Exploring Syntactic Relation Patterns for Question Answering

Abstract. In this paper, we explore the syntactic relation patterns for opendomain factoid question answering. We propose a pattern extraction method to extract the various relations between the proper answers and different types of question words, including target words, head words, subject words and verbs, from syntactic trees. We further propose a QA-specific tree kernel to partially match the syntactic relation patterns. It makes the more tolerant matching between two patterns and helps to solve the data sparseness problem. Lastly, we incorporatethe patterns into a Maximum Entropy Model to rank the answer candidates. The experiment on TREC questions shows that the syntactic relation patterns help to improve the performance by 6.91 MRR based on the common features.

• Type: Conference
• Year: 2005
• Where: IJCNLP
• Authors: Dan Shen, Geert-Jan M. Kruijff, Dietrich Klakow
• Keywords: IJCNLP 2005, Natural Language Processing, Opendomain Factoid Question, Syntactic Relation Patterns, Syntactic Trees