Let's learn how to use the Shogun, Dlib, and Shark-ML libraries to create a simple MLP for a regression task. The task is the same for all series samples—MLP should learn cosine functions at limited intervals. In this book's code samples, we can find the full program for data generation and MLP training. Here, we'll discuss the essential parts of the programs that are used for the neural network's API view. Note that the activation functions we'll be using for these samples are the Tanh and ReLU functions. We've chosen them in order to achieve better convergence for this particular task.