-
Notifications
You must be signed in to change notification settings - Fork 47
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix minimization of improvement-based MC acquisition functions #465
Conversation
@AdrianSosic is there any way that we can (either automatically or manually) test this fix now or did you already verify that this now behaves as intended? |
are we certain that the best_f does not need to be manually handled for non-MC acqfs? I.e. is it all included by setting |
From what I guess, yes. Indeed, have had longer conversation here and see also here. Will add a test though to verify the behavior and ping you again once it's there |
fdeb2da
to
ab0da62
Compare
Not all MC acquisitions functions use best_f (e.g. qUCB), causing a KeyError in these cases
qNegIntegratedPosteriorVariance does not accept an objective
ab0da62
to
f29c4b5
Compare
This PR fixes a critical bug introduced in #340 that has been present since version
0.10.1
and adds a corresponding test. The bug occurs when using improvement-based Monte Carlo acquisition functions (such as the defaultqLogExpectedImprovement
) in with a single numerical target inMIN
mode. The cause was a missing inversion of thebest_f
reference value.