DEEP LEARNING-BASED INTRUSION DETECTION SYSTEMS: A COMPREHENSIVE SURVEY OF FOUR MAIN FIELDS OF CYBER SECURITY

The security flaws in cyber security have always put the users and organizations at risk, which as a result created catastrophic conditions in the network that could be either irreversible or sometimes too costly to recover. In order to detect these attacks, Intrusion Detection Systems (IDSs) were born to alert the network in case of any intrusions. Machine Learning (ML) and more prominently deep learning methods can be able to improve the performance of IDSs. This article focuses on IDS approaches whose functionalities rely on deep learning models to deal with the security issue in Internet of Things (IoT), wireless networks, Software Defined Networks (SDNs), and Industrial Control Systems (ICSs). To this, we examine each approach and provide a comprehensive comparison and discuss the main features and evaluation methods as well as IDS techniques that are applied along with deep learning models. Finally, we will provide a conclusion of what future studies are possibly going to focus on in regards to IDS, particularly when using deep learning models.