The Text-to-Knowledge Group (T2K) conducts research in Natural Language Processing (NLP),1 ranging from classical machine learning based text enrichment systems to deep learning based models. A central theme of our research is the development of creative new algorithms for processing text to solve important problems in human language technology, with a link to industry through collaborative projects.
User-generated textual information is the most commonly used type of data on the web but it is highly unstructured, as natural language text. NLP systems discover structured information from such text that enables much richer querying and data mining (e.g., in news articles, based on automatically extracted named entities, relations, dates). With our T2K team, we have a strong track record in this field of information extraction (in domains including news, human resources, biomedical), as well as text classification tasks (e.g., sentiment). More recently, we have also worked on conversational agents and generative models, e.g., for educational applications.
Follow us on the Ghent University NLP page on X. While you're at it, you may also want to follow our colleagues from the Language and Translation Technology Team (LT3), also working on NLP.
1: Note that our research group moved towards NLP from the information retrieval domain, which we initially worked on almost a decade ago.