Name of the participant: Arne Binder
Description of the IT research project: Every year, millions of new scientific articles are published worldwide, and the trend is rising [1]. Researchers and scientific reviewers increasingly find themselves in the situation that the quantity of publications, even in their respective fields of specialisation, is no longer sufficiently easy to grasp. Automated feeds or recommender systems, such as those from Arxiv Sanity or Semantic Schoolar, promise to remedy this situation, but usually work with simple procedures that do not do justice to the complex nature of scientific texts. Most importantly, they do not provide information on why a proposed publication is considered relevant, what the important arguments are and how they relate to other statements. Argumentative structure is not taken into account by these tools, although arguments are the relevant basic elements in scientific discourse.
Argument Mining (AM) refers to the automated analysis of the argumentative structure of texts. In this project, machine learning-based AM techniques are explored for application in the scientific domain, with the ultimate goal of developing prototypes that make these structures visible and usable for the user. The automated identification of claims and their argumentative dependencies has the potential to pre-process large volumes of publications in order to then efficiently address practical questions such as: Which assertions are made in a publication under consideration? Which ones support them, which ones attack them? In which publications is a particular claim shown or refuted? Which similar claims have already been made in other publications? The aim is to make access to relevant knowledge much more efficient by answering these and similar questions, and ultimately to increase the transparency of scientific discourse.
Software Campus partners: DFKI, Holtzbrinck Publishing Group
Implementation period: 01.03.2021 – 28.02.2023
[1] cf. https://ncses.nsf.gov/pubs/nsb20206/executive-summary.
