Introduction to deep learning

Introduction to Deep Learning, basic theory and full coding examples.

Part 1 - Linear perceptron and the concept of learning

Discussion is on the structure of a neuron and the algorithmical concept concept behind learning. The coded example single linear neuron or perceptron, learning the logic OR and AND Gates.

• Inspiration - The Biological Neuron Model
• Perceptron - An Artificial Neural Network
• Decision Units - Activation functions
• Learning - the Bias and weights
• Perceptron Example - Logic OR

Part 2 - Non Linear Perceptron

How to extend learning,with deep networks and non linear activation functions.

• Why deep learning ?
• Non Linear Activation function
• Handwritten Digit Classification (Mnist )

Part 3 - Multiple Layer Neural Networks

Delta rule for learning, Gradient Descent and backpropagation

• How to train Multiple Layer Neural Networks
• Gradient Decent - Minimize loss function iteratively
• Delta rule for learning
• Backpropagation
• Train and predict MLP for Logic OR - example

Part 4 - Keras Mnist example

Build and visualize Simple feed forward dense network with Mnist data

• Load and normalize data set
• Create the model
• Train the model
• Evaluate Model
• Predict

Part 5 - Convolution Neural Network

What is convolution network ?

• Intuition and examples
• Basic terms and building blocks of CNN
• Keras CNN to classify Mnist data

utils_plot.py

Few simple python plotting functions