FactRank began as a term project of Brecht Laperre, Ivo Merchiers and Rafael Hautekiet, in the course Knowledge and the Web offered by Prof. Dr. Bettina Berendt at the Department of Computer Science at the University of Leuven. Their goal was to develop a tool that would help the fight against fake news and the spread of misinformation.
As a first step, they developed a machine learning model that was able to automatically detect check-worthy factual statements. If such claims can be reliably identified, they can be forwarded to a fact-checking process for verification. The model was trained on 1800 manually classified statements.
This yielded a working prototype, which was improved during the course of 2019. The tool is now more accurate and draws on a larger dataset of ~7000 labeled statements. This was made possible by a grant of € 39.869 that was awarded to FactRank by the Vlaams Journalistiek Fonds in December 2018.
This new phase of the project, which was finalised in March 2020, is the result of Flemish-Dutch teamwork. Led by freelance journalist Dr. Jan Jagers (fact-checker of Knack Magazine and lecturer in journalism at Vrije Universiteit Brussel) the team also comprised Dr. Peter Burger and Dr. Alexander Pleijter (Nieuwscheckers.nl, University of Leiden), Prof. Dr. Peter Van Aelst (University of Antwerp), Prof. Dr. Bettina Berendt (KU Leuven and TU Berlin, Germany), and machine learning engineer Rafael Hautekiet.
Since 2021 Textgain has taken over development of FactRank. They maintain the scrapers and the database of statements in the context of the BENEDMO project.
Links
