Model-driven Development for Adapting Question Answering Systems to Restricted Domains

A Question Answering (QA) system must provide concise answers from large collections of documents to questions stated by the user in natural language. However, although many QA systems for open domain exist, its adaptation for restricted domains (such as those of healthcare, agri culture, transportation, or science) is a far from trivial task. The principal problem is that domain experts ask more precise questions (and expect more precise answers), including specific terminology, and this is costly to integrate into a QA system. To overcome this drawback, this paper presents an innovative approach based on model-driven software development. It uses restricted-domain resources to automatically and effortlessly adapt open-domain QA systems in order to make them useful in restricted-domain scenarios. Finally, a set of experiments has been carried out to show the approach’s applicability.