On the Scalability of the Answer Extraction System "ExtrAns"

Abstract

This paper reports on the scalability of the answer extraction system ExtrAns. An answer extraction system locates the exact phrases in the documents that contain the explicit answers to the user queries. Answer extraction systems are therefore more convenient than document retrieval systems in situations where the user wants to find specific information in limited time. ExtrAns performs answer extraction over UNIX manpages. It has been constructed by combining available linguistic resources and implementing only a few modules from scratch. A resolution procedure between the minimal logical form of the user query and the minimal logical forms of the manpage sentences finds the answers to the queries. These answers are displayed to the user, together with pointers to the respective manpages, and the exact phrases that contribute to the answer are highlighted. This paper shows that the increase in response times is not a big issue when scaling the system up from 30 to 500 documents, and that the response times for 500 documents are still acceptable for a real-time answer extraction system.