Cyber-Physical System Security Enhanced with Kolmogorov-Arnold Network

Strategic Insights -

Researchers at K.N. Toosi University of Technology have developed a lightweight and efficient solution for detecting cyber attacks in critical infrastructure systems. By employing the Kolmogorov-Arnold network (KAN), the team has achieved high classification accuracy while minimizing computational overhead, making it a practical solution for real-time CPS security. This breakthrough has significant implications for the protection of industrial systems from potential threats. The research has been peer-reviewed and published in the International Journal of Critical Infrastructure Protection.

Key Takeaways:

  • The Kolmogorov-Arnold network (KAN) is a lightweight and efficient alternative to conventional models for attack detection in cyber-physical systems (CPSs).
  • KAN achieves high classification accuracy while minimizing computational overhead, making it a practical solution for real-time CPS security.
  • The research eliminates the need for complex feature extraction and preprocessing, preserving data integrity and enabling faster decision-making.
  • The KAN model is evaluated on three datasets: SWaT, WADI, and ICS-Flow, demonstrating superior performance in detecting cyber attacks across binary and multi-class tasks.
  • The KAN model reduces the need for extensive data preprocessing, which can introduce drawbacks such as loss of critical information, reduced interpretability, and increased latency.
  • The research has been peer-reviewed and published in the International Journal of Critical Infrastructure Protection.

Statistics:

  • The Kolmogorov-Arnold network (KAN) achieves high classification accuracy, with performance evaluated on three datasets: SWaT, WADI, and ICS-Flow.
  • The KAN model demonstrates superior performance in detecting cyber attacks across binary and multi-class tasks, with results indicating a significant reduction in computational overhead.
  • The research eliminates the need for complex feature extraction and preprocessing, reducing the risk of data loss and enabling faster decision-making.

Sources:

  • NewsRx. Investigators from K.N. Toosi University of Technology Target Engineering (Using Kolmogorov-arnold Network for Cyber-physical System Security: a Fast and Efficient Approach). Journal of Engineering. September 1, 2025; p 1299.
  • International Journal of Critical Infrastructure Protection. Using Kolmogorov-arnold Network for Cyber-physical System Security: a Fast and Efficient Approach. International Journal of Critical Infrastructure Protection, 2025;50.
  • Elsevier. International Journal of Critical Infrastructure Protection. www.journals.elsevier.com/international-journal-of-critical-infrastructure-protection/