TensorFlow is not a library to directly build neural networks. In a similar way, NumPy is not a library to perform data mining; it just does the heavy lifting and is generally used from another library. TensorFlow contains a built-in library, referred to as TensorFlow Learn to build networks and perform data mining. Other libraries, such as Keras, are also built with this in mind and use TensorFlow in the backend.

Keras implements a number of modern types of neural network layers and the building blocks for building them. In this chapter, we will use convolution layers which are designed to mimic the way in which human vision works. They use small collections of connected neurons that analyse only a segment of the input values - in this case, an image. This allows the network to deal with standard alterations such as dealing with translations of images. In the case of vision-based experiments, an example of an alteration dealt with by convolution layers is translating the image...