Tutorial on GIN

[adam2392]

Is your feature request related to a problem? Please describe.
Before implementing a native version, a first step could be to demonstrate the usage of GIN by for example replicating one of the simulation results in the original paper. One should note that the algorithm only has guarantees when all observed variables have no direct causal relationships with each other.

It should be noted in the tutorial.

Afterwards, a native implementation (#87) can take place that can use the unit tests and the tutorial to check for correctness.

Describe the solution you'd like
Add a notebook to docs/tutorials, or add a short python script example to examples/. I think since this is "explaining an algorithm", this would be more "in-depth" and should therefore be a tutorial.

Additional context
cc: @robertness wanna add this, since you're more familiar?