Answering end-user questions, queries and searches on Wikipedia and its history

Knowledge bases (KBs) encoded using RDF triples deliver many benefits to applications and programmers that access the KBs on the web via SPARQL endpoints. In this paper, we describe and compare two user-friendly systems that seek to make the universal knowledge of Web KBs available to users who neither know SPARQL, nor the internals of the KBs. We first describe CANaLI, that lets people enter Natural Language (NL) questions and translates them into SPARQL queries executed on DBpedia. CANaLI removes the ambiguities that are often present in NL communication by requiring the use of a Controlled NL and providing on-line knowledge-driven question-completion that shows alternate correct interpretations. While CANaLI is a very powerful NL system, which placed first in the 2016 competition on Question Answering over Linked Data QALD-6, even more powerful user-friendly interfaces are available to users who enter questions and queries on web-browsers. In particular, the SWiPE system provides a wysiwyg interface that lets users specify powerful queries on the Infoboxes of Wikipedia pages in a query-by-example fashion. Thus, in addition to those supported in CANaLI, we now have queries with (i) complex aggregates, (ii) structured conditions combined with keyword-based searches, and (iii) temporal conditions on Cliopedia, a historical knowledge base that captures the evolution of Wikipedia entities and properties. These systems demonstrate that semi-curated web document corpora and their KBs are making possible the seamless integration through user-friendly interfaces of (i) NL question answering, (ii) structured DB queries, and (iii) information retrieval. These were once viewed as distinct functions supported by different enabling technologies.