0

I have studied neural network theory and know how they work on a basic level. I have completed a few exercises on the coursera and I know how to define layers and train the network in Python but I can only do this on the sample datasets present. So I made a game using pygame and wanted to make a neural network to train for it but I don't understand how I can make that work.

I know what my input layers will be, I know what my output layer would be, I just can not understand how to implement a reward system to improve the network. I understand how to make test data and train it against it but that's supervised learning which is not what I need, I want to implement q-learning in this project

user9252284
  • 1
  • 1
  • Hi welcome to SO, Your question in its current state very vague. There are ton of tutorials in this topic across the internet. These might interest you [PLE](https://pygame-learning-environment.readthedocs.io/en/latest/), [similar SO qestion](https://stackoverflow.com/questions/58974034/pygame-and-open-ai-implementation), [OpenAI](https://openai.com/) – Nijeesh Joshy Dec 24 '19 at 14:12

1 Answers1

1

Check out NEAT(NeuroEvolution of Augmenting Topologies).

https://neat-python.readthedocs.io/en/latest/index.html

It is a method for evolving artificial neural networks with a genetic algorithm. And it is a python module that can be imported and used.

Jad Haddad
  • 11
  • 2