BayesianPEIsCorrect test fails probabilistically

[tcNickolas]

Describe the bug
BayesianPEIsCorrect test fails probabilistically.
Similarly to microsoft/QuantumLibraries#444, we should fix the seed used in the test, or run several tries in a loop and say that the test passes if at least one of the runs succeeds.

To Reproduce
The last failure I've seen.

[cgranade]

At least for fixing the seed, that's unfortunately not currently supported (as per your earlier issue at microsoft/qsharp-runtime#947). In the meantime, adding this to the maintenance backlog to consider your suggestion of looping several times.