Deep-Learning in Python by Keras on top of Tensorflow

Contents

• Forward Propagation Neural Networks

• Activation Function (RELU)

• Applying the network to many observations/rows of data

• Multi-Layer Neural Network

• Loss Function

• Mean Squared Error (MSE), Scaling up loss function to multiple data points

• Calculating Slopes Using Gradient Descent

• Calculating Error Using Learning Rate

• Making multiple updates to weights

• Backward Propagation Neural Networks

• Building Models in Keras (Specify Architecture, Compile Model, Fit Model, Prediction)

• Classification Models in Keras (Titanic Dataset)

• Using Models in Keras (Save, Reload, and add extra features for classification problems)

• Model Optimization via Stochastic Gradient Descent(SGD) with different Learning Rates

• Model Validation in Keras Using Early Stopping Monitor

• Model Comparison Using Matplotlib in order to check the validation score