STAI CDT student Gabriel Freedman presented a paper at the 11th Workshop on Argument Mining (ArgMining 2024), co-located with the 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) last year. The paper, ‘Detecting Scientific Fraud Using Argument Mining‘, was nominated for best paper at the workshop.
The paper introduces a new way of trying to detect fake academic papers that have somehow managed to be published in journals.
As Gabriel explains, “This is a growing problem and can have a lot of negative consequences – for instance healthcare providers may base their treatments on findings in academic papers. The method we developed looks at the quality of the arguments contained within these papers, using machine learning models to output a rating for how good or bad the arguments are. Our thinking behind it was that fraudulent papers are likely to be based on poor quality arguments, whereas real papers are more likely to contain coherent and well-structured arguments. Luckily, we found that this is in fact the case for quite a lot of papers, however it is not a perfect solution and there is still a lot of work to be done on this very important and challenging problem”.
Gabriel really enjoyed the opportunity to gain exposure to other researchers and their work. Attending the workshop also allowed Gabriel to get an in depth feel for the field of Argument Mining. A highlight for Gabriel was also the opportunity to explore the location of the conference, Thailand. Gabriel said, “Thailand is an amazing and vibrant country, and I was lucky enough to see some of the islands before I left”.
He also shared that, “While I really enjoy the everyday experience of doing research for my PhD, it was a great motivation to be able to attend the conference and share my work with other people. It adds a very human element to the field of AI, which I think we all can agree is a good thing!”.
The paper was co-authored with Francesca Toni, and you can read it here: Detecting Scientific Fraud Using Argument Mining – ACL Anthology