Passage retrieval in a question answering system.
Juri Pyykkö, Rebecka Weegar, Pierre Nugues · Lund University Publications (Lund University) · 2014
In this paper, we describe a passage retrieval component for a questioning answering system and we evaluate its performance on Swedish documents. We used a corpus of questions and answers transcribed from the Swedish board game Kvitt eller dubbelt and, as source for the passages, we used the articles of the Swedish version of Wikipedia. We show that Wikipedia is a suitable knowledge source to answer the game questions. For answers consisting of one word, we could extract passages of text that contained the right answer for 91% of the questions and when these answers corresponded to an entity, we found and ranked correct answers for up to 75% of the questions.