Lisa Beinborn

Lisa Beinborn is a Professor for Human-Centered Data Science at the University of Göttingen, leading the Human-Centered Data Science group. Her research bridges natural language processing with cognitive science, focusing on multilingual models and interpretability.

• PhD in Computer Science (2016), Technische Universität Darmstadt
• MSc in Computational Linguistics (2010), Saarland University & Bolzano, Italy
• BSc in Computational Linguistics (2008), Saarland University & Barcelona, Spain

Her research explores cognitive plausibility in NLP, analyzing how language models process language differently from humans. Key areas include multilingual model interpretability, semantic drift, eye-tracking, and readability prediction.

Recent work examines input representation stability in neural models, cross-lingual transfer of complexity, and aligning language models with human cognitive patterns. Her team has presented findings at EMNLP, CoNLL, ACL, and CoLING.

• VENI Grant for "Interpretability of Transfer in Multilingual Models"
• Early Career Partnership by Royal Dutch Academy of Science
• "Most Interesting Paper" Award at BabyLM Challenge
• "Best Project Award" by Network Institute

She has taught courses like Language as Data and Advanced NLP at University of Göttingen, VU Amsterdam, and TU Darmstadt. Her group collaborates with institutions like Gemeente Amsterdam and NT2 on multilingual text simplification and learner correction.