With the help of AI and deep learning methods, medical image data, biosignals and other patient data are analyzed and automatically interpreted, e.g. to automatically delimit pathologically changed image areas, detect diseases and predict patient-specific disease progression. In particular, aspects important for medicine, such as the optimization of neuronal networks with highly limited training quantities, as well as the evaluation of the safety and explainability of the results are examined. The developed methods and systems are developed and evaluated in close cooperation with medical cooperation partners on the basis of practice-relevant questions.