Web of Data: Describing and Linking Entities

Abstract

The Web bears the potential of being a universal source of knowledge used to answer questions, retrieve facts, solve problems, or create new knowledge. Many major scientific discoveries have been made possible by recognizing the connections across domains or by integrating insights from several sources [Gruber, 2008]. This process requires accessing deep insights involving the actual subject matter of scientific domains that clearly goes beyond simple associations of words supported by the existing Web infrastructure. The emerging Web of data aims to support a global data infrastructure, in which things of the real world (i.e., entities) are described on the Web by data rather than documents. This generation of the Web infrastructure is expected to transform the way structured information is exploited across domains at a large scale, and thus plays the role of a key-enabling technology for the forthcoming data-driven economy.¹ The Web of data has been proposed as a stripped-down version of the W3C specifications for the Semantic Web [Heath and Bizer, 2011] and has been boosted by the proliferation of scientific datasets made available worldwide according to the fourth paradigm of science [Hey et al., 2009], as well as of high-quality open encyclopedias, like Wikipedia.²