SPARQL2NL: converting SPARQL queries to natural language

Over the last years, Semantic Web and Linked Data technologies have reached the backend of a considerable number of applications. While experts can easily access the data stored in these backends via the W3C standard SPARQL, most lay users do not understand this query language. Thus, they mostly have to rely on forms and query builders. We address this drawback by presenting a framework that allows converting virtually any SPARQL 1.0 query into natural language. Our framework implements a bottom-up approach that consists of normalizing the input query, converting it to natural language and reducing the query to make easily understandable. As our approach is generic, it not only allows transforming queries but also to represent the answers to queries in natural language. Therewith, our approach can be used to enable users accessing triple stores without even having to deal with SPARQL or RDF.

Demo Source Code

Survey (not active anymore): http://blog.aksw.org/2012/sparql2nl-survey/

Survey Winners (those who agreed that we can publish their names): Ben Companjen, Thimo Thoeye, Luke Opperman, Jeffrey Putnam, Tobbe Sjogren, Ghalem Ouadjed

AutoSPARQL TBSL Demo User Interface (using SPARQL2NL)

Involved CITEC members:

Dr. Christina Unger

Project Team

News

No recent news available.