Example solution for causal reasoning in probability trees

In the following lines, it looks like `see(cut)` and `do(cut)` should be replaced by `see(cut_data)` and `do(cut_data)`. This doesn't affect P(Bias|do(D)), but it does affect P(Bias|D).

https://github.com/deepmind/deepmind-research/blob/7488a1f70af89760e335a5c05afed5fe3d06a45e/causal_reasoning/Causal_Reasoning_in_Probability_Trees.ipynb#L3633-L3651