LASIGE’s Andre Lamurias, Diana Sousa and Francisco M. Couto publish at IEEE Access

Date: 04/09/2020

Biomedical questions are a challenge to most automatic Question-Answering approaches, due to the complexity of the domain and limited availability of training sets. In the paper “Generating Biomedical Question Answering Corpora from Q&A Forums” the researchers Andre Lamurias, Diana Sousa and Francisco M. Couto present a method to automatically extract question-article pairs from Q&A web forums, that can be used to train a deep learning QA system. The proposed framework extracts from selected forums questions and the respective answers that contain citations and can be used to update BiQA corpus from the same forums as new posts are made, and from other forums that support their answers with documents. The paper was published at IEEE Access, a top-ranked journal (h5-index=119).The BiQA corpus and the framework used to generate it are available here.