Studying the Reinforcement Learning techniques for the problem of intrusion detection

The intrusion detection problem is one of the core problems in cyber-security. In recent years, there are many research studies that leverage the power of supervised machine learning algorithms to detect the attacking flows from the Internet. In this study, we study the reinforcement learning techniques to solve the problem. Our experiments using the latest realworld dataset showed that the current reinforcement learning algorithms in general cannot achieve the same performance as the well studies supervised machine learning algorithms but they do have different advantages, such as the dynamic nature.