Upcoming Research Talk on Wikidata: Collaborative Modeling of Incomplete Knowledge and Disagreement

The computer science department by way of our research group is hosting a colloquium talk (free and open to the public):

  • by: Daniel Kinzler, Wikimedia Deutschland
  • about: Collaborative Modeling of Incomplete Knowledge and Disagreement
  • on: Monday, 18. April 2016, 16:15 Uhr
  • at: Cauerstr. 11, 91058 Erlangen, Room 01.150-128

Abstract: Wikidata is a knowledge base anyone can edit – a machine readable Wikipedia, so to say. Over the three years since we launched Wikidata, it has become the world’s largest general purpose knowledge base. To make this possible, we had to think about scalability on every level – not only for the technical systems that deliver content over the web, but also for the processes data structures needed for collaborative modeling of the world’s knowledge. One key idea that allows us to scale this process so well is the notion that we are not modeling reality, but knowledge about reality. “Knowledge” in this context means verifiable claims by authorities. As such, it’s not absolute: it my change, be contested, uncertain, or dependent on context. When designing the data model for Wikidata, we tried to strike a balance that allows us to model such uncertain knowledge, while still keeping the data model simple enough to be usable by casual contributors.

Speaker: Daniel Kinzler is the lead developer of the Wikidata project at Wikimedia Germany. He has been active on Wikipedia since 2004 and contributed to MediaWiki since 2005. He has a diploma in Informatics with a thesis about data extraction from Wikipedia.