Natural language is the primary form of communication between human beings. Apart from spoken language, written language is the most important medium for the documentation and transmission of knowledge. In the age of digitization and the information society, the exchange of knowledge via written or spoken language has increased dramatically. Modern communication technologies allow the almost limitless production, storage and distribution of digital information. In this context, the research fields of Computational Linguistics and Natural Language Processing (NLP) have a long history in studying methods and techniques that enable computers to process and analyze large amounts of natural language data.
The research at our chair focuses on exploiting textual data with natural language processing techniques. We are working in both NLP subfields, namely Natural Language Understanding (NLU) as well as Natural Language Generation (NLG). It is our goal to build interdisciplinary teams with both scientists and practitioners from various industry sectors, including the health, mobility, legal, software engineering and scholarly publishing domain. Our legal tech projects are good examples for these interdisciplinary teams. We involve practitioners from the legal domain in our projects to develop solutions needed by and suited for the legal users. Together with our collaboration partners, we strive to build intelligent NLP solutions which have a real-world impact on users in digitized markets.
Furthermore, our chair also focuses on Transfer Learning aspects of Deep Learning. Training a good model usually requires a lot of data samples and computing power. However, word embedding models, such as transformers, offer generic pre-trained language models that can be later fine-tuned on a specific downstream task. These models are trained on very large datasets and can be reused from any researcher or company for their own problems, without the need for big datasets or large amounts of computing resources.
