The University of Sheffield TREC 2002 Q&A System

M. Greenwood,Ian Roberts,R. Gaizauskas
Abstract:The system entered by the University of Sheffield in the question answering track of TREC 2002 represents a significant development over the Sheffield system entered into TREC-8 [9] and TREC9 [15], although the underlying architecture remains the same. The essence of the approach is to pass the question to an information retrieval (IR) system which uses it as a query to do passage retrieval against the text collection. The top ranked passages output from the IR system are then passed to a modified information extraction (IE) system. Syntactic and semantic analysis of these passages, along with the question, is carried out to identify the “sought entity” from the question and to score potential matches for this sought entity in each of the retrieved passages. The potential matches are then combined or discarded based on a number of criteria. The highest scoring match is then proposed as the answer to the question.
What problem does this paper attempt to address?