What is the simplest but still useful NN example you know?

Question

I find neural networks pretty interesting, but most useful examples I know of require too much training data to be useful when it comes to trying to understand the whole process. I used a DOS program over 20 years ago in school to train an NN to do a simple table lookup. That isn't very exciting of course, since it is pretty obvious that a simple array would do much better.I am wondering, are there any small, self-contained NN examples that actually do something useful, while still being small enough to play with an implementation in my favourite programming language?

tgflynn · Accepted Answer

I'm not sure what exactly you're looking for. Dealing with large amounts (by some definition) of training data is at the heart of the "whole process", so I'm not sure what you could learn without that element.One of the simpler data sets available is the MNIST set of labelled hand-written digits. These days it should be easy to build and train a 3-layer NN for classifying those images.

sgillen · Answer

You can make some toy robots do things. See https://gym.openai.com/envs/#classic_controlThe networks and data required to solve some of these is really small. I&rsquo;m sure you can find some simple RL code out there to poke around in.

thedevindevops · Answer

Can I clarify something, are you looking to learn more about how NNs work or are you looking for something you can run/train in your own local machine?