An Overview of Explainable Artificial Intelligence for Cyber Security

Abstract

The rapid development of the Internet, various form of host and network attack have emerged, to detect and recognize different categories of attack Intrusion Detection System (IDS) have deployed as a defensive tool to detect attacks. However IDS based on manual and traditional techniques as signatures of known attacks and deviation of normal activity have become obsolete in the field of cyber security. Recently Artificial Intelligent (AI) especially Machine Learning (ML) and Deep Learning (DL) techniques are applied in IDS to construct a model which can be able to detect variety of attacks in real time. This work aims to provide an overview of various type of IDS, AI especially their two branches ML and DL. We also explain the importance of their conjunction in cyber security. Furthermore the different public dataset and various metrics used to analyze, compared and evaluate a ML and DL techniques for Intrusion Detection has been presented. Finally a series of discussion showed how AI enforce the effectiveness of cyber security. At the end of this chapter the different challenges of application of AI in Cyber Security are explored.KeywordsCyber securityArtificial IntelligenceDeep learningIntrusion detection systemAttackSignature detectionAnomaly detectionDatasetPerformance evaluation