This project aims the development of methods for the field of Question Answering (QA) and the implementation of a system that will employ them. Natural Language Processing techniques are intended to be explored in order to perform transformations in small portions of text, which should make them easier for processing, while keeping their original meaning. Thus, a QA system can have more flexibility when dealing with questions and text bases that talk about the same subjects, but written differently. In particular, the system is intended to be directed to the domain of chemical and contamination risk in the milk production chain, because this project is carried out in partnership with a research network about the domain and which involves other institutions.
News published in Agência FAPESP Newsletter about the scholarship: