You are here
Consumer Health Question Answering to Automatically Support NLM Customer Services
In 2012, given the growth of customer questions and advances in natural language processing research, the Director of the NLM, Dr. Donald A.B. Lindberg, launched a project to build a system that would assist the NLM staff in responding to customer requests. Such a tool would classify customer requests and automatically direct them to the appropriate area of the library so that they could be answered in a more efficient manner. As a research and development division of NLM, the Lister Hill National Center for Biomedical Communications (LHNCBC) was tasked with the development of such a tool. This initiative resulted in development of the Consumer Health Information Question Answering (CHIQA) system.
In this report, we briefly present the overall architecture of the CHIQA system and the results of its integration with the customer support services. We then discuss in depth our research on various aspects of automated question understanding and answering. This includes co-reference resolution, spelling correction, and question decomposition, as well as the answer generation methods needed for a consumer health question answering system.