Challenges for Model-Based Diagnosis

Since the seminal works by Reiter and de Kleer and Williams published in the late 80's, Model-based Diagnosis has been a significant area of research. This has been motivated by the fact that MBD assists us in tackling a challenge that we face almost on a daily basis, i.e., by MBD allowing us to reason in a structured manner about the root causes for some encountered problem. MBD achieves this in an intuitive, complete and sound way, based on the central idea of investigating the compliance of some observed behavior with a model that describes how a system should behave - given this or that input scenario and parameter set. Over the last 40 years, MBD has been adopted for a multitude of applications, and we saw the emergence of a diverse set of algorithmic, optimizations, as well as extensions to the initial theoretical concepts.We argue that MBD remains highly relevant, with numerous scientific challenges to tackle as we face increasingly complex diagnostic problems. We discuss several such challenges and suggest related topics for PhD theses that have the potential to significantly contribute to the state-of-the-art in MBD research.