />2 Project Managers<br />1 SystemAdministrator<br />1 Secretary<br /> 50+ PastCollaborators<br /> 10+ visitors<br />
  • 3. Research Areas<br />2004<br />2008<br />1995<br />1997<br />2000<br />
  • 4. Beforewestart…<br />Howmany of youhaveeverheardabouttheword “Ontology”?<br />And howmany of you do actuallyknowwhatitmeans?<br />4<br />
  • 5. Comingtotermswithontologies and semantics<br />An ontology is an engineering artifact, which provides: <br />A vocabulary of terms<br />A set of explicit assumptions regarding the intended meaning of the vocabulary. <br />Almost always including concepts and their classification<br />Almost always including properties between concepts<br />Shared understanding of a domain of interest <br />Agreement on the meaning of terms<br />Formal and machine manipulable model of a domain of interest<br />Besides...<br />The meaning (semantics) of such terms is formally specified<br />New terms can be formed by combining existing ones<br />Can also specify relationships between terms in multiple ontologies<br />5<br />
  • 6. Example: Anontologyaboutsatellites<br />6<br />
  • 7. Outline<br />Introduction<br />What I willbetalkingabout and what I willnot…<br />Therewereseveraloptionsthat I exploredbeforeselectingtheonethatyouwillbehearing in thistalk…<br />7<br />
  • 8. Option 1: The Semantic Gap<br />The lack of coincidencebetweentheinformationthatone can extractfromthesensory data and theinterpretationthatthesame data has for a user in a givensituation<br />8<br />A.W.M. Smeulders, M. Worring, S. Santini, A. Gupta, R. Jain: Content-based image retrieval at the end of the early years, IEEE PAMI, 1349–1380, 2000.<br />However, I alreadyassumedthatEbroulwouldbetalking a lotaboutit in hisopeningkeynote (as he did). <br />Besides, I havenotworked at allonthelow-levelpart, so itmaybedifficultfor me toprovideyouwith a goodinsightonthe (many) open problems in thisarea<br />
  • 9. Option 2: MPEG-7 and the Semantic Web<br />ISO standard since December 2001<br />Main components:<br />Descriptors (Ds) and Description Schemes (DSs)<br />DDL (XML Schema + extensions)<br />Concern all types of media<br />A good number of ontologies developed around it<br />9<br />Part 5 – MDSMultimedia Description Schemes<br />
  • 10. Option 2: MPEG-7 and the Semantic Web<br />However, thetalkmay:<br /> Be a bit boring and tootechnical<br />Maylackthemix of state of the art and visionthataninvitedtalkshouldnormallyhave<br /> And MPEG-7 isnotusedtoomuch<br />…So I willcoveronlysomeaspects of thislater, when I talkabout multimedia ontologies.<br />10<br />
  • 11. Option 3: Canonical Processes of Media Production (and semantics, obviously)<br />For example….<br />http://www.cewe-photobook.com<br />Application for authoring digital photo books<br />Automatic selection, sorting and ordering of photos<br />Context analysis methods: timestamp, annotation, etc.<br />Content analysis methods: color histograms, edge detection, etc.<br />Customized layout and background<br />Print by the European leader photo finisher company<br />11<br />Credits: Raphaël Troncy, LyndaHardman<br />
  • 12. CeWe Color PhotoBook Processes<br />My winter ski holidays with my friends<br />Credits: Raphaël Troncy, LyndaHardman<br />
  • 13. CeWe Color PhotoBook Processes<br />Credits: Raphaël Troncy, LyndaHardman<br />
  • 14. CeWe Color PhotoBook Processes<br />Credits: Raphaël Troncy, LyndaHardman<br />
  • 15. CeWe Color PhotoBook Processes<br />Credits: Raphaël Troncy, LyndaHardman<br />
  • 16. CeWe Color PhotoBook Processes<br />Credits: Raphaël Troncy, LyndaHardman<br />
  • 17. CeWe Color PhotoBook Processes<br />Credits: Raphaël Troncy, LyndaHardman<br />
  • 18. Semantics can be important in the process<br />Credits: Raphaël Troncy, LyndaHardman<br />
  • 19. Option 3: Canonical Processes of Media Production<br />However, some of youprobablyattended Raphaël Troncy’stalklastyear (available in slideshare)<br />19<br />
  • 20. In summary…<br />I decidedtotalkaboutsomethingthat I havebeenworking in forthelastcouple of years, and which combines<br />Semantics (of course, thisisthekeyexpertise of ourgroup)<br />Mainlyannotation, Linked Data and a bit of Multimedia OntologyEngineering<br />Social networks, collaboration, sharing and collectiveintelligence<br />Exploiting home networks and online multimedia sites<br />And, obviously, multimedia<br />And hence I stillleaveoutmanyinterestingtopics (e.g., semanticsin user interfaces)<br />20<br />
  • 21. Outline<br />Introduction<br />What I willbetalkingabout and what I willnot<br />Sem-UPnP-Grid<br />Sharing multimedia contentacrosshomesthroughsemanticannotations<br />Credits: Mariano Rico and Adrián Siles (UPM), Víctor Méndez and José Manuel Gómez-Pérez (iSOCO), José Manuel Palacios and Mónica Pérez (TID)<br />Sem4Tags<br />Tagdisambiguation in Flickr<br />M3 Ontology(onlyif time permits)<br />A semanticbackboneforour multimedia-relatedwork<br />Conclusions and outlook<br />21<br />
  • 22. Internet<br />Motivation<br />Multimedia resources in Web2.0 are stored in centralised servers.<br />You lose some of yourrights as anauthorwhenyouuploadtheseresourcestothese servers.<br />Privacyproblems.<br />Poorannotations and metadata.<br />Theseresourcescannotbesharedwithotherresources in your home.<br />22<br />UpGrid<br />
  • 23. Multimedia Content SharingwithUpGrid<br />23<br />Annotation:<br />“Ángel onthebeach”<br />Reasoning:<br /><ul><li> “Ángel is my son”
  • 24. “Pedro is my brother”
  • 25. “Juan is my brother”
  • 26. -------------------------------
  • 27. Ángel is my nephew</li></ul>Juan<br />P2P<br />Semantic-basedquery:<br />“multimedia contentrelatedto my nephew”<br />Annotation:<br />“Ángel playing soccer”<br />Pedro<br />Additionalsemanticinformation:<br /><ul><li> “Ángel is my son”
  • 28. “Pedro is my brother”</li></ul>Additionalsemanticinformation:<br />“Juan is my brother”<br />
  • 29. Architecture<br />
  • 30. Architecture (anotherviewonit)<br />
  • 31. Snapshotsfromtheapplication<br />Checkhttp://www.youtube.com/results?search_query=UPnPGrid<br />
  • 32. Summary<br />Aneffectivemeansforsharing multimedia contentsacrosshomes, avoiding Web2.0 siteswhereyourrightsmaybecompromised<br />However, itisstill a prototype, and no serioususabilitytesting has been done<br />Muchworkstillneeded in ordertogointo a real system<br />And endusersfinditdifficulttoprovideannotations<br />Do you imagine yourparents and grandparentsannotatingphotos and videos likethat?<br />Let’sseehowthiscouldbeamelioratedwiththenextpart of ourpresentation.<br />27<br />
  • 33. Outline<br />Introduction<br />What I willbetalkingabout and what I willnot<br />Sem-UPnP-Grid<br />Sharing multimedia contentacrosshomesthroughsemanticannotations<br />Sem4Tags<br />Tagdisambiguation in Flickr<br />Credits: Héctor Andrés García Silva<br />M3 Ontology(onlyif time permits)<br />A semanticbackboneforour multimedia-relatedwork<br />Conclusions and outlook<br />28<br />Egresado de laUniversidad del Valle<br />
  • 34. Introduction<br />Social Tagging Systems<br />Web 2.0 applications <br />Applications for storing, sharing, and discovering information resources.<br />Users assign tagsto identify information resources<br />Tags are used to search/discover resources<br />29<br />
  • 35. Introduction<br />Folksonomy<br />Emerging classification scheme from social tagging systems <br />Folk: People, Taxonomy: Classification<br />Represented by: Users, Tags, Resources<br />Taxonomy<br />Folksonomy<br /><ul><li>Top-down
  • 36. Controlled Vocabulary
  • 37. Hierarchical structure
  • 38. Exclusive/Restrictive
  • 39. Expensive to maintain
  • 40. Bottom-up (user created)
  • 41. No fixed vocabulary
  • 42. No Hierarchical structure
  • 43. No Exclusive/Flexible
  • 44. Low cost</li></ul>30<br />
  • 45. Introduction<br />Why is tagging so popular?<br />Reduce cognitive burdens <br />it’s easy to use<br />Users don´t need any special skill or experience<br />The benefits of tagging are immediate<br />Future retrieval<br />Contribution and sharing<br />Attract Attention<br />Self Presentation<br />Opinion Expression<br />31<br />
  • 46. Introduction<br />However<br />Tags can be ambiguous <br />Polysemy: partyas a celebration as opposed to partyas a political organization<br />Synonym: party and celebration <br />Morphological variations: <br /> party, parties, partying, partyign<br />Plurals<br />Acronyms<br />Conjugated verbs<br />Misspelling<br />Compound words<br />Political party, PoliticalParty, Political_party,<br /> Political-Party, etc.<br />Detail/granularity level<br />A general tag as partyin contrast to a specific tag as banquet.<br />32<br />
  • 47. Motivation<br />The problem: Morphological variations, synonyms, granularity, and polysemy hamper information retrieval processes based on folksonomies. <br />Systems ignore resources tagged with morphological variationsor synonyms of that tag, as well as the resources tagged with more generic or more specific tags<br />710.659 results<br />8.661.581 Results<br />33<br />
  • 48. When searching with polysemous tags, all the resources tagged with that tag are retrieved without taking into account the tag sense the user was looking for.<br /> (e.g., Query flickr with bank results in photos about financial institutions, river edges, fog banks, and sand banks, etc. )<br />34<br />Motivation<br />
  • 49. Motivation<br />What if we associate tags with semantic entities?<br />http://morpheus.cs.umbc.edu/aks1/ontosem.owl <br />#non-work-activity<br />We can avoid the aforementioned pitfalls<br />#organization<br />#special-occasion<br />#political-entity<br />#party<br />#Celebration<br />#political-party<br />#Coalition<br />#federation<br />#Birthday<br />#Anniversary<br />uk, tories, party, conservative, speech <br />party, balloons, colors, bar, crowd<br />35<br />
  • 50. State of the Art: Semantic Grounding of Cross-Lingual Folksonomies<br />Garcia HA, Corcho O, Alani H, Gómez-Pérez A. Review of the state of the art: Discovering and Associating Semantics to Folksonomies. Knowledge Engineering Review (in press)<br />None of the analyzed approaches deals with multilingual tags<br />36<br />
  • 51. Semantic Grounding of Cross-Lingual Folksonomies<br />MSR: a Multilingual Sense Repository based on Wikipedia and enriched with semantic information taken from DBpedia.<br />Terms and <br />frequency<br />Banco<br />Bank<br />http://dbpedia.org/resource/Bank<br />Terms and <br />frequency<br />Banco<br />Cardumen<br />Swarm<br />http://dbpedia.org/resource/Swarm<br />Banco de <br />Arena<br />http://dbpedia.org/resource/<br />SandBank<br />Terms and <br />frequency<br />Sandbank<br />37<br />
  • 52. Semantic Grounding of Cross-Lingual Folksonomies<br />Sem4Tags: A process for Associating Semantics to Tags.<br />38<br />Dinero,<br />Calle,<br />Santander,<br />Money,<br />Madrid,<br />Atm, <br />cajero<br />Europe<br />Euro <br />Finance<br />Central bank<br />awesomePic<br />Nikon ..<br />Bank<br />Banco<br />http://dbpedia.org/resource/Bank<br />
  • 53. Semantic Grounding of Cross-Lingual Folksonomies<br />Disambiguation activity<br />The candidate senses and the tag context are represented as vectors. <br />The vector components are the set of most frequent terms in each Wikipedia page representing a sense.<br />For each sense the values of the vector are calculated using TF-IDF.<br />For the tag context the values in each position are 1 or 0 if the corresponding term appears in the tag context. <br />The tag context vector is compared against each sense vector using the cosine of the angle as similarity measure. <br />The most similar sense to the tag context is selected as the one representing the meaning of the analyzed tag<br />39<br />39<br />
  • 54. Semantic Grounding of Cross-Lingual Folksonomies<br />Disambiguation activity<br />We use the information of the wikipedia default sense for a term. <br />Sim(TagContext, Sensei)= λ*Cosine + β*defaultSense<br />We experimentally defined β = 0,2 and λ = 0.8<br />We attempt to use DBpedia semantic information in the disambiguation activity:<br />Sim(TagContext, Sensei)= λ*Cosine + β*defaultSense + δ*SemanticInfo<br />Studies have shown that tags in flickr refers mainly to: Locations, Time, Given Names, Potography related subjects among others. <br />We use DBpedia and YAGO relations to classify the senses according to this categories.<br />However, we found that not all the senses related to a term have the same amount of relations. (e.g. Madrid is not a city)<br />40<br />
  • 55. Let’s try it<br />http://robinson.dia.fi.upm.es:8080/SemanticTagsWebApp/index.jsp<br />Whatdoes “bernabeu” mean ifitscontextis…?<br />estadio, madrid, fútbol<br />41<br />
  • 56. Experiment<br />Baseline: Directly associate tags with DBpedia resources<br />Look for spaces and replace them with ' _‘.<br />For tags in English:<br />Create a URI of the form http://en.wikipedia.org/wiki/tag<br />Query DBpedia using the http://xmlns.com/foaf/0.1/page relation<br />For tags in Spanish:<br />Create a URI of the form http://es.wikipedia.org/wiki/tag<br />Query DBpedia using the http://dbpedia.org/property/wikipage-es relation<br />42<br />
  • 57. Experiment<br />Approaches:<br />Baseline: Selection of the sense without a disambiguation activity.<br />Sem4Tags: For each sense we use the whole Wikipedia article as source for frequentterms.<br />Sem4TagsAC: Same as Sem4Tags including the selection of the Active Context.<br />Sem4TagsAbs: For each sense we use the Wikipedia article abstract (extracted from DBpedia) as source for frequent terms.<br />Sem4TagsAbsAC: Same as Sem4TagsAbs including the selection of the Active Context.<br />43<br />
  • 58. Experiment<br />Initial Data Set<br />Wide range of Users, photos, and tags.<br />764 photos uploaded by 719 users to Flickr that have been tagged with tags describing tourist places in Spain<br />12.4 (+/- 7.85) tags per photo<br />9484 tagging activities (TAS) : <user,photo,tag><br />4135 distinct tags where used<br />Processed Data Set<br />From each photo we processed on average 2 tags <br />2260 taggingactivities (TAS)<br />44<br />
  • 59. Experiment<br />Evaluation Campaign<br />41 Evaluators<br />Evaluate semantic associations produce by each approach: <user; tag; photo; DBpedia resource; language><br />Three different evaluators evaluated each semantic association.<br />Questions:<br />Able to identify the tag meaning (known or Unknown)<br />Tag language (English, Spanish, Both, other)<br />The tag correspond to a Named entity<br />According to the identified tag language they evaluate the semantic association in terms of<br />Highly related, Related, Not Related.<br />45<br />
  • 60. Experiment<br />Results<br />Evaluators identified the semantics of the 87% of TAS (known)<br />62.6 % of TAS were considered in English<br />87.7% of TAS were considered in Spanish<br />Agreement among evaluators (Fleiss’ kappa statistics):<br />k=0.76 for highly related<br />K=0.71 for the related case/highly related case<br />46<br />
  • 61. Experiment<br />Precision and RecallforHighlyRelevantresults<br />47<br />English<br />Spanish<br />
  • 62. Experiment<br />Conclusions<br />Baseline obtained high precision, however it was able to find semantic resources for just a fraction of the analyzed data set:<br />Baseline: 27.7% in English and 19.4% in Spanish.<br />Sem4Tags: 79.1 % in English and 81.4% in Spanish<br />All approaches obtained better precision with named entities than with unnamed entities. <br />Sem4Tags and Sem4TagsAC are the approaches that obtained the best results in terms of Precision and Recall. <br />Sometimes Sem4TagsAC obtains better P@1 values but the improvements are supported by no or low statistical evidence. <br />Sem4TagsAbs and Sem4TagsAbs are clearly the worst approaches. <br />48<br />
  • 63. Outline<br />Introduction<br />What I willbetalkingabout and what I willnot<br />Sem-UPnP-Grid<br />Sharing multimedia contentacrosshomesthroughsemanticannotations<br />Sem4Tags<br />Tagdisambiguation in Flickr<br />M3 Ontology(onlyif time permits)<br />A semanticbackboneforour multimedia-relatedwork<br />Conclusions and outlook<br />49<br />
  • 64. 50<br />Ontología M3<br />
  • 65. There are already multimedia ontologies<br />MDS Upper Layer represented in RDFS<br />2001: Hunter<br />Later on: link to the ABC upper ontology<br />MDS fully represented in OWL-DL<br />2004: Tsinaraki et al., DS-MIRF model<br />MPEG-7 fully represented in OWL-DL<br />2005: Garcia and Celma, Rhizomik model<br />Fully automatic translation of the whole standard<br />MDS and Visual parts represented in OWL-DL<br />2007: Arndt et al., COMM model <br />Re-engineering MPEG-7 using DOLCE design patterns<br />However, their requirements are not always clear nor have they been developed with clear methodological guidelines<br />
  • 66. 52<br />Knowledge Resources<br />Ontological Resources<br />O. Design Patterns<br />3<br />4<br />O. Repositories and Registries<br />5<br />6<br />Flogic<br />RDF(S)<br />OWL<br />Ontological Resource<br />Reuse<br /> O. Aligning<br /> O. Merging <br />5<br />6<br />2<br />Ontology Design<br />Pattern Reuse<br />Non Ontological Resource<br />Reuse<br />4<br />3<br />6<br />Non Ontological Resources<br />2<br />Ontological Resource<br />Reengineering<br />7<br />Glossaries<br />Dictionaries<br />Lexicons<br />5<br />Non Ontological Resource<br />Reengineering<br />4<br />6<br />Classification<br />Schemas<br />Thesauri<br />Taxonomies<br />Alignments<br />2<br />RDF(S)<br />1<br />Flogic<br />O. Conceptualization<br />O. Implementation<br />O. Formalization<br />O. Specification<br />Scheduling<br />OWL<br />8<br />Ontology Restructuring<br />(Pruning, Extension, <br />Specialization, Modularization)<br />9<br />O. Localization<br />1,2,3,4,5,6,7,8, 9<br />Ontology Support Activities: Knowledge Acquisition (Elicitation); Documentation; <br />Configuration Management; Evaluation (V&V); Assessment<br />NeOnMethodology<br />
  • 67. 53<br />Ontology Requirements Specification (I)<br /><ul><li>Non-functional ontology requirements:
  • 68. Characteristics not related to the ontology content </li></li></ul><li>54<br />Ontology Requirements Specification (II)<br /><ul><li>Functional ontology requirements:
  • 69. Content specific requirements referred to the particular knowledge to be represented by the ontology
  • 70. Requirements in natural language
  • 71. in the form ofCQs
  • 72. in the form of sentences (General Characteristics)
  • 73. Strategies: (1) Top-Down, (2) Bottom-Up, and (3) Middle out</li></li></ul><li>55<br />Ontology Requirements Specification (III): Functional Requirements on M3<br />Perspectiva Multidominio<br />Perspectiva Multimedia<br />Perspectiva Multilenguaje<br />
  • 74. 56<br />Ontology Requirements Specification (IV): ORSD<br />Perspectiva Multidominio<br />M3<br />Perspectiva Multimedia<br />Perspectiva Multilenguaje<br />
  • 75. 57<br />Knowledge Resources<br />Ontological Resources<br />O. Design Patterns<br />3<br />4<br />O. Repositories and Registries<br />5<br />6<br />Flogic<br />RDF(S)<br />OWL<br />Ontological Resource<br />Reuse<br /> O. Aligning<br /> O. Merging <br />5<br />6<br />2<br />Ontology Design<br />Pattern Reuse<br />Non Ontological Resource<br />Reuse<br />4<br />3<br />6<br />Non Ontological Resources<br />2<br />Ontological Resource<br />Reengineering<br />7<br />Glossaries<br />Dictionaries<br />Lexicons<br />5<br />Non Ontological Resource<br />Reengineering<br />4<br />6<br />Classification<br />Schemas<br />Thesauri<br />Taxonomies<br />Alignments<br />2<br />RDF(S)<br />1<br />Flogic<br />O. Conceptualization<br />O. Implementation<br />O. Formalization<br />O. Specification<br />Scheduling<br />OWL<br />8<br />Ontology Restructuring<br />(Pruning, Extension, <br />Specialization, Modularization)<br />9<br />O. Localization<br />1,2,3,4,5,6,7,8, 9<br />Ontology Support Activities: Knowledge Acquisition (Elicitation); Documentation; <br />Configuration Management; Evaluation (V&V); Assessment<br />NeOnMethodology<br />
  • 76. 58<br />Scheduling using gOntt<br />gOntt<br />Life cycle model selection<br />I need to schedule the development of the M3 ontolgy network<br />Scenarios selection<br />
  • 77. 59<br />Scheduling using gOntt (II)<br />gOntt<br />
  • 78. 60<br />Knowledge Resources<br />Ontological Resources<br />O. Design Patterns<br />3<br />4<br />O. Repositories and Registries<br />5<br />6<br />Flogic<br />RDF(S)<br />OWL<br />Ontological Resource<br />Reuse<br /> O. Aligning<br /> O. Merging <br />5<br />6<br />2<br />Ontology Design<br />Pattern Reuse<br />Non Ontological Resource<br />Reuse<br />4<br />3<br />6<br />Non Ontological Resources<br />2<br />Ontological Resource<br />Reengineering<br />7<br />Glossaries<br />Dictionaries<br />Lexicons<br />5<br />Non Ontological Resource<br />Reengineering<br />4<br />6<br />Classification<br />Schemas<br />Thesauri<br />Taxonomies<br />Alignments<br />2<br />RDF(S)<br />1<br />Flogic<br />O. Conceptualization<br />O. Implementation<br />O. Formalization<br />O. Specification<br />Scheduling<br />OWL<br />8<br />Ontology Restructuring<br />(Pruning, Extension, <br />Specialization, Modularization)<br />9<br />O. Localization<br />1,2,3,4,5,6,7,8, 9<br />Ontology Support Activities: Knowledge Acquisition (Elicitation); Documentation; <br />Configuration Management; Evaluation (V&V); Assessment<br />NeOnMethodology<br />
  • 79. 61<br />Reusing Ontological Resources: Comparative Analysis (I)<br />
  • 80. 62<br />Reusing Ontological Resources: Comparative Analysis<br />
  • 81. Outline<br />Introduction<br />What I willbetalkingabout and what I willnot<br />Sem-UPnP-Grid<br />Sharing multimedia contentacrosshomesthroughsemanticannotations<br />Sem4Tags<br />Tagdisambiguation in Flickr<br />M3 Ontology(onlyif time permits)<br />A semanticbackboneforour multimedia-relatedwork<br />Conclusions and outlook<br />63<br />
  • 82. Conclusions and outlook<br />Weallagreethat…<br />Multimedia UGC has beenone of thebasis of Web2.0<br />The use of semantics can provide…<br />Betterunderstanding of thedomain and of theircontent<br />Heavyweight: addressingthesemantic gap automatically<br />Ligthweight: allowinguserstoannotate<br />Middleweight: from free tagstoknowledge<br />Betterexploratorynavigation and serendipity<br />Interconnecting multimedia contentwiththeLinked Data cloud<br />However, privacyissues are still a majorbarrierfor a largeruptake, especiallyforsomepopulationsegments<br />Allowing P2P exchangebetween “known” homes, whileexploitingsemantic-basedsearch<br />64<br />
  • 83. Combining Multimedia and Semantics<br />Oscar Corcho (ocorcho@fi.upm.es)<br />Universidad Politécnica de Madrid<br />http://www.oeg-upm.net/<br />LACNEM 2010, Cali, ColombiaSeptember 9th 2010<br />Credits: Adrián Siles, Mariano Rico, Víctor Méndez, Hector Andrés García-Silva, María del Carmen Suárez-Figueroa, Ghislain Atemezing, Raphaël Troncy<br />WorkdistributedunderthelicenseCreativeCommonsAttribution-Noncommercial-Share Alike3.0<br />http://www.slideshare.net/ocorcho<br />

×