A Review of Deep Learning based Intrusion Detection Systems

In recent years, proliferation of network attacks and cyber threats has underscored the critical need for robust and efficient Intrusion Detection Systems (IDS) to safeguard digital assets and sensitive information. Deep Learning (DL) has emerged as a promising technology for enhancing accuracy and effectiveness of IDS due to its ability to automatically extract intricate patterns and anomalies from large-scale network data. This review paper provides a comprehensive survey of DL-based IDS, offering insights into their strengths, limitations, and recent advancements. Due to the necessity of cyber security, numerous researches have been undertaken to categorize the data moving through the network. Intrusion detection is the prominent method of cyber security to identify the intruders and lessen the network hazard. The present study work has concentrated on the necessity and procedure of IDS using DL mechanisms. Reliability and feasibility of IDS depend on performance and accuracy of deep learning. Present research has conducted a review study of conventional research work and focused on the methodology used during IDS detection and classification using ML and DL. The limitation of conventional research work is also presented in problem statement to propose a better deep learning approach in the future.