Toward Automated Penetration Testing Intelligently with Reinforcement Learning

As the world was moving into the generation of Artificial Intelligent (AI), many sectors utilise AI for various types of tasks. In this research, an intelligent automated penetration testing system is introduced in this thesis. The research is about implementing a machine learning technique called reinforcement learning to predict the use of Metasploit Framework’s module and achieve the best performance and result while conducting automated penetration testing with the Metasploit framework. In this research, two learning algorithms have been implemented, Q-learning and Deep Q-learning. In this research, Q-learning has achieved a notable result and also discovered a flaw of the purposed method in this research