Query-Based Summarization using Rhetorical Structure Theory

TL;DR: Existing question answering systems—which aim at finding precise answers to questions—can be improved by exploiting summarization techniques to extract more than just the answer from the document in which the answer resides.

Abstract: Research on Question Answering is focused mainly on classifying the question type and finding the answer. Presenting the answer in a way that suits the user’s needs has received little attention. This paper shows how existing question answering systems—which aim at finding precise answers to questions—can be improved by exploiting summarization techniques to extract more than just the answer from the document in which the answer resides. This is done using a graph search algorithm which searches for relevant sentences in the discourse structure, which is represented as a graph. The Rhetorical Structure Theory (RST) is used to create a graph representation of a text document. The output is an extensive answer, which not only answers the question, but also gives the user an opportunity to assess the accuracy of the answer (is this what I am looking for?), and to find additional information that is related to the question, and which may satisfy an information need. This has been implemented in a working multimodal question answering system where it operates with two independently developed question answering modules.