Add EnsembleMapSaasGP #3035
Closed
Add EnsembleMapSaasGP #3035
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
meta-cla
bot
added
the
CLA Signed
|
@saitcakmak has exported this pull request. If you are a Meta employee, you can view the originating Diff in D83701925. |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #3035 +/- ##
=======================================
Coverage 99.98% 99.98%
=======================================
Files 216 216
Lines 20468 20504 +36
=======================================
+ Hits 20464 20500 +36
Misses 4 4 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Summary: Introduces a simple `EnsembleMapSaasGP` model that will replace `get_fitted_map_saas_ensemble` (which fits individual non-ensemble models and combines them into a fully Bayesian GP). The model internally is a batched `ExactGP`, which behaves just like a multi-output `SingleTaskGP`. The `posterior` method is overwritten to produce a `MixtureGaussianPosterior`, which retains the old behavior of the ensemble model. The benefit of this model class is that it can be fit just like any other GP model, using `ExactMarginalLogLikelihood` and `fit_gpytorch_mll`. As such, it is fully compatible with Ax's MBM setup (as long as `allow_batched_models=False`). Reviewed By: sdaulton Differential Revision: D83701925
saitcakmak
force-pushed
the
export-D83701925
branch
from
e423c6d to
a61ee05
Compare
saitcakmak
added a commit
to saitcakmak/botorch
that referenced
this pull request
Oct 2, 2025
Summary: Introduces a simple `EnsembleMapSaasGP` model that will replace `get_fitted_map_saas_ensemble` (which fits individual non-ensemble models and combines them into a fully Bayesian GP). The model internally is a batched `ExactGP`, which behaves just like a multi-output `SingleTaskGP`. The `posterior` method is overwritten to produce a `MixtureGaussianPosterior`, which retains the old behavior of the ensemble model. The benefit of this model class is that it can be fit just like any other GP model, using `ExactMarginalLogLikelihood` and `fit_gpytorch_mll`. As such, it is fully compatible with Ax's MBM setup (as long as `allow_batched_models=False`). Reviewed By: sdaulton Differential Revision: D83701925
|
This pull request has been merged in 0650dcc. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
Introduces a simple
EnsembleMapSaasGPmodel that will replaceget_fitted_map_saas_ensemble(which fits individual non-ensemble models and combines them into a fully Bayesian GP). The model internally is a batchedExactGP, which behaves just like a multi-outputSingleTaskGP. Theposteriormethod is overwritten to produce aMixtureGaussianPosterior, which retains the old behavior of the ensemble model.The benefit of this model class is that it can be fit just like any other GP model, using
ExactMarginalLogLikelihoodandfit_gpytorch_mll. As such, it is fully compatible with Ax's MBM setup (as long asallow_batched_models=False).Differential Revision: D83701925