Assessing Machine Learning Enabled Anomaly Detection Models for Real Time Cyberattack Mitigation in Optical Fiber Communication Systems.

The increasing complexity and data throughput of optical fiber communication systems have made them critical yet vulnerable components of modern digital infrastructure. With the rapid growth of high-speed networks, ensuring cybersecurity in these systems requires intelligent, adaptive, and real-time mitigation strategies. This review examines the application of machine learning (ML)-enabled anomaly detection models for identifying and mitigating cyberattacks in optical fiber communication environments. It highlights how supervised, unsupervised, and reinforcement learning algorithms—such as Support Vector Machines (SVM), Random Forests, Deep Neural Networks (DNN), and Autoencoders—enable real-time detection of network anomalies, signal disruptions, and malicious intrusions. Furthermore, the paper explores the integration of hybrid ML frameworks combining statistical signal processing with deep learning for enhanced detection accuracy and low false alarm rates. Special emphasis is placed on the challenges of model interpretability, scalability, and latency in large-scale fiber networks, alongside the role of edge computing and federated learning in decentralized security monitoring. The study also evaluates emerging trends such as graph-based anomaly detection, explainable AI (XAI), and transfer learning approaches for resilient optical network protection. By synthesizing current methodologies, datasets, and performance metrics, this review provides a comprehensive perspective on the state-of-the-art in ML-driven anomaly detection and outlines research directions for achieving secure, autonomous, and self-healing optical communication systems.