Machine Learning in processing?

Ater

Hi there!

You hear about neural networks from almost everywhere now. Curiosity suggests me to try it. As I know processing, I thought I can for example write some simple game like arcanoid, and leave the controls to some Library or API that gonna use reinforcement algorithm and train itself to play my game.

I looked through some Java Libraries, particulary BURLAP it has good tutorials, but I wonder will I be able to use it in processing. It has its own visualization interface, but it looks like it is pretty simple to bind parameters to processing world. However, I'm not sure if it would work with processing draw() cycles, there is an interface to manage states, but I did not find something like step() that I could call from draw() in processing. I'm not experienced enough to go through the source code, although it is open source.

Before I continue my research, I was wondering, If someone tried something like I want?

It looks like Shiffman works on some tutorials: https://github.com/shiffman/Machine-Learning-Processing but I'm so impatient.

Thanks!

jeremydouglass

Some recent interesting discussions on machine learning -- you might reach out to those users already working in this area, like @SeanS6 --

• https://forum.processing.org/two/search?Search="machine+learning"

CharlesDesign

I wrote an article/tutorial about create Neural network from scratch in processing which you might find interesting. https://medium.com/typeme/lets-code-a-neural-network-from-scratch-part-1-24f0a30d7d62