✳flâneur — a map of the web's best reading

3Blue1Brown - What is backpropagation really doing?

3blue1brown.com · 2,762 words · saved by 1 readers

Here we tackle backpropagation, the core algorithm behind how neural networks learn. If you followed the last two lessons or if you’re jumping in with the appropriate background, you know what a neural network is and how it feeds forward information. We’ve been looking at the classic example of recognizing handwritten digits. I’ve been showing a network with... Our basic neural network has four layers: An input layer, an output layer, and two hidden layers in between. We’ve also discussed gradient descent, so you should know that when people describe a network as “learning,” they mean finding the weights and biases that minimize a certain cost function. We use gradient descent to walk downhill to a local minimum of the cost function. To calculate the cost of a single training example, you take the output that the network gives, along with the output you wanted it to give, and you add up the squares of the differences between each component: The cost for a single training example is the

What is backpropagation really doing? | 3Blue1Brown Neural Networks Chapter 4 W h a t i s b a c k p r o p a g a t i o n r e a l l y d o i n g ? An overview of backpropagation, the algorithm behind how neural networks learn. Nov 3, 2017 Lesson by Grant Sanderson Text adaptation by Josh Pullen Source Here we tackle backpropagation, the core algorithm behind how neural networks learn. If you followed the last two lessons or if you're jumping in with the appropriate background, you know what a neural network is and how it feeds forward information. Quick Recap We've been looking at the classic exa

Explore this link on the map →

related reading