Application of Deep Learning Techniques in Cyber-Attack Detection

Abstract

Cyber security is an emerging technology that provides the requisite process and security framework to protect information and resources across the network. Cyber-attack is one of the critical problems of cyber security. To ensure security, a large number of the defensive systems and softwares are available such as firewalls, IDS, and antivirus for protecting information over the network. But due to change in the network algorithms, new attacks are introduced that disrupt the normal functioning of the network. Hence, there is a need of efficient system to detect and mitigate cyber-attacks. The conventional techniques of attack detection such as heuristic, fuzzy, and rule-based techniques failed to provide efficiency on large datasets; therefore, deep learning techniques were introduced. Deep learning techniques provide better performance on large datasets, and as well as they are helpful in detection of unknown attacks. This study focuses on the recent advancement in attack detection system using deep learning techniques and also presents the comparative study of techniques with performance evaluation on different attack datasets.