Three, then Saarbrücken-based, scientists were awarded for a publication issued 11 years ago. The committee acknowledged with the SEOUL Test of Time Award the long-lasting and sustainable influence of the ontology “YAGO”. Since a decade, this semantic and formal representing data base extracts and links text-based entries from Wikipedia and other internet sources.
The project YAGO – Yet Another Great Ontology – was publically introduced by an scientific paper of Fabian Suchanek, Gjergji Kasneci und Gerhard Weikum presented during the International World Wide Web Conference in 2007. This is the most important conference for research about the World Wide Web.
YAGO is a light-weight and extensible ontology that automatically searches, finds and extracts relations between entities in internet sources like Wikipedia. The scientists use a carefully designed combination of rule-based and heuristic extractions methods. The relations found by the automatic search algorithm have very low error rate. Meanwhile, YAGO contains more than 100 million facts about 10 million persons, locations, organizations, and works, structured in hundreds of thousands semantic classes.
Gerhard Weikum comments the recognition by the Test of Time Award: “In 2006, when we started with the YAGO project, many said that it is hopeless and useless. Today, there are huge knowledge graphs at all large internet players supporting important online services with background information. For example, when searching for people, products or events with both Google and Microsoft Bing, you often get precise answers that come from the respective knowledge bases.”
YAGO is jointly developed and continuously improved by scientists at the Max Planck Institute for Computer Science and the Télécom ParisTech University in Paris. The knowledge base itself and the code for its automatic construction are freely available as open source.
From the laudatio:
YAGO was among the first projects to extract semantic knowledge at large scale from Wikipedia. Together with DBpedia, it is one of the pioneering contributors of the web of data. Over the years, several versions of YAGO were released continuously, contributing to the open linked data content and structure, and supporting many applications.
Test of Time Awards are given by a number of computer science organizations. In contrast to outstanding first published findings, they honor trend-setting research results after they have proven their relevance in the long run. Therefore, Test of Time Awards focus on the long-term influence of a research topic.
Webpage Gerhard Weikum
Prof. Dr. Gerhard Weikum