Getting Started
Course Information
Introduces the course structure, learning objectives, and the prerequisites needed to get started.
Credits
A comprehensive list of credits for the sources that inspired and informed the course material. Useful for further reading.
Unit 1: Building a Neural Network
The Structure of Neural Networks
Introduces the fundamental building blocks of neural networks, from individual neurons and activation functions to the layered architecture of feedforward networks.
Digit Classifier: Neural Network Architecture
A hands-on lab to implement the basic structure of a neural network, preparing it to classify handwritten digits.
Training Neural Networks
Explores how neural networks learn using cost functions, gradient descent, and the backpropagation algorithm.
Digit Classifier: Training a Neural Network
This lab applies the concepts of backpropagation and stochastic gradient descent to train the digit classifier network.
Behind the Scenes of Training
Explores the intuition behind the training algorithms and discusses what a neural network actually learns, from simple features to complex patterns.
Unit 2: Improving a Neural Network
Activation Functions, Cost Functions, and Gradient Descent
This chapter compares activation functions, cost functions, and gradient descent variants to get a better understanding of how each affects performance.
Initialization, Data. and Regularization
Covers how quality and quantity of data and proper weight initialization influence learning, and introduces regularization techniques to combat overfitting and improve generalization.
Hyperparameters, Debugging, and Practicality
Discusses often-overlooked parts of setting up a network such as the hyperparameters and practicality, and addresses the trickiness of debugging a neural network.