Machine learning, neural networks and backpropagation in Processing

edited November 7 in General

I am interested in learning to build multi-layered neural networks. I have only been programming for two months or so and I only know processing, and have grown quite used to its syntax.

My question is directed towards those with experience in the field: is building neural networks with at least one hidden layer and training them with backpropagation possible and optimal to do in processing? I have noticed a neural networks library in processing. Does anybody have experience with it, what parts of building the network does it make easier? If it is doable in processing, what is a good place to start learning?

I am aware that it’s a pretty advanced subjext but I have all the time and enthusiasm needed to learn it. Do I have to learn a new language like straight-up Java or Python though? Is learning it in processing too impractical?

Comments

  • @randomdude --

    Re: "possible and optimal to do in Processing"

    Possible?

    Yes. Here are some recent posts in the forum that you might find interesting as examples:

    https://forum.processing.org/two/search?Search=neural

    Optimal?

    If you mean large-throughput, high-speed performance than no, none of the Processing modes (Java, JavaScript, Python(transpiled), R(transpiled)...) are optimized for raw speed. But that usually isn't required... is that a requirement for your project?

Sign In or Register to comment.