Neural Network Assisted IDS/IPS: An Overview of Implementations, Benefits, and Drawbacks
Kyle Rozendaal, Thivanka Dissanayake-Mohottalalage, Akalanka Bandara Mailewa · International Journal of Computer Applications · 2022
Modern IDS use inflexible knowledge bases, rule sets and rely on human interaction for successful threat mitigation.While this approach to network and hardware security has been effective in the past, the explosion of large data breaches in the past few years reveals a lack of effective detection for unknown or undocumented threats.We infer that a change in detection and prevention of cybercrime needs to start at the system level and use more intelligent methods of attack detection and prevention: Neural Networks and Artificial Intelligence assisted IDS.This paper gives a broad overview of the modern state of IDS/IPS systems, discusses the benefits and drawbacks of modern implementation, gives a broad overview of current research into the field of neural networkbased IDS, and discusses benefits and drawbacks of NNIDS systems.Finally, we conclude with a few examples of modern implementations of NNIDS and areas for future study in the field.