From Enriched Museum Collections to Social Web and TV: Seminar at BBC

Preview:

DESCRIPTION

Presentaiton by Lora Aroyo and Guus Schreiber July 7, 2010

Citation preview

Integra(ng  Social  Web  &  TV    with  help  of  seman(cs  

Lora  Aroyo  &  Guus  Schreiber  Computer  Science  

VU  University  Amsterdam  

Seman(c  Web  @  VU  Amsterdam  

•  40  people,  two  groups:  Web  &  Media  (Schreiber),  Knowledge  Representa(on  &  Reasoning  (van  Harmelen)  

•  A  few  ongoing  projects:  – europeana.eu:  EU  culture  portal  – NL  projects  on  access  to  cultural  heritage:  CHIP,  Agora  – EU  NoTube:  Web  &  TV  seman(c  integra(on  

– PrestoPrime:  user-­‐generated  annota(ons  and  content  for  TV  archives  

– EU  LarKC:  plaRorm  for  massive  distributed  reasoning  

The  Linked  Data  Web:    typed  resources  and  links  

URL URL

Web link

ULAN

Henri Matisse

Dublin Core

creator

Painting “Woman with hat SFMOMA

Demo  using  linked  data  (RPI,  Hendler)  

The  power  of  simple  alignments  “Tokugawa”

SVCN period Edo

SVCN is local in-house ethnology thesaurus

AAT style/period Edo (Japanese period) Tokugawa

AAT is Getty’s Art & Architecture Thesaurus

hVp://e-­‐culture.mul(median.nl/demo/search    

Europeana  Thought  Lab  cloud  

From  metadata  to    seman(c  metadata  

1. Make vocabulary interoperable

2. Align metadata schema

3. Enrich metadata 4. Align vocabulary

SKOS EDM

Enriching  the  metadata  

Resul(ng  seman(c  annota(on    

NoTube 1st review

Personalized  Rijksmuseum  

http

://ch

ip-p

roje

ct.o

rg

05-06 May 2010

NoTube 1st review 15

Personalized  Rijksmuseum  

http

://ch

ip-p

roje

ct.o

rg

Mobile  Museum  Tour  

http://www.chip-project.org/demo/mobileguide/index.jsp

Crowdsourcing:  Video  Tagging  Games  

http

://w

aisd

a.nl

NoTube:    Making  Television  More  Personal  

hVp://www.notube.tv    

Acronym  and  consor(um  

NoTube  slogan:  Pu#ng  the  user  back  in  the  driving  seat  

Observa(ons:    

• Personalized  services  are  now  common    

• But:  user  data  is  s(ll  under  control  of  separate  applica(ons  

• Result:  user  is  faced  with  mul(tude  of  distributed  personal  data,  hidden  in  tons  of  inaccessible  cookies  

NoTube  building  blocks  (1)  

1.  TV  metadata  services  EPG  metadata  grabbers    

–  from  170+  channels  (issue:  channel  URLs)  –  metadata  format:  TV  Any(me    –  real-­‐(me  service    

2.  Metadata  enrichment    – Add  links  to  external  Web  vocabularies  and  repositories:  Lupedia  service  

NoTube  building  blocks  (2)  

3.  Linked  Open  Data  for  TV  – Access  services  to  major  vocabularies    – Alignment  services  between  major  vocabularies,  where  needed  (e.g.  genre  typologies)  

4.  User  acKvity  streams  – Standard  for  ac(vity  stream  representa(on,  i.e.  Atom  Ac(vity  Stream  

– Access  services  to  ac(vity  streams,  e.g.  YouTube,  TwiVer,  ..  

– Trusted  access  to  “friend”  informa(on,  e.g.  implementa(on  of  standard  like  OAuth  2.0  

NoTube  building  blocks  (3)  

5.  User  profiling  Services  for  genera(ng  user  preferences    –  “Beancounter”  

 –  abstrac(ons  from  ac(vity  stream    

User-­‐model  representa(on  based  on  FOAF,  i.e.  weighted  interests  and  considering  context  

6.  Recommender  services    Collabora(ve  recommenders,  e.g.  preferences  of  friends  

Content-­‐based  recommenders,  e.g.  program  about  Alma  Mahler  

 program  about  Walter  Gropius    Experiment  with  mix  of  these  recommenders  for  single  users  and  small  groups  of  users,  e.g.  families,  friends  

Usage  scenario  &  demo  

http://notube.tv

A world of opportunities is opening!

Thank  you  Ques(ons?  

Recommended