Welcome to IDM 4: Data Science, a unit within the information and data management (IDM) department at the Stabi Berlin.

We conduct research and development for the digitisation, analysis and curation of our collections as data, with a focus on machine learning and artificial intelligence. We are active in the Digital Humanities and collaborate closely with the Stabi Lab.

A main focus of our work lies in computer vision tasks like document layout analysis and OCR/HTR for historical documents. But we also work with language models for named entity recognition and linking, or with multilingual models for image classification and image similarity search.

Another important aspect of our work is the curation of digitised cultural heritage and its publication as data for research and artificial intelligence, under consideration of ethical, legal and social issues, and within the context of the surrounding policy frameworks. We actively participate in the relevant task forces and working groups of e.g. Europeana or AI4LAM and technical committees for standards such as ALTO or the DFG Practical Guidelines.

We are experienced in the planning, agile management and successful implementation of research projects at various levels of scale and funding, with national and international partners. Some of our past projects include OCR-D (funded by DFG), Qurator (funded by BMBF) and Mensch.Maschine.Kultur (funded by BKM).

We share all outputs openly. Have a look at the resources below or contact us to find out more.