Inspiration

Nowadays, bigdata and machine learning techniques are expanding relentlessly. We are sure you are not currently aware of the level of processement your data suffers each day, or even perhaps each minute! In order to let you know more about those techniques, we have designed this project.

What it does

This website gives the user some insight about cutting-edge machine learning techniques, highlighting neural networks. Users can teach both a regular NARX and an extremely interesting genetic NARX customized image filters, make a perceptron try to build any image depeding on the location of its bits and finally test the efficacy of a regressive perceptron regarding color classifciation

How we built it

This project is solely based on an Angular JS front end in which all the logic resides. Since it doesn't store any sort of information, databases are not needed. Additonally, since all the services reside on the front end, there is no need to use an API as well. This is an extremely powerful project done in a very minimalistic environment.

Challenges we ran into

Although the project was of a complex nature, we surpassed all the problems that stepped on our way. The main adversities we faced were: the choice of neural netowork to use in each case and its parameters. The other big issue was to keep this project as simple as possible.

Accomplishments that we're proud of

We managed to create an easy to use webpage with powerful and interesting neural networks. Also we had the satisfaction to bring these technologies closer to people. This is why we are so proud of this project.

What we learned

We have learned how to implement neural networks in Javascript, using new sorts of technologies and improved the knowledge of the ones we already knew, enhancing our experience.

What's next for Neulter

As it's only a webclient application, turning it into a server application of some sort with additional services would be the path to be followed.

You can check our work at our GitHub organization in the link you can find below.

Share this project:

Updates