FACO: Fuzzy Ant Colony Optimization for Attack Detection in Smart Water Security using Few-Shot Learning

Speakman, Lee; Gaber, Tarek; Hamed, Ahmed; Nicho, Mathew; Galeela, Mohamed

Abstract

In the realm of smart water utilities, the increasing sophistication of cyber threats presents a significant challenge to the security and operational integrity of water and wastewater treatment facilities. These facilities, heavily dependent on Industrial Control Systems (ICS), are vulnerable to cyberattacks, potentially resulting in severe consequences for public health and safety. This paper introduces an innovative approach that integrates Fuzzy Logic, Ant Colony Optimization (ACO), and Few-Shot Learning to enhance the detection and response to cyber threats in smart water utilities. A novel version of the ACO algorithm, called FACO (Fuzzy ACO), is proposed to optimize the rule base of the Fuzzy Logic system, improving efficiency and reducing computational overhead. Few-Shot Learning is employed to address the challenge of detecting novel attack vectors with limited historical data, a common constraint in cybersecurity datasets. By combining these three techniques, the results demonstrate that our attack detection model is accurate, achieving a 95% accuracy rate, and computationally efficient with a reduction rate of 60%. The evaluation, including statistical analysis, further indicates the superiority of our approach over traditional methods in terms of detection accuracy and computational time. Thus, it is recommended that our proposed model be considered as a valuable tool in the arsenal against cyber threats in smart water utilities.

• Presentation Conference Type: Conference Paper (published)
• Conference Name: International Joint Conference on Neural Networks (IJCNN)
• Start Date: Jun 30, 2024
• End Date: Jul 5, 2024
• Acceptance Date: Apr 1, 2024
• Online Publication Date: Jun 30, 2024
• Publication Date: Jun 30, 2024
• Deposit Date: Mar 17, 2025
• Peer Reviewed: Peer Reviewed
• ISBN: 9798350359312
• Keywords: Small data, few-shot learning, Industrial Control Systems, Fuzzy Logic, Ant Colony Optimization