Per Herald Borgen has created a very interesting tutorial on how to build a neural network in the browser using Brain.js. In his example he uses the classical flower dataset that's been so widely implemented in many machine learning tutorials.
The network will learn to classify image flowers into three categories:
- Iris Setosa
- Iris Virginica
- Iris Versicolor.
According to Borgen:
"Setting up the network is actually super simple. Brain.js provides a high-level interface for doing this. This is perfect for developers who want to start playing around with neural networks without having to learn the underlying theory first. I'm personally a big fan of this kind of top-down learning." [source]
The parameters that are readily configurable for the net are:
- the activation function
- the number of hidden layers
- the number of iterations (for the learning process)
- the learning rate.
The code is available for modification and for testing directly in the browser on the scrimba website, which is actually really cool - as it lets you develop and test code of multiple files (like a project) in the browser. If you want to play with this yourself, please visit the link below:
To stay in touch with me, follow @cristi
Cristi Vlad Self-Experimenter and Author
Good nice new post shearing thank you pelese my vote halp me
yes baby. nature is a network in browser of brain when any question in enter the mind then brain working for answer depends on nature.
good nature post
Nice intro to a machine learning library in Javascript. Upvoted and followed. :)