【读书1】【2017】MATLAB与深度学习——代价函数与学习规则(1)

代价函数与学习规则（Cost Function and Learning Rule）

本节简要说明了代价函数是什么，以及它如何影响神经网络的学习规则。

This section briefly explains what the costfunction is and how it affects the learning rule of the neural network.

该函数也被称为损失函数或目标函数。

It is also called the loss function andobjective function.

代价函数是一个与优化理论相联系的数学概念。

The cost function is a rather mathematicalconcept that is associated with the optimization theory.

你不必了解这个概念。

You don’t have to know it.

然而，如果你想要更好地理解神经网络的学习规则，那么最好是能够深刻理解它。

However, it is good to know if you want tobetter understand the learning rule of the neural network.

这并不是一个很难理解的概念。

It is not a difficult concept to follow.

代价函数与神经网络的监督学习有关。

The cost function is related to supervisedlearning of the neural network.

第二章阐述了神经网络的监督学习就是通过调整权值来减小训练数据误差的过程。

Chapter 2 addressed that supervisedlearning of the neural network is a process of adjusting the weights to reducethe error of the training data.

在这种情况下，代价函数是神经网络误差的度量。

In this context, the measure of the neuralnetwork’s error is the cost function.

神经网络的误差越大，则代价函数的值也越大。

The greater the error of the neuralnetwork, the higher the value of the cost function is.

神经网络的监督学习主要有两种代价函数类型。

There are two primary types of costfunctions for the neural network’s supervised learning.

图3-9 训练输出与正确输出之间的差异越大，则误差越大The greaterthe difference between the output and the correct output, the larger the error

显然，代价函数值与误差成正比。

It is clearly noticeable that the costfunction value is proportional to the error.

这种关系非常直观，不需要进一步解释。

This relationship is so intuitive that nofurther explanation is necessary.

神经网络的大多数早期研究都采用这种代价函数来推导学习规则。

Most early studies of the neural networkemployed this cost function to derive learning rules.

前一章的增量规则是从该函数导出，而且反向传播算法也是如此。

Not only was the delta rule of the previouschapter derived from this function, but the back-propagation algorithm was aswell.

回归问题仍然使用这个代价函数。

Regression problems still use this costfunction.

——本文译自Phil Kim所著的《Matlab Deep Learning》

更多精彩文章请关注微信号：