Review07 [coursera] Machine learning - Stanford University - Andrew Ng

最新推荐文章于 2025-03-15 10:57:38 发布

MoMo_qyc

最新推荐文章于 2025-03-15 10:57:38 发布

阅读量237

点赞数 1

CC 4.0 BY-SA版权

分类专栏： course assignment 文章标签： Machine Learning

本文链接：https://blog.youkuaiyun.com/MoMo_flamboyant/article/details/89500827

course 同时被 2 个专栏收录

40 篇文章

订阅专栏

assignment

33 篇文章

订阅专栏

本文探讨了神经网络在表示逻辑函数方面的能力，特别是在处理二进制输入时的表现。文中指出，使用sigmoid激活函数的神经网络的隐藏单元激活值总处于0到1之间，并纠正了一个常见的误解：两层神经网络无法实现XOR函数，实际上XOR是通过NOT XOR来实现的。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

Neural Networks: Representation

Any logical function over binary-valued (0 or 1) inputs x1 and x2 can be (approximately) represented using some neural network. T

The activation values of the hidden units in a neural network, with the sigmoid activation function applied at every layer, are always in the range (0, 1). T

A two layer (one input layer, one output layer; no hidden layer) neural network can represent the XOR function. F XOR is actually implemented as NOT XOR.