MaltQA : an open-domain question answering system

Abstract

This dissertation considers a surface-based question answering system for an open-domain solution. Its main purpose is to provide a question answering system for education purposes but the solution proposed has wider applications. This document analyzes the current progress that has been done in this area so far, and describes a methodology for answering questions by using infom1ation retrieved from very large collections of text. The solution proposed is based on indexing techniques and surface based natural language processing that identify paragraphs from which an answer can be extracted. The MaltQA system that is presented in this document contains some innovative ideas in improving the accuracy of QA systems. These include an FAQ module that is able to identify similar questions that have been asked before, an acronym dictionary, a question classifier module that is based on the Winnow algorithm and an answer extraction algorithm that achieves a good accuracy when applied both on locally indexed and web documents. The approach presented here will not solve all the problems associated with question answering, but has the objective to provide a solution that is feasible, achieves reasonable accuracy and can return an answer in an acceptable time limit. Various techniques are discussed including question analysis, question reformulation, term extraction, answer extraction and other methods for answer pinpointing. The lessons learned during the development of MaltQA are presented and future enhancements are identified on how to improve the system even further.