4 ms·
Reasoning on Web Data: Algorithms and Performance [pdf]
- jcr 11y agoI found the submitted paper on the home page of one of its authors, Ioana Manolescu. [1] According to her page, the paper is designed to be more of a tutorial. Having a rough understanding of the Resource Description Framework (RDF) [2], RDF Schema (RDFS) [3], and RDF Query Language (RDFQL) [4] certainly helps, but is not required. If you think about the paper in terms of how a major search engine like Google works, and you remember the occasional "index updates" they used to do a few times a year, you'll see the "saturation" or "closure" approach mentioned in the paper. i.e. > "compile the knowledge into data, as follows. Infer all possible facts derived from the base data and by the semantic constraints prior to querying the database; add these facts to the database; then, process queries on the (enriched) database ignoring the constraints which have lead to it." Since Google no longer does occasional "index updates" and now updates continuously, it seems they may have moved to something similar to the alternate "reformulation" approach mentioned in the paper: > "The alternative technique is called reformulation. Here, the database is left unchanged, while queries are modified (reformulated) to take into account all the known semantic constraints." [1] http://pages.saclay.inria.fr/ioana.manolescu/ http://pages.saclay.inria.fr/ioana.manolescu/ [2] https://en.wikipedia.org/wiki/Resource_Description_Framework https://en.wikipedia.org/wiki/Resource_Description_Framework [3] https://en.wikipedia.org/wiki/RDFS https://en.wikipedia.org/wiki/RDFS [4] https://en.wikipedia.org/wiki/RDF_query_language https://en.wikipedia.org/wiki/RDF_query_language