A Survey of Artificial Intellegence Techniques to Detect Intrusion Attack in Communication Network

Networks are a huge part of modern life, and cyber security has become a crucial area of research. An intrusion detection system (IDS) is a device that examines network traffic to confirm security, availability, and integrity. Despite years of research, conventional IDSs continue to struggle with enhancing detection performance, lowering False Alarm Rates (FAR), and identifying novel assaults. Many academics have concentrated on constructing IDSs that leverage Artificial Intelligence (AI) technologies to overcome the difficulties mentioned above. With extreme accuracy, AI algorithms can reliably identify the main distinctions among normal and anomalous data. Furthermore, AI approaches are very generalizable, allowing them to identify previously undiscovered assaults. IDS methods based on ML, DL, and meta-heuristic optimization have recently been proposed as workable alternatives for spotting intrusions throughout the network. This survey presents an IDS taxonomy that classifies and summarizes AI-based IDS research using data items as the primary dimension. This article offers a thorough analysis of recent studies on the application of AI in IDSs, including surveys and suggestions for cutting-edge techniques. Following that, recent developments and advances in AI-based network IDS are explored, as well as the proposed strategy, assessment metrics, and dataset selection. This study highlighted many research obstacles and presented future direction for research in developing AI-based Network IDS using the inadequacies of the existing approaches.