Question-Answering via Enhanced Understanding of Questions

We describe a machine learning centered approach to developing an open domain question answering system. The system was developed in the sum- mer of 2002, building upon several existing machine learning based NLP modules developed within a unified framework. Both queries and data were pre-processed and aug- mented with pos tagging, shallow parsing informa- tion, and some level of semantic categorization (be- yond named entity) using a SNoW based machine learning approach. Given these as input, the sys- tem proceeds as an incremental constraint satisfac- tion process. A machine learning based question analysis module extracts structural and semantic constraints on the answer, including a fine classi- fication of the desired answer type. The system continues in several steps to identify candidate pas- sages and then extracts an answer that best satis- fies the constraints. With the available machine learning technologies, the system was developed in six weeks with the goal of identifying some of the key research issues of QA and challenges to it.