Welcome to the website of the Computational Linguistics Group (CLAUSE) at Bielefeld University.

Generally, we work on natural language processing and on learning models for language generation & understanding from data. We would like to develop machines that use language as flexibly and smoothly as humans do. This is why we are particularly interested in computational modeling of language use, visual language grounding, reference, pragmatics and dialogue.

From left to right: Clara Lachenmaier, Ronja Utescher, Özge Alaçam, Marc Brinner, Nazia Attari, Simeon Junker, Sina Zarrieß, Henrik Voigt
Not pictured: Sanne Hoeken, Judith Sieker, Bastian Bunzeck

News

• We are hosting Bialogue 2025, the 29th Workshop on the Semantics and Pragmatics of Dialogue, in Bielefeld (September 3–5, 2025)
• Jan ‘25: Bastian presented his paper Small Language Models Also Work With Small Vocabularies: Probing the Linguistic Abilities of Grapheme- and Phoneme-Based Baby Llamas (with Daniel Duran, Leonie Schade and Sina Zarrieß) at COLING 2025 in Abu Dhabi
• Dec ‘24: Bastian’s article The richness of the stimulus: Constructional variation and development in child-directed speech (with Holger Diessel, U of Jena) has been published in First Language
• Nov ‘24: We presented 5 main conference papers and 4 workshop papers at EMNLP in Miami:
  • Main conference: Hateful Word in Context Classification (Sanne, Sina & Özge), Eyes Don’t Lie: Subjective Hate Annotation and Detection with Gaze (Özge, Sanne & Sina), Rationalizing Transformer Predictions via End-To-End Differentiable Self-Training (Marc & Sina), The Illusion of Competence: Evaluating the Effect of Explanations on Users’ Mental Models of Visual Question Answering Systems (Judith, Simeon, Ronja & Sina), Evaluating Diversity in Automatic Poetry Generation (Sina)
  • GenBench: The SlayQA benchmark of social reasoning: testing gender-inclusive generalization with neopronouns (Bastian & Sina)
  • BlackboxNLP: How LLMs Reinforce Political Misinformation: Insights from the Analysis of False Presuppositions (Judith, Clara & Sina, non-archival), Fifty shapes of BLiMP: syntactic learning curves in language models are not uniform, but sometimes unruly (Bastian & Sina, non-archival)
  • BabyLM challenge: Graphemes vs. phonemes: battling it out in character-based language models (Bastian & Sina)
• Simeon & Sina received the best paper award at INLG 2024 for their paper Resilience through Scene Context in Visual Referring Expression Generation
• Oct ‘24: Bastian presented his paper Fifty shapes of BLiMP: syntactic learning curves in language models are not uniform, but sometimes unruly at the MILLing conference in Gothenburg
• Sep ‘24: Bastian presented his ongoing work on Constructions in child-directed speech at the 10th International Conference of the German Cognitive Linguistics Association
• Jun ‘24: We hosted the 3rd annual NLG in the Lowlands workshop
• Mar ‘24: Clara presented her late breaking report Towards Understanding the Entanglement of Human Stereotypes and System Biases in Human–Robot Interaction at the International Conference on Human Robot Interaction (HRI) 2024 in Boulder (Colorado)!
• Dec ‘23: We presented four papers at EMNLP 2023 (and adjacent workshops) in Singapore:
  • Methodological Insights in Detecting Subtle Semantic Shifts with Contextualized and Static Language Models (Sanne & Özge)
  • Towards Detecting Lexical Change of Hate Speech in Historical Data (Sanne, Sina & Özge)
  • When Your Language Model Cannot Even Do Determiners Right: Probing for Anti-Presuppositions and the Maximize Presupposition! Principle (Judith & Sina)
  • GPT-wee: How Small Can a Small Language Model Really Get? (Bastian & Sina)