Welcome to the website of University of Passau’s Natural Language Processing (CAROLL) Group! We are a dynamic research group at the Faculty of Computer Science and Mathematics. We focus on Computational Rhetoric techniques applied to Sentiment Analysis, Offensive language modelling and detection, Legal Tech, Argumentation Mining and other areas of NLP.

To this end, we combine this approach with argument mining, also a research area within NLP. Argumentation mining is already used in social media to identify offensive language. However, the technology has so far reached its limits. She does not recognize irony or unobtrusive insults. The team will test whether the Passau approach is capable of this in heated debates on the topics of the climate crisis and vaccination criticism.

We are also collaborating with our fellow researchers in Media Bias Group which is a research group that is focusing on uncovering media bias or unbalanced coverage.

The project on which this research group is based is funded by the German Federal Ministry of Education and Research (BMBF) along with the University of Passau.


06. July 2021

Dr. Jelena Mitrovic gave a talk about Neural Language Models for Abusive Language Detection at The AI Seminar organized by the Mathematical Institure of the Serbian Academy of Sciences and Arts.

22. June 2021

Dr. Jelena Mitrovic gave a Keynote Speech at The French-German Summerschool on AI with Industry 2021 organized by the University of Passau, Ecole Normale Supérieure Paris-Saclay and with the support from Siemens AI Lab.

21. June 2021

Rahmona Kühn and Tobias Milz submitted two posters on their research, Fantastic Figures and How To Find Them and Analysis of a German Legal Citation Network, in the context of the 2021 French-German Summer-school on Artificial Intelligence with Industry.

28. April 2021

Dr. Jelena Mitrović has participated in 10 Minuten Soziologie zum Thema Stress, 10 Minuten Rechtswissenschaft zum Thema Digitalisierung. For the recorded session, use this link.

19. March 2021

Our Paper Automated identification of bias inducing words in news articles using linguistic and context-oriented features was accepted for publication at the Information Processing and Management Journal.

... see all News