验证的一种变化形式是交叉验证。
A variation of validation iscross-validation.
根据训练方法进行分类，机器学习可以分为有监督学习、无监督学习和强化学习。
Depending on the training method, MachineLearning can be supervised learning, unsupervised learning, and reinforcementlearning.
下表给出了这些学习方法的训练数据格式。
The formats of the training data for theselearning methods are shown here.
有监督学习可分为分类和回归，这取决于模型的使用。
Supervised learning can be divided intoclassification and regression, depending on the usage of the model.
分类用于确定输入数据属于哪个类别。
Classification determines which group theinput data belongs to.
分类的正确输出用类别形式给出。
The correct output of the classification isgiven as categories.
相反，回归预测估计值并获取训练数据中正确的输出值。
In contrast, regression predicts values andtakes the values for the correct output in the training data.
第二章 神经网络（Chapter 2 Neural network）
本章介绍广泛应用于机器学习的神经网络。
This chapter introduces the neural network,which is widely used as the model for Machine Learning.
神经网络具有悠久的发展历史以及大量的研究成果。
The neural network has a long history ofdevelopment and a vast amount of achievement from research works.
有许多书籍纯粹集中在神经网络的讲解上。
There are many books available that purelyfocus on the neural network.
随着最近对深度学习的研究热度不断增长，神经网络的重要性也显著增加。
Along with the recent growth in interestfor Deep Learning, the importance of the neural network has increasedsignificantly as well.
我们将简要回顾相关的和实用的技术，以更好地了解深度学习。
We will briefly review the relevant andpractical techniques to better understand Deep Learning.
对于那些刚学习神经网络的新手，我们从基本原理开始。
For those who are new to the concept of theneural network, we start with the fundamentals.
首先，我们来看看神经网络是如何与机器学习相关的。
First, we will see how the neural networkis related to Machine Learning.
机器学习的模型可以通过各种形式实现。
The models of Machine Learning can beimplemented in various forms.
神经网络就是其中之一。
The neural network is one of them.
很简单，不是吗？
Simple isn’t it?
图2-1说明了机器学习与神经网络之间的关系。
Figure 2-1 illustrates the relationshipbetween Machine Learning and the neural network.
图2-1 机器学习与神经网络的关系The relationship betweenMachine Learning and the neural network
注意，我们用神经网络代替模型，用学习规则代替机器学习。
Note that we have the neural network inplace of the model, and the learning rule in place of Machine Learning.
在神经网络的应用下，确定模型（神经网络）的过程称为学习规则。
In context of the neural network, theprocess of determining the model (neural network) is called the learning rule.
——本文译自Phil Kim所著的《Matlab Deep Learning》
更多精彩文章请关注微信号：
