Chrysolites Tech Academy

Video by freeCodeCamp.org · taught by Brandon Rohrer · streamed from YouTube. All rights remain with the creator.

Lesson 1: How Deep Neural Networks Work

About this lesson

This beginner course by Brandon Rohrer explains how deep neural networks work, aiming to make you comfortable both with the concepts and with the mathematics behind them, which is presented visually and built up step by step. It is made of a series of connected, illustrated lectures. The first explains the fundamentals: how a single artificial neuron combines its inputs with weights, how neurons are arranged into layers and how a network of very simple parts can represent complicated patterns. The next looks at what neural networks can learn and how they learn it, explaining error, gradient descent and the way backpropagation passes responsibility for an error back through the layers. The course then explains how convolutional neural networks work, showing how filters, rectified linear units and pooling let a network recognise patterns in images wherever they appear. Recurrent neural networks and long short-term memory units follow, showing how networks can handle sequences such as text by remembering what came earlier, and how gates decide what to keep and what to forget. A lecture that demystifies deep learning discusses what these networks really do well and where their limits are, and another considers how robotics might move machines closer to human intelligence. The course closes with an in-depth second look at convolutional networks. Afterwards, read Chapter 1 of the course book, which works through a backpropagation step by hand.

Install Chrysolites Tech Academy

Android / Chrome / Edge: open the browser menu and choose Install app or Add to Home screen.

iPhone / iPad (Safari): tap the Share button, then Add to Home Screen.