Python All-in-One For Dummies

Chapter 2

Building a Neural Network in Python

IN THIS CHAPTER

How machines learn

Understanding the basics of machine learning

Teaching a machine to learn something

Using TensorFlow to do machine learning

Neural networks and various other models of how the brain operates have been around as long as people have been talking about AI. Marvin Minsky, introduced in our last chapter, started mainline interest in modeling neurons with his seminal work in perceptrons in 1969. At the time, there was widespread “irrational exuberance” about how the perceptron was going to make AI practical very quickly. This attracted a good deal of venture capital to the area, but when many of these ventures failed, investment in neural networks dried up. It is like the concept of fashion in science. What’s popular, sells.

Fast forward 30 years. There is now renewed interest in neural networks. Better models have been built, but the really important thing is that we now have real, useful, and economical applications based on neural networks. Will this lead to another bubble? Most assuredly, but this time ...