koolo233 / neuralnetworks

Python implementation of Neural Networks

NeuralNetworks

Python implementation of Neural Networks

网页打开jupyter可能存在公式渲染错误的问题，在VS Code中利用Markdown Preview Enhanced进行渲染则可以保证正常。

具体操作为：

• 安装VS Code以及Markdown Preview Enhanced插件
• 打开README.md并启用Markdown Preview Enhanced渲染
• 切换到需要进行渲染的Jupyter Markdown Block并进入编辑模式

ToDo list

• Perceptron
• Adaptive Linear Neuron(Adaline)
• Basic Automatic Differentiation(AD)
• Multilayer Perceptrons(Fully Connected Neural Network)
• Radial-Basis Function Networks(RBF)
• Hopfield
• Jordan
• Elman
• Hamming Neural Network
• Self-Organizing Feature Map(SOFM)
• Cerebellar Model Articulation Controller(CMAC)
• Deep Boltzmann Machine(DBM)
• Deep Belief Network(DBN)
• Convolutional Neural Networks(CNN)
• Recurrent Neural Network(RNN)
• Graph Neural Network(GNN)
• Graph Convolutional Netword(GCN)
• Deep Q-Learning(DQN)
• Neural Architecture Search(NAS -> ENAS)

Result

Perceptron

• and

• or

• Nor(x1)

• Nor(x2)

• xoR

• 线性可分样本簇

• 线性不可分样本簇

Adaptive Linear Neuron(Adaline)

• and

• or

• Nor(x1)

• Nor(x2)

• xoR

• 线性可分样本簇

• 线性不可分样本簇

Basic Automatic Differentiation(AD)

Multilayer Perceptrons(Fully Connected Neural Network)

• and

• or

• Nor(x1)

• Nor(x2)

• xoR

• 线性可分样本簇

• 线性不可分样本簇

• 非线性回归

Radial-Basis Function Networks(RBF)

• 分类

• 回归

Hopfield

• 联想记忆

Elman

• 回归1

• 回归2

HammingNN

• 联想记忆

SOFM

• 保序映射

• 聚类

Cerebellar Model Articulation Controller(CMAC)

• 拟合函数