We continue our mini-series on SUMMETIX being used in scientific work. As a spin-out from TU Darmstadt, SUMMETIX is committed to supporting cutting-edge AI and LLM research.
This week, as part of the world-leading international NLP conference ACL (63rd Annual Meeting of the Association for Computational Linguistics) in Vienna, the 12th Workshop on Argument Mining takes place. One of the papers presented is by our long-time collaboration partner, the Dialogue Systems Group at Ulm University, together with Tensor AI Solutions GmbH.
The paper is named “Automatic Identification and Naming of Overlapping and Topic-specific Argumentation Frames” and is motivated by the use of frames to mitigate biases and filter bubbles in argumentative dialogue. Identifying frames significantly contributes to discourse understanding and can be used to reduce toxic and harmful discussions on the web and elsewhere. The authors of the paper show that a combination of aspect term extraction and weighting with c-TF-IDF can reliably describe frames in arguments “from the wild”. They use our Aspect Detection API (api.summetix.com) as part of the frame naming process.
