The First Cry of Atom Today is the first day of the rest of my life.

Yes, You are Malala

<meta property=”og:image” content=”/images/posts/2014-02-14-you-are-malala/malala.jpg” /> We, Japanese have built high level democratic society. We have an universal suffrage system, the rights to discuss through a lot of type media, and the constitution which guarantees our liberty of expression. The end of the Edo era, we have kep... Read more

NN with nodejs

2 layer newral network is added to n42 This network is simple newral network which can trained throught gradient descent optimization calculation. It is the same algorithm to the one of denoised autoencoder used by n42. So implementation itself was not diffucult. The code is showed below. /** * Training weight parameters with supervised le... Read more

Restricted Boltzmann Machine in node

n42 has restricted boltzmann machine. Restricted boltzmann machine is a kind of autoencoder in order to extract charasteristics. In this article, I would leave out the detail of restricted boltzmann machine. I want to introduce how the RBM is written in node. /* * Restricted Boltzmann machine * * @module n42 * @class RBM * @author ... Read more

Stacked denoised autoencoder with nodejs

I developed deep leanring module which enables you to use stacked denoised autoencoder in nodejs. This is called n42. You can train with deep learning algorithm very easily. https://npmjs.org/package/n42 How to use This is how to use it. var n42 = require('n42'); // input data // This is made of sylvester matrix var input ... Read more

Autoencoder with node.js

Auto encoder is used for deep learning. Auto encoder extract characteristics of data through unsupervised learning. This is a kind of newral network. By using an auto encoder, you don’t have to be in trouble with choicing extracting algorithm, or doing yourself. Therefore, in deepleanring field, this algorithm is used very actively. There are m... Read more