This chapter addresses many, although not all, security challenges: the base-rate fallacy, the data availability issue, and the active attacker problem. The machine learning that are relevant to cybersecurity are classification and clustering is very important. Applications to control and automation, such as self-driving cars and flying and coordinating drones are also areas of application for machine learning. The linear classifier is the geometrically simplest of all classifiers and the easiest to draw. The classifier simply defines a line that splits the data so that one side of the line corresponds to one class and the other corresponds to the other class. In the 1970s and 1980s one of the most popular methods of artificial intelligence was expert systems. The perceptron model is inspired from the cell in our brains, the biological neuron. Neural network models are a class of models that are widely used for classification and regression tasks.