Neuro-Symbolic Models for Ensuring Cybersecurity in Critical Cyber-Physical Systems

Abstract

This paper presents the results of a comprehensive study on the application of the neuro-symbolic approach for detecting and preventing cyber threats in railway systems, a critical component of cyber-physical infrastructures. The increasing complexity and integration of physical systems with digital technologies have made such infrastructures vulnerable to cyberattacks, where breaches can result in severe consequences, including system failures, financial losses, and threats to public safety and the environment. The objective of this study was to assess the effectiveness of the neuro-symbolic approach, which combines artificial neural networks with symbolic algorithms, in detecting and mitigating cyber threats in dynamic environments. The methodology involved simulating various cyberattack scenarios on a test architecture for railway system security, followed by applying the neuro-symbolic model for threat detection and response. Results showed that the neurosymbolic approach demonstrated high accuracy in detecting cyber threats and was particularly effective in adapting to new and unknown types of attacks. Compared to traditional methods, this approach significantly improved detection efficiency and response speed. The findings confirm that the neuro-symbolic approach enhances cybersecurity, particularly in critical infrastructures like railway systems, and contributes to more reliable protection of data related to passengers and transported goods. Further research will focus on optimizing the implementation of these algorithms and expanding the range of practical applications to other critical sectors.