[ DEPRECATED ] This project uses an old version of ml5.js.

ML PONG

Machine Learning Pong it's an experimental version of the famous computer game Pong that aims to educate users on how machine learning tasks are trained through a fun gaming experience. The user is able to teach, in realtime, a neural network and then play the game in the browser using a set of assigned emojis as reference.

How it works

ML Pong uses the P5 DeepLearn library https://github.com/cvalenzuela/p5-deeplearn-js to train the controllers and predict the output.

Demo

Video Demo version runs in 2 different windows browsers and is sending the data through web sockets. (The repo for this project combines the game and the controller in one single page)

ML PONG

Usage

Currently version of the game runs the game and the controller at the same page.*

• Clone Repository.
• Create private directory inside of /server and add your cert.pem and key.pem certificates.
• Replace your credentials in theserver.js file.
• Run $ npm install
• Start $ npm start

Project will run at port 3000.

Test

https://am7673.itp.io:3000

License

MIT