I'm not familiar with physics, especially fundmental physics. But this theory is really attractive to me, so please let me make some discussion on this topic.
So is it possible to use genetic algorithm or even AI algorithms, such as gradient descent or even GAN, to somehow 'fit' this true universe?
To achieve this, we may need some evaluation functions which may evaluate the simulated graph, that is generated by a given method.
For example, we generate a universe U by a given rule set S. And we already have some evaluate functions F1 ~ Fn, which in ture universe R, we know F1(R) = H1 ... Fn(R) = Hn.
So we can tell how much this generated universe U is real by calculating Fi(U) and compare the real world results to Hi.
Is it possible?