add eta to get_acquisition_function #1541
Conversation
facebook-github-bot
added
the
CLA Signed
Codecov Report
@@ Coverage Diff @@
## main #1541 +/- ##
=========================================
Coverage 100.00% 100.00%
=========================================
Files 153 153
Lines 13598 13598
=========================================
Hits 13598 13598
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
|
I also fixed in this PR the linting issue from main. |
| @@ -61,8 +62,13 @@ def get_acquisition_function( | |||
| constraints: A list of callables, each mapping a Tensor of dimension | |||
| `sample_shape x batch-shape x q x m` to a Tensor of dimension | |||
| `sample_shape x batch-shape x q`, where negative values imply | |||
| feasibility. Used when constraint_transforms are not passed | |||
| as part of the objective. | |||
| feasibility. Used only for qEHVI and qNEHVI. | |||
There was a problem hiding this comment.
@sdaulton, @saitcakmak I guess this special handling is needed b/c we don't yet have posterior transforms implemented for MOO? Do we have any concrete plans to change this? It would be nice to make this interface more consistent.
There was a problem hiding this comment.
Hmm, I thought the reason is that in qNEHVI and qEHVI the constraints are defined directly in the acquisition function whereas in the other acqfs, we define the constraints as part of the objective and then give the objective to the constructor of the acquisition function. Or is there another reason?
There was a problem hiding this comment.
I think @jduerholt is correct here. IIRC, in q(N)EHVI we use constraints to weight the HVI, whereas in other acqfs, we feasibility weight the samples directly (via the objective).
|
@Balandat has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
@Balandat has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
1 similar comment
|
@Balandat has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
Motivation
This PR adds the
etaparameter for the constraints also to get_acquisition_function.Have you read the Contributing Guidelines on pull requests?
Yes.
Test Plan
Unit tests.