Hilde Weerts

Hi @matthew-so - could you perhaps open separate issues for the different sections? In that way other prospective contributors can easily identify that they can pick it up!

Hi @bramreinders97: awesome! Could you open up a new issue in which you refer back to this issue? Thanks!

Thanks a lot for letting us know, @matthew-so!

That would be really useful - I have assigned the issue to you, @alliesaizan! Let us know if anything is unclear.

Hi, @Hadyark! Relabeling by swapping elements seems reasonable to me. A possible caveat is that the meaning of the proportions in the leafs is lost after swapping, so we might...

> I only handled the case of a binary tree because it was similar to my work. If it is necessary to adapt it for multiclass classification, it will be...

To avoid re-running all tests we can also hardcode a few re-tries to OpenML calls, perhaps with a short time delay. That usually fixed my problems with API calls during...

I think so. We can always re-open if we still encounter any issues.

Reopened this issue (see https://github.com/fairlearn/fairlearn/pull/732#issuecomment-824583523) as well as the one the at OpenML repo (https://github.com/openml/OpenML/issues/1092). @riedgar-ms you mentioned you've previously had some retry logic there, what was the reason that...

I don't know of a specific build, but @romanlutz mentioned you've encountered this issue several times over the past few weeks (see https://github.com/fairlearn/fairlearn/pull/732#issuecomment-824583523).