Name of participant: Tanise Ceron
Project’s name: MULTIVIEW: Classifying and ranking perspectives in a set of documents
Project description:Recommendation systems nowadays only take into account the user’s history in order to suggest new items. This may encourage the development of the so-called “filter bubbles” online given that users end up reading from similar viewpoints. MULTIVIEW aims to burst this bubble by diversifying perspectives in news recommendation. For that, we will develop methods for modeling not only stance and polarity in texts but also taking aspects of a news topic (e.g. historical, background) in news articles. This will allow, for example, news consumers to read from different viewpoints, and therefore, have a broader understanding of events or current affairs in the world.
Software Campus Partner: University of Stuttgart und Holtzbrinck Publishing Group
Implementation period: 01.03.2024 – 28.02.2026