V. Punyakanok, D. Roth, and W. Yih
6 January 2004
We describe an approach for answer selection in a free form question answering task. In order to go beyond the keyword based matching in selecting answers to questions, one would like to incorporate both syntactic and semantic information in the question answering process. We achieve this goal by representing both questions and candidate passages using dependency trees, and incorporating semantic information such as named entities in this representation. The sentence that best answers a question is determined to be the one that minimizes the generalized edit distance between it and the question tree, computed via an approximate tree matching algorithm. We evaluate the approach on question-answer pairs taken from previous TREC Q/A competitions. Preliminary experiments show its potential by significantly outperforming common bag-of-word scoring methods.
In Proceedings of AI&Math 2004 (Special session: Intelligent Text Processing)