[Causal Inference] Bug found in precedence mincut algorithm (fig. 4 in paper)

There appears to be a typo in the precedence mincut algorithm as outlined in the paper. The error seems to have been corrected in the example notebook, however reproducing the algorithm directly from the paper results in logical errors.

Specifically, lines 15 and 16 in figure 4 both refer to the _cause_ mincut, where they should refer to the _effect_ mincut.

[source: arxiv.org](https://arxiv.org/pdf/2010.12237.pdf)

This error appears to have been fixed in the example notebook, where the true and false sets of the _effect_ mincut are referenced instead:
https://github.com/deepmind/deepmind-research/blob/7488a1f70af89760e335a5c05afed5fe3d06a45e/causal_reasoning/Causal_Reasoning_in_Probability_Trees.ipynb?short_path=1867df5#L267-L271