Chapter 10. Neural Networks and MLP with TensorFlow and Keras

The neural network is a modeling technique that was inspired by the structure and functioning of the brain. Just as the brain contains millions of tiny interconnected units known as neurons, the neural networks of today consist of millions of tiny interconnected computing units arranged in layers. Since the computing units of neural networks only exist in the digital world, as against the physical neurons of the brain, they are also called artificial neurons. Similarly, the neural networks (NN) are also known as the artificial neural networks (ANN).

In this chapter, we are going to further expand on the following topics:

• The perceptron (artificial neuron)
• Feed forward neural networks
• MultiLayer Perceptron (MLP) for image classification
  • TensorFlow-based MLP for MNIST image classification
  • Keras-based MLP for MNIST classification
  • TFLearn-based MLP for MNIST classification
• MLP for time series regression