Enabling a Question-Answering System for COVID Using a Hybrid Approach Based on Wikipedia and Q/A Pairs

Chicaiza, Janneth Bouayad-Agha, NadjetThe research on COVID-19 disease has produced much information, but there are more questions than certainty. This proposal aims to contribute by providing reliable and updated answers to questions aimed at the general public. To achieve this goal, we design a question-answering architecture that leverages two information sources of different nature, controlled-official and open-collaborative. Thus, the system can answer several questions that the community may have about COVID. During the experimentation, we found that thanks to knowledge graphs, information retrieval, and NLP methods, the system can provide explainable answers; i.e., they obtain direct answers and can browse into enriched responses.