書名： Python Deep Learning
作者名： Ivan Vasilev Daniel Slater Gianmario Spacagna Peter Roelants Valentino Zocca
本章字?jǐn)?shù)： 234字
更新時間： 2021-07-02 14:31:03

Neural Networks

In Chapter 1, Machine Learning – an Introduction, we introduced a number of basic machine learning(ML) concepts and techniques. We went through the main ML paradigms, as well as some popular classic ML algorithms, and we finished with neural networks. In this chapter, we will formally introduce what neural networks are, describe in detail how a neuron works, see how we can stack many layers to create a deep feedforward neural network, and then we'll learn how to train them.

In this chapter, we will cover the following topics:

The need for neural networks
An introduction to neural networks
Training neural networks

Initially, neural networks were inspired by the biological brain (hence the name). Over time, however, we've stopped trying to emulate how the brain works and instead we focused on finding the correct configurations for specific tasks including computer vision, natural language processing, and speech recognition. You can think of it in this way: for a long time, we were inspired by the flight of birds, but, in the end, we created airplanes, which are quite different. We are still far from matching the potential of the brain. Perhaps the machine learning algorithms in the future will resemble the brain more, but that's not the case now. Hence, for the rest of this book, we won't try to create analogies between the brain and neural networks.

官术网_书友最值得收藏!

Python Deep Learning

Neural Networks