AI Assisted Certification of Medical Software
- August 1, 2022 – July 31, 2025
- funded by Federal Ministry of Education and Research (BMBF)
Safety certificates for medical software are so far unstructured and often spread over several documents. This hinders the efficient processing of these documents in the formal certification of medical products relying on software. An automation of the creation and the approval of safety certificates may lead to a significant speed-up in certification of medical products.
The goal of KIMEDS is the establishment of the method of the Assurance Case as an ontology-based safety documentation for automated verification and validation of risk-free medical software components. An assisting and explainable system is based on the regulatory foundations and enacts as a validating instance of the Assurance Case over the lifecycle of medical software. The creation of the safety documentation begins in the early stages of software development, which is then guided by a continuous check of validity through the system we develop. This system also supports the certification of Assurance Cases, leading to a significant increase in efficiency of the overall certification process.
Tractable Diversity: Scalable Multiperspective Ontology Management via Standpoint EL
Proceedings of the 32nd International Joint Conference on Artificial Intelligence, IJCAI 2023, to appear
Efficient Dependency Analysis for Existential Rules
Proceedings of the 15th Alberto Meldenzon International Workshop on Foundations of Data Management (AMW'23). Santiago, Chile, May 2023
An Existential Rule Framework for Computing Why-Provenance On-Demand for Datalog
Proceddings of the 6th International Joint Conference on Rules and Reasoning (RuleML+RR 2022), to appear. Springer