Setting Up Machine Learning Model

Note - These are my notes on DeepLearning Specialization Improving Deep Neural Networks (Week 1)

Table of Contents¶

1  Setting up your machine Learning Application

1.1  Train / Dev / Test sets

1.2  Bias / Variance

1.3  Basic Recipe for Machine Learning

2  Regularizing your neural network

2.1  Regularization

2.2  Why regularization reduces overfitting?

2.3  Dropout Regularization

2.4  Understanding Dropout

2.5  Other regularization methods

3  Setting up your optimization problem

3.1  Normalizing inputs

3.2  Vanishing / Exploding gradients

3.3  Weight Initialization for Deep Networks

3.4  Numerical approximation of gradients

3.5  Gradient Checking

3.6  Gradient Checking Implementation Notes

Setting up your machine Learning Application¶

Train / Dev / Test sets¶

Bias / Variance¶

Basic Recipe for Machine Learning¶

Regularizing your neural network¶

Regularization¶

Why regularization reduces overfitting?¶

Dropout Regularization¶

Understanding Dropout¶

Other regularization methods¶

Setting up your optimization problem¶

Normalizing inputs¶

Vanishing / Exploding gradients¶

Weight Initialization for Deep Networks¶

Numerical approximation of gradients¶

Gradient Checking¶

Gradient Checking Implementation Notes¶

Credits to the teacher