OxBioNLP: bias-free systematic literature reviews and evidence discovery for COVID-19 research
COVID-19 research is being rapidly produced, but the volume and lack of structure makes this research difficult to navigate. The COVID-19 Open Research Dataset (CORD-19), for example, is a resource of over 400,000 scholarly articles, including over 150,000 with full text. Meetings with lead scientists in the fields of research methods, systematic reviews, infection diseases and medical virology enabled the team to identify the key 'pain points' that scientists need to resolve in their active research of systematic reviews on COVID-19. OxBioNLP-v1.0 addresses these issues by curating the latest data on COVID-19 research, extracting data in real time from world-leading open data lakes, and processing the title, abstract and full-text of the articles using state-of-the-art NLP approaches. It stores the processed data after efficient indexing, provides a bespoke clustering-based search and knowledge discovery features, and displays data using advanced data visualisation techniques.
PIs: Thomas Lukasiewicz, Omer Gunes