Year: 2026
Type: book-chapter
Abstract: The growth of cloud computing, IoT, and 5G/6G has led to a sharp rise in network traffic. As a result, networks are now more exposed to threats like DDoS attacks, MITM attacks, and unusual traffic patterns. Over the years, several methods have been used to detect such anomalies, including signature-based techniques, machine learning, and deep neural networks. However, each of these comes with its own drawbacks, such as poor handling of time-series data, high energy use, and weak performance against new or unknown threats. Recently, neuromorphic computing and Spiking Neural Networks (SNNs) have emerged as promising tools for processing spatio-temporal data. Still, not enough attention has been given to how these techniques can be applied to network anomaly detection. While a number of surveys have looked into neuromorphic computing in the context of AI, most of them do not specifically address the unique challenges that arise in network security. To fill this gap, this survey reviews studies published between 2020 and 2026 that focus on neuromorphic approaches for detecting anomalies in networks. Our work offers four key contributions. First, it presents a clear classification of network anomalies and how they relate to neuromorphic detection methods. Second, it reviews SNN models and learning rules used in cybersecurity. Third, it highlights major challenges, including scalability, lack of proper datasets, and hardware limitations. Finally, it suggests possible future directions for using neuromorphic systems in real-time, low-power detection for edge devices and IoT applications.
Source: IntechOpen eBooks
Author: Hossein Seilani
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204247969