Neural network visualization

Your browser does not support the <canvas> element.

|

This is a JavaScript implementation of a neural network, the Multilayer Perceptron.
This example in particular is inspired by the one featured in the CS231n class on Convolutional Neural Networks at Standford University.
It is a simple feed-forward neural network with only 1 hidden layer, of variable size.
The network here is training itself to classify the points on the graph according to their class (represented by their color). The colors of the background represent the area of classification as currently evaluated by the network.