The mathematicians at the Center for Technomathematics (ZeTeM) at the University of Bremen are on the trail of the secrets of Artificial Intelligence (AI). It is now widely known that deep learning and other AI processes achieve amazing results for a multitude of applications. But why do they work and how can the structures of trained neural networks be translated back into scientifically and mathematically formulated findings? ZeTeM deals with these questions and is funded with € 2 million by the Klaus Tschira Foundation as part of the BMBF funding measures “Mathematics for Innovations” and “Computational Life Sciences”.

In spring 2020, new projects will start in the Technomathematics group of Prof. Dr. Dr. hc Peter Maaß, who in cooperation with the industrial partners EWE, Siemens, Engineering System International, Deutsche Bahn, Bruker Daltonik, Proteopath, ProCon X-Ray, atacama blooms and Volkswagen combine basic research on AI with industrial applicability. A key point of research are invertible network architectures that allow the results to be interpreted directly in the respective specialist discipline. Now, however, the employees at ZeTeM have a particular problem: There are not enough qualified candidates for research in this area. In the medium term, the planned realignment of the technical mathematics course towards industrial mathematics and data analysis should help.