NLP

DITLab focuses on text-based methods, data-intensive research (collection, processing and computational analysis of large datasets) and the long-term development of methodological expertise and teaching. We analyse large bodies of text using corpus-based methods and Natural Language Processing (NLP), bringing together expertise from different fields in areas such as text mining and topic modelling. We also focus on the use, application and development of Large Language Models (LLMs) and deep learning algorithms (neural networks) in research and teaching, as well as data visualization and network analysis.

TurkuNLP is a leading Finnish research group in natural language processing (NLP) and digital language studies. Known for its strong interdisciplinary focus ranging from digital humanities to BioNLP and large language models, the group develops advanced language technology tools and resources, and runs a number of projects targeting language use in digital collections. TurkuNLP is internationally recognized for its work on dependency parsing, with widely used tools such as the Turku Neural Parser Pipeline and contributions to Universal Dependencies. Further, TurkuNLP works on the development of large language models such as FinGPT and Viking, and pioneers in structuring web corpora through web register (genre) identification.