This page lists videos made by other people that explain neural networks. It is for a reader who wants a second explanation beside this chapter, or who learns better from watching than from reading. Nothing in the chapter depends on these videos, so you can skip this page and start at neural networks and neurons, which is where the chapter itself begins.
The videos come from two sources. 3Blue1Brown is a YouTube channel that explains mathematics with animated drawings. MIT OpenCourseWare publishes recordings of lectures given at the Massachusetts Institute of Technology (MIT).
Contents#
1. Four videos to watch first#
These four cover the same ground as this chapter and the one after it, so they are the ones shown here to be played on the page. Watch them in the order given.
The first video draws a network that reads a picture of a handwritten digit. It explains what a neuron, a layer, a weight and a bias are, which are the same words that neural networks and neurons introduces.
The second and third videos are the first two lectures of an MIT course on deep learning. The first lecture introduces neural networks and starts on how they are trained.
The second lecture goes on with training, and then builds a network with Keras and TensorFlow, which are code libraries for building neural networks. Its link starts the recording 1 hour, 5 minutes and 59 seconds in.
The fourth video explains gradient descent, which is the method that finds good values for the weights. It belongs with the next chapter, which starts at the score of being wrong.
These four explain what a network is and how it is trained. The next section lists two shorter videos that go one step further.
2. Two more videos#
These two follow on from the videos in four videos to watch first, so they are listed as links rather than shown on the page.
- Backpropagation, intuitively (3Blue1Brown, Deep learning chapter 3) explains how a network works out which way to change each of its weights. It follows straight on from the gradient descent video.
- Large language models explained briefly (3Blue1Brown) explains in a few minutes how the models behind chatbots use neural networks.
The chapter itself starts on the next page, neural networks and neurons, which says what a neural network is in one sentence.