Today we announce two BSc theses produced by Bosch’s students in the context of our MEDINA project. We refer to “Modell-Diebstahl für Zeitreihenprognosen in Bezug auf ein Cybersecurity Governance Framework für Künstliche Intelligenz” (Acker V., Hochschule Albstadt-Sigmaringen, Germany, Aug-2022) and “Framework für Cybersecurity-Metriken für die Einhaltung von vorschriften” (Habeck T. O., Hochschule der Medien, Germany, Apr-2023).
In his thesis, Mr. Acker explores the idea of assessing the cybersecurity of Artificial Intelligence models in order to detect vulnerabilities which are specific to this technology. The resulting assessment methodology (applicable to so-called time-series AI models) can become an AI-specific Evidence Collector in the MEDINA framework, which is the basis for the upcoming Horizon Europe COBALT project (project number 101119602).
Related to the research work from Mr. Habeck, we find the proposal to develop a metrics management framework for Bosch which can assess the compliance level of our internal IT systems (cloud and on-premises). His work is a direct extrapolation of MEDINA framework in an industrial context, where real-world experiences are leveraged. The resulting framework (consisting of processes and proposed tools), is currently being analyzed by our operational teams at Bosch to assess its feasibility of implementation.
Both BSc works are directly related to Bosch’s exploitation and sustainability plans for MEDINA.