Certification and Foundations of SafeMachine Learning Systems in Healthcare

Modern Deep Learning systems in healthcare have the potential to make diagnostic decisions of similar quality to attending doctors. However, there are concerns about their transparency, robustness, fairness, and reliability. The goal of this project is to address these issues by exploring the interdependencies and potential trade-offs of different aspects (fairness, accuracy, interpretability, and privacy) as well as their ethical implications in concrete healthcare applications. The aim is to establish guidelines for the certification of such systems and thus enable their safe use in medicine.