High-Precision Semantic Search by Generating and Testing Questions.
William J. Black, C. J. Rupp, Chikashi Nobata, John McNaught, Jun’ichi Tsujii, Sophia Ananiadou · Research Explorer (The University of Manchester) · 2010
A novel way of allowing users to query large collections of scientific papers, looking for facts rather than documents, is presented. This is done in such a way that from ordinary search terms, queries are generated and tested, and only those known to match facts expressed within individual document sentences are proposed to the user, who then selects from a drop-down list.