Summary

The Semantic Web as proposed by Tim Berners-Lee is a vision of the Web as a universal, distributed database that connects and information and knowledge from different sources. All information should be encoded in a form that can be understood and interpreted by computer programs. However the Semantic Web lacks free information about general concepts of the world like places, people, activities and so one. This information is collected in Wikipedia but in natural language for humans. In the recent years more and more methods and projekts have been started to use Wikipedia as a main source of general information for the Semantic Web: Semantic MediaWiki provides typed links and attributes to explicitely encode semantics in Wiki pages, dbpedia tries to extract semantic that is implicitely contained in Wikipedia, and templates are more and more used to uniformly specify information in Wikipedia articles. But the formalization of knowledge in fixed data fields and strictly defined properties is a problem especially if information is controversial and fuzzy. Natural language provides many methods to deal with disambiguation, fuzzyness and discrepancy. One of Wikipedia's fundamental principles, the Neutral Point of View requires to fairly represent and without bias all significant views on a given topic. Additionally Wikipedia provides discussion pages and version history. If semantic information is extracted from Wikipedia to be used in other applications, this context gets lost: only one view can practically be considered. This presentation will show examples of Semantic Wikipedia and why it is both unavoidable and useful to Wikipedia. In the second part the conflict with NPOV will be explained to start a discussion about the responsibility of Wikipedia's community in the Semantic Web.

Outline

  1. The Semantic Web
  2. The role of Wikipedia
  3. The Problem
  4. Possible solutions