Confirmed Knowledge Acquisition Using Mails Posted to a Mailing List

Confirmed Knowledge Acquisition Using Mails Posted to a Mailing List Yasuhiko Watanabe, Ryo Nishimura, and Yoshihiro Okada Ryukoku University, Seta, Otsu, Shiga, 520-2194, Japan watanabe@rins.ryukoku.ac.jp Abstract. In this paper, we first discuss a problem of developing a knowledge base by using natural language documents: wrong informa- tion in natural language documents. It is almost inevitable that natural language documents, especially web documents, contain wrong informa- tion. As a result, it is important to investigate a method of detecting and correcting wrong information in natural language documents when we develop a knowledge base by using them. In this paper, we report a method of detecting wrong information in mails posted to a mailing list and developing a knowledge base by using these mails. Then, we describe a QA system which can answer how type questions based on the knowl- edge base and show that question and answer mails posted to a mailing list can be used as a knowledge base for a QA system. 1 Introduction Because of the improvement of NLP, research activities which utilize natural language documents as a knowledge base become popular, such as QA track on TREC [1] and NTCIR [2]. However, these QA systems assumed the user model where the user asks what type questions. On the contrary, there are a few QA systems which assumed the user model where the user asks how type question, in other words, how to do something and how to cope with some problem [3] [4] [7]. There are several difficulties in developing a QA system which answers how type questions, and we focus attention to two problems. First problem is the difficulty of extracting evidential sentences by which the QA system answers how type questions. It is not difficult to extract evidential sentences by which the QA system answers what type questions. For example, question (Q1) is a what type question and “Naoko Takahashi, a marathon runner, won the gold medal at the Sydney Olympics” is a good evidential sentence for answering question (Q1). (Q1) Who won the gold medal in women’s marathon at the Sydney Olympics? (DA1–1) Naoko Takahashi. It is not difficult to extract this evidential sentence from natural language doc- uments by using common content words and phrases because this sentence and question (Q1) have several common content words and phrases. On the contrary, it is difficult to extract evidential sentences for answering how type questions R. Dale et al. (Eds.): IJCNLP 2005, LNAI 3651, pp. 131–142, 2005. c ⃝ Springer-Verlag Berlin Heidelberg 2005 132 Y. Watanabe, R. Nishimura, and Y. Okada only by using linguistic clues, such as, common content words and phrases. For example, it is difficult to extract evidential sentences for answering how type question (Q2) because there may be only a few common content words and phrases between the evidential sentences and question (Q2). (Q2) How can I cure myself of allergy? (DA2–1) You had better live in a wooden floor. (O2–1–1) Keep it clean. (O2–1–2) Your room is always dirty. (DA2–2) …

Content truncated to 500 words