Docstoc

Slide 1 - Vrije Universiteit Amsterdam

Document Sample
Slide 1 - Vrije Universiteit Amsterdam Powered By Docstoc
					Social Networks and
 the Semantic Web
   The Perfect Storm




          Peter Mika // Vrije Universiteit, Amsterdam
Hurricane Katrina…
… hit New Orleans…
…and rendered
most of the city
uninhabitable…
…residents took refuge…
…all over the United States…
…looking for the missing…
…mostly on the Web.
What’s the problem?
         What doesn’t work
But we still have search, right?
  Search on what?
  Search how?


    New orleans west bank found nine year old niece of Theresa


The non-solution: crisis.gov
  Throw away the Web:
   robustness, openness, innovation
   The Katrina PeopleFinder
Part social, part technical effort:
  The PeopleFinder Information Format
  Extracting and entering records
  katrinalist.net




                                   XML
                                repository
       Where do we stand?
Comprehensive database for search,
 visualization and developing additional
 services

Information loss
  A single schema proves to be a bottleneck

Limited by human effort
 Entering records
 Removing duplicate records
 Matching of safe and missing persons
               In this Thesis…

Social network mining from the Web
   Statistical methods
   Evaluation


Semantic-based representation of personal and social
   information

Aggregation of information using machine intelligence

Different scenario
   Semantic Web research community
…replacing human effort




         RDF/OWL
         repository
...how the story ends…
               In the end…
No usage of semantic technology in PF
  An emergency is not the time to educate
  ShelterFinder: finding duplicate shelter records


PeopleFinder shuts down
  Most found their loved ones
  Cease and Desist threats
  Data donated to a larger repository (Red Cross,
   Microsoft)

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:13
posted:1/22/2012
language:
pages:17