A Review on Generative Intelligence in Deep Learning based Network Intrusion Detection
Mohammad Ali, Ifiok Udoidiok, Fuhao Li, Jielun Zhang · 2024
The incorporation of generative intelligence into deep learning-based intrusion detection systems (IDS) has become a viable method for improving cybersecurity. While existing surveys have extensively covered the use of Generative Adversarial Networks (GANs), there remains a gap in exploring other generative models, such as Variational Autoencoders (VAEs) and autoregressive models, within this domain. This paper presents a detailed review of the integration of these underexplored generative models within IDS that offer novel approaches to addressing challenges like zero-day attack detection, dataset imbalance, and adversarial robustness. By analyzing existing literature and recent advancements, this study emphasizes the potential of these generative techniques in improving IDS performance. Additionally, the paper identifies critical challenges in current research, proposes future research directions, and discusses the ethical and practical considerations of deploying generative models in real-world network intrusion detection.