Testing an Entity Ranking Function for English Factoid QA

Much progress has been made in the English question-answering task since it was initiated in TREC-8 and our last participation in TREC-2001. QA is a complex semantics-oriented task, and it is necessary that much linguistic processing, auxiliary resources, and learning steps are needed to come up with adequate performance to the task [1]. Chinese language factoid QA was first introduced during NTCIR-5 and -6 [2,3] and in which we participated. Because Chinese QA was new with little training data and Chinese linguistic tools and resources were not as readily available as in English, we employed a simple answer ranking technique that depends only on surface term usage statistics in questions and document sentences [4,5]. The outcome has been surprisingly satisfactory, achieving results ~80% of the best [2,3]. We were curious how this approach may perform for English factoid questions, comparing with median result for example. The program was modified to support English in this TREC-2007 QA task. As in Chinese, we made little use of linguistic tools or training data, and no auxiliary resources. It can form a basis result from which more sophisticated tools may enhance.