Mastering Machine Learning for Penetration Testing

Reinforcement learning

Previously (especially in the first chapter), we explored the different machine learning models: supervised, semi-supervised, unsupervised, and reinforcement models. Reinforcement machine learning models are important approaches to building intelligent machines. In reinforcement learning, an agent learns through experience, by interacting with an environment; it chooses the best decision based on a state and a reward function:

A famous example of reinforcement learning is the AI-based Atari Breakout. In this case, the environment includes the following:

• The ball and the bricks
• The moving paddle (left or right)
• The reward ...