A Survey on Machine Learning and Explainable AI Techniques for the Detection of Zero Day Web Attacks

The increase in internet usage has increased the scope of cyber threats. Among them, zero day web attacks are the most crucial type of cyber attacks. Zero day web attacks are the unknown attacks and detecting them is the greatest challenge. Therefore, an intrusion detection system that is capable of detecting such zero day web attacks with this experimental setup is highly required. It is intended to use machine learning algorithms over the collected CICIDS2017 dataset in order to detect the unknown attacks. Later, techniques of Explainable AI can be implemented on the best performing model for gaining clear insights on the models decision thus increasing the transparency over the model. This paper majorly focuses on developing a clear understanding on zero day attack detection and the techniques applied for their detection by referring to different research work.