Chapter 7
Introducing Neural Networks
IN THIS CHAPTER
Considering the perceptron
Dealing with complex data
Developing strategies for overcoming overfitting
You may have heard the term neural network in reference to artificial intelligence. The first thing you need to know is that the correct term is Artificial Neural Network (ANN) because no one has discovered any method of recreating a real brain, which is where the concept of a neural network comes from. Chapter 2 of this book describes the various approaches to deep learning, of which ANNs are one. You find the term shortened in this book because everyone else is using the short term, but you need to know that ANN is actually the correct term and that they’re the work of the connectionist tribe. (See the “Discovering the five main approaches” section of Chapter 2 for a discussion of the five tribes of machine learning and the approaches they’ve developed to solving problems.)
After you get past the whole idea that your computer lacks a brain — at least a real brain — you can begin to appreciate the perceptron, which is the simplest type of neural network. The perceptron is the focus of many of the neural network pictures you ...