Deepak semantic web_iitd

Preview:

DESCRIPTION

 

Citation preview

Hacking  with  Seman/c  Web  

Deepak  Shevani  Developer  @  Yahoo!  

What’s  in  here?  

•  Evolu/on  of  the  web  •  Poorly  Solved  Informa/on  Needs  •  Seman/c  Web  Technologies  •  Linked  Data  •  GeIng  Structured  Informa/on  from  Web.  

Few  Content  Creators!    Majority  Consumers!  WEB  1.0    

hKp://www.flickr.com/photos/leandrociuffo/3665883373/  

WEB  2.0    

Web  as  a  pla7orm  hKp://www.flickr.com/photos/lambertwm/4737580179/  

Ofoto   Flickr  

Personal  Website   Blogging  

Britannica  Online   Wikipedia  

Directories(taxonomy)   Tagging(“folksonomy”)  

Content  Management  Systems  

Wikis  

WEB  1.0    vs    WEB  2.0    

WEB  3.0    

hKp://www.flickr.com/photos/markhillary/337685031  

Which  direc/on  will  it  take?  

Seman/c  Web  

Pervasive  Web  

Ar/ficial  Intelligence  Personaliza/on  

Virtual  Web   WEB  3.0    

Could  be  anything!  

Tim  Berners  Lee  –  Inventor  of  the  WWW  

Web   was   designed   as   an   informaCon  space,  for  humans  as  well  as  machines.  The   informaCon   on   web   must   be  explicit  for  machines,  so  that  they  take  part   in   reasoning   and   solving   human  problems  -­‐  TBL  

A  Web  of  Documents  rather  than  Data!  

Today’s  Web  

Poorly  Solved  Informa/on  Needs  

•  Multiple interpretations –  Apple

•  Long tail queries –  Roja (I meant a south indian actress)

•  Imprecise or overly precise searches –  Brad Pitt –  pictures of strong adventures people

•  Searches for descriptions –  countries in Africa –  25 year old computer engineer living in Bangalore –  Reliable smart phone under 15,000 rupees

THE  SOLUTION  

Seman/c  Web  

Publish  data  on  the  Web  

•  Linked  Data:  linking  data  similar  to  how  we  link  documents  on  the  Web  

•  Query  databases  over  the  Web  

Architectural  Challenges  

•  A  common  format  for  sharing  data  •  Sharing  the  meaning  of  data  •  Infrastructure  

Current  Researches  &  Other  Efforts  

•  Seman/c  Web  research  into  knowledge  representa/on  and  reasoning,  data  integra/on,  data  quality  and  many  other  topics  

•  Community  effort  (Linked  Data  movement)  

Linked  Data  cloud:  interlinked  RDF  datasets  on  the  Web  

hKp://linkeddata.org/  

DBPedia  

•  Dbpedia  is  dataset  that  contains  much  of  the  structured  data  in  Wikipedia  – Data  from  the  info-­‐boxes  – Links  between  Wikipedia  pages  – Categories  – Disambigua/on  and  redirect  pages  

•  Links  to  other  datasets  

Fetching  individual  resources  

•  Use  your  web  browser  •  hKp://dbpedia.org/resource/Yahoo  redirects  to  

hKp://dbpedia.org/page/Yahoo    •  You  can  plug  in  this  URI  into  other  Linked  Data  browsers  

•  HTTP  GET  to  fetch  data  – Using  curl:  add  Accept:  applicaCon/rdf+xml  for  RDF  and  enable  redirect  

•  curl  -­‐L  -­‐H  'Accept:applica/on/rdf+xml'  'hKp://dbpedia.org/resource/Berlin’  

•  Data  dumps  –  hKp://wiki.dbpedia.org/Datasets  

References  

•  hKp://www.slideshare.net/tompraison  •  hKp://inkdroid.org/journal/2010/06/04/the-­‐5-­‐stars-­‐of-­‐open-­‐linked-­‐data/  

•  hKp://www.freebase.com/  •  hKp://dbpedia.org/About  

 

Thank  you  J