A Survey on the Applications of Deep Learning in Network Intrusion Detection Systems to Enhance Network Security
Fatima M. Anis, Maryam Alabdullatif, Sarah Aljbli, Mohammad Ali A. Hammoudeh · IEEE Access · 2025
Network security breaches continue to grow in both complexity and impact, making intrusion detection a critical component of modern defense strategies. Intrusion Detection Systems (IDS) monitor system and network activity to identify malicious behavior and stop it before causing harm. causes harm. Among them, Network-based IDS (NIDS) is particularly important because it analyzes network traffic at the packet or flow level, enabling protection across a wide range of infrastructure. Deep Learning (DL) has recently emerged as a transformative approach for NIDS by learning complex representations of normal and malicious traffic patterns without the need for extensive manual feature engineering. This survey provides a review of recent progress in DL for NIDS and introduces a taxonomy with four categories. These include reconstruction and generative models that detect anomalies, Transformer-based sequence models for capturing temporal dependencies, convolutional and deep neural networks for supervised classification, and hybrid or deployment-oriented approaches that combine techniques or support real-world deployment needs. This survey reviews 31 studies published between 2020 and 2025, covering their datasets, feature sets, model architectures, and reported performance. A comparative analysis highlights common practices and emerging trends. Finally, we identify open challenges related to data quality, scalability, adversarial robustness, privacy, and interpretability, and we outline promising directions for future research. This work contributes an updated review, a structured framework for comparison, and practical insights to guide the development of more secure and effective NIDS.