Material Detail

FactForge: Data Service and the Value of Inferred Knowledge

FactForge: Data Service and the Value of Inferred Knowledge

This video was recorded at European Data Forum (EDF), Copenhagen 2012. Linked Open Data movement is maturing. Not only LOD cloud increases by billions of triples yearly, but also technologies and guidelines about how to produce LOD fast, how to assure their quality, and how to provide vertical oriented data services are being developed (LOD2, LATC, baseKB). Little is said however about how to include reasoning in the LOD framework, and about how to cope with its diversity. In this talk we will present FactForge, a reason-able view on the web of data, which comprise a segment of LOD cloud, e.g. DBPedia, Freebase, Geonames, Wordnet, NY Times, Musicbrainz, Lingvoj, Lexvo, CIAFactbook, loaded in a single repository (OWLIM), and forming a compound dataset, on which inference is performed. This results in 40% increase of the knowledge available for querying to about 15 billion statements. The diversity of LOD makes their use and querying extremely challenging, as one has to be intimately familiar with the schemata underlying each dataset. Initiatives and research projects like, UMBEL, BLOOMS+, ALOCUS which try to involve the notion of a golden standard at schema level to allow better interoperability of LOD and the WWW in general, are indicative for the search of a solution along these lines. The new version of FactForge which will be shown in this talk and in the making for several years now, aligns with these views. It is supplied with a reference layer of the upper-level ontology PROTON, which is mapped to the ontologies of the LOD datasets in FactForge, making their instances accessible via PROTON concepts and properties. This reference layer makes loading of the LOD ontologies unnecessary, optimizing the reasoning processes, and allows for quick and seamless data integration of new datasets with the entire LOD segment of FactForge. It also ensures better interfacing with other components via SPARQL as the queries are more compact and easy to formulate, faster response times, because of less joins are employed, and a wealth of inferred knowledge across the datasets, which allows for real journey of knowledge discovery, and navigation from different stand points. FactForge is the largest body of general knowledge and LOD on which inference is performed. We will present applications which make use of FactForge and emphasize the role of inferred knowledge in them produced by the reason-able views, and will argue for a new paradigm of data services, based not only on linked data verticals but also on inferred knowledge.


  • Editor Reviews
  • User Rating
  • Comments
  • Learning Exercises
  • Bookmark Collections
  • Course ePortfolios
  • Accessibility Info

More about this material


Log in to participate in the discussions or sign up if you are not already a MERLOT member.