Antonios Gouglidis

Ebtisaam Alharbi, Leandro Soriano Marcolino, Qiang Ni et al.

Federated Learning (FL) enables collaborative model training across multiple devices while preserving data privacy. However, it remains susceptible to backdoor attacks, where malicious participants can compromise the global model. Existing defence methods are limited by strict assumptions on data heterogeneity (Non-Independent and Identically Distributed data) and the proportion of malicious clients, reducing their practicality and effectiveness. To overcome these limitations, we propose Robust Knowledge Distillation (RKD), a novel defence mechanism that enhances model integrity without relying on restrictive assumptions. RKD integrates clustering and model selection techniques to identify and filter out malicious updates, forming a reliable ensemble of models. It then employs knowledge distillation to transfer the collective insights from this ensemble to a global model. Extensive evaluations demonstrate that RKD effectively mitigates backdoor threats while maintaining high model performance, outperforming current state-of-the-art defence methods across various scenarios.

Igor Ivkic, Patrizia Sailer, Antonios Gouglidis et al.

Cyber-Physical Systems (CPS) are formed through interconnected components capable of computation, communication, sensing and changing the physical world. The development of these systems poses a significant challenge since they have to be designed in a way to ensure cyber-security without impacting their performance. This article presents the Security Cost Modelling Framework (SCMF) and shows supported by an experimental study how it can be used to measure, normalise and aggregate the overall performance of a CPS. Unlike previous studies, our approach uses different metrics to measure the overall performance of a CPS and provides a methodology for normalising the measurement results of different units to a common Cost Unit. Moreover, we show how the Security Costs can be extracted from the overall performance measurements which allows to quantify the overhead imposed by performing security-related tasks. Furthermore, we describe the architecture of our experimental testbed and demonstrate the applicability of SCMF in an experimental study. Our results show that measuring the overall performance and extracting the security costs using SCMF can serve as basis to redesign interactions to achieve the same overall goal at less costs.

Antonios Gouglidis, Christos Grompanopoulos, Anastasia Mavridou

Usage control models provide an integration of access control, digital rights, and trust management. To achieve this integration, usage control models support additional concepts such as attribute mutability and continuity of decision. However, these concepts may introduce an additional level of complexity to the underlying model, rendering its definition a cumbersome and prone to errors process. Applying a formal verification technique allows for a rigorous analysis of the interactions amongst the components, and thus for formal guarantees in respect of the correctness of a model. In this paper, we elaborate on a case study, where we express the high-level functional model of the UseCON usage control model in the TLA+ formal specification language, and verify its correctness for <=12 uses in both of its supporting authorisation models.