I am new in the field of AI. I am working to create the flappy bird using Genetic Algorithm. After reading and seeing some examples, I saw that most implementations use a Neural Network + Genetic Algorithm and after certain generations, you achieve a very good agent that survives very long.
I currently struggling to implement the Neural Network since I have never taken a Machine Learning course. On many examples that I have read neural networks require training inputs and outputs. For the flappy bird, I can't think of output since you don't really know if the action of flapping will benefit you or not.
In the example that I followed, Synaptic.js is used and it is pretty straight-forward. However, in Python, I can't find a simple library that will initialize randomly and adjust the weights and biases depending on the good agents that survive longer.
What would be the right way to implement this Neural Network without having a training dataset?
Is there any way to create Flappy Bird without using Neural Networks, just Genetic Algorithm?