Fix qNEI with Derivative Enabled BO

[yyexela]

Motivation

Derivative enabled GPs allow for faster convergence rates. This is well studied in the literature and makes intuitive sense since a lot more information is provided with each objective function evaluation. I want code to be available for myself and other future researchers who want to explore derivative enabled GPs.

This PR resolves an error with the qNEI acquisition function for derivative enabled GPs. For more information, see here.

Have you read the Contributing Guidelines on pull requests?

Yes, this doesn't affect the outcome of the unit tests, though when I run the unit tests on an unmodified branch of main I get 110 failed, 853 passed, 1165 warnings.

Test Plan

I have some files from an issue (1679) I opened which show a simple case where a derivative-enabled Gaussian Process fails to find the qNEI value at a point where an equivalent non-derivative-enabled Gaussian Process does not fail.

• shared.py.txt
• Simple_Repro_non_D_2.py.txt
• Simple_Repro_non_D.py.txt

Adding the fix in this PR removes the error.

Please let me know if I should add a specific unit test for this.

Related PRs

N/A

[codecov]

Codecov Report

Merging #1716 (d239247) into main (eaa6fb2) will not change coverage.
The diff coverage is 100.00%.

❗ Current head d239247 differs from pull request most recent head fcb9584. Consider uploading reports for the commit fcb9584 to get more accurate results

@@            Coverage Diff            @@
##              main     #1716   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files          170       170           
  Lines        14632     14633    +1     
=========================================
+ Hits         14632     14633    +1     

Impacted Files	Coverage Δ
botorch/posteriors/gpytorch.py	100.00% <100.00%> (ø)

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[facebook-github-bot]

@Balandat has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[SebastianAment]

Thank you for putting up this PR @yyexela! There's one code formatting error that currently prevents us from merging this in. Could you change the order of the imports to be alphabetical from

    BlockDiagLinearOperator,
    LinearOperator,
    SumLinearOperator,
    DenseLinearOperator,

to

    BlockDiagLinearOperator,
    DenseLinearOperator,
    LinearOperator,
    SumLinearOperator,

Thanks again!

[yyexela]

Hey @SebastianAment , no problem! The fix should be pushed, let me know if there's anything else I can do!

[facebook-github-bot]

@SebastianAment has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@esantorella has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@esantorella merged this pull request in 0b744dd.