Neural Networks and Deep Learning by Michael Nielsen

neuralnetworksanddeeplearning.com AI Learning & Guides PublicationFree

This online book teaches neural networks and deep learning through explanations and exercises.

Neural Networks and Deep Learning by Michael Nielsen homepage screenshot

About Neural Networks and Deep Learning by Michael Nielsen

Neural Networks and Deep Learning is an online book by Michael Nielsen that teaches the core concepts of neural networks and deep learning. Its main aim is a solid understanding of principles rather than a hazy grasp of many ideas. Readers need some programming experience, but they do not have to be professional programmers. The book covers perceptrons, backpropagation, regularization and convolutional networks, and it includes exercises. The code is written in Python, and a small neural network library is built as the book proceeds. It is licensed for non-commercial sharing, so commercial use requires contacting the author.

Key features

Covers perceptrons, sigmoid neurons and gradient descent
Covers backpropagation and overfitting with regularization
Introduces convolutional networks and the vanishing gradient problem
Builds a small neural network library in Python through the book

Best for: Readers with some programming experience who want to understand the core concepts of neural networks and deep learning.

Good for

Includes interactive JavaScript elements that the text often expects readers to use

Watch out for

Code is written in Python 2.7
Not a tutorial for a particular neural network library
No PDF or print edition
Last update listed is December 2019
No published solutions to the exercises and problems