您的位置: 首页  >  IT文章  >  andrew ng machine learning week4 神经网络

andrew ng machine learning week4 神经网络

分类: IT文章 • 2022-04-23 20:03:55
吴恩达机器学习全套笔记博客地址
 

一、模型表示

计算机视觉中一个像素点可能需要很多特征向量来表示，于是我们需要用其他的形式来计算如此巨大的输入量

andrew ng machine learning week4 神经网络

二、神经元与神经网络

andrew ng machine learning week4 神经网络

神经元受到来自其他神经元所传递来的输入，数目达到一定的阈值以后，便会产生输出

andrew ng machine learning week4 神经网络

andrew ng machine learning week4 神经网络

输入层-隐藏层-输出层三层结构：

andrew ng machine learning week4 神经网络

a的右上角脚标代表的是layer 

andrew ng machine learning week4 神经网络

andrew ng machine learning week4 神经网络

andrew ng machine learning week4 神经网络

向量化的表示：

andrew ng machine learning week4 神经网络

andrew ng machine learning week4 神经网络

andrew ng machine learning week4 神经网络

三、实际应用

and

andrew ng machine learning week4 神经网络

or 

andrew ng machine learning week4 神经网络

取反：

andrew ng machine learning week4 神经网络

异或

andrew ng machine learning week4 神经网络

四、多分类问题

手写数字识别，输出的y是one-hot representation的一种，代表是某一个种类

andrew ng machine learning week4 神经网络

相关推荐