Chair for Natural Language Processing

The Chair for Natural Language Processing focuses on enabling machines to comprehend human languages, particularly written text. Their research emphasizes deep learning and representation learning techniques to precisely model the meaning of natural language across various languages. They prioritize multilingual representation learning and the transfer of models for different NLP tasks. While recent advancements in deep learning have propelled NLP forward, they recognize that such progress often excludes speakers of low-resource languages and those lacking computational resources. Additionally, the environmental impact of training large neural models and the perpetuation of societal biases are concerns. The group at WüNLP addresses these challenges by striving for equitable, socially responsible, and sustainable language technology through sustainable, modular, and sample-efficient models, fair and unbiased NLP practices, and multilingual NLP, particularly targeting low-resource languages.

chair website