Expose validation for TL experiments and fetching of candidate TL sources through AxService #4615
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
|
@Cesar-Cardoso has exported this pull request. If you are a Meta employee, you can view the originating Diff in D84102485. |
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #4615 +/- ##
==========================================
- Coverage 96.52% 96.51% -0.01%
==========================================
Files 568 568
Lines 58958 59227 +269
==========================================
+ Hits 56909 57163 +254
- Misses 2049 2064 +15 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Cesar-Cardoso
force-pushed
the
export-D84102485
branch
from
afcc527 to
d388f46
Compare
Cesar-Cardoso
added a commit
to Cesar-Cardoso/Ax
that referenced
this pull request
Dec 11, 2025
…rces through AxService (facebook#4615) Summary: ## Core Data Structures (`ax/core/auxiliary.py`) Added new dataclasses for auxiliary experiment handling: - **`AuxiliaryExperimentMetadata`** - Abstract base class for metadata - **`PreferenceExplorationMetadata`** - For PE experiments (stores `overlap_metrics`) - **`TransferLearningMetadata`** - For TL experiments (stores `overlap_parameters`) - **`AuxiliaryExperimentValidation`** - Validation result containing `is_valid`, `invalid_reason`, and `metadata` ## Parameter Compatibility (`ax/core/parameter.py`) Added parameter compatibility checking: - New `is_compatible_with()` method on the base `Parameter` class - New abstract `_is_domain_compatible()` method with implementations for: - **`RangeParameter`**: Compatible if both are RangeParameters (bounds don't need to overlap) - **`ChoiceParameter`**: Compatible if they have the same set of values - **`FixedParameter`**: Compatible if they have the same value - **`DerivedParameter`**: Always returns `False` ## Search Space Overlap (`ax/core/search_space.py`) - Added `get_overlapping_parameters()` method to identify compatible overlapping parameters between two search spaces ## Experiment Validation (`ax/core/experiment.py`) - Added `validate_auxiliary_experiment()` method that validates a source experiment against a target experiment for a given purpose (currently supports `TRANSFERABLE_EXPERIMENT`) ## Storage Layer Changes (`ax/storage/sqa_store/load.py`) Moved and refactored transfer learning identification logic: - Added `_query_historical_experiments_given_parameters()` - queries DB for historical experiments with specific parameters - Added `identify_transferable_experiments()` - finds transferable experiments based on parameter overlap - Added `load_candidate_source_auxiliary_experiments()` - loads candidate source experiments for a target experiment Reviewed By: lena-kashtelyan Differential Revision: D84102485
Cesar-Cardoso
added a commit
to Cesar-Cardoso/Ax
that referenced
this pull request
Dec 11, 2025
…rces through AxService (facebook#4615) Summary: ## Core Data Structures (`ax/core/auxiliary.py`) Added new dataclasses for auxiliary experiment handling: - **`AuxiliaryExperimentMetadata`** - Abstract base class for metadata - **`PreferenceExplorationMetadata`** - For PE experiments (stores `overlap_metrics`) - **`TransferLearningMetadata`** - For TL experiments (stores `overlap_parameters`) - **`AuxiliaryExperimentValidation`** - Validation result containing `is_valid`, `invalid_reason`, and `metadata` ## Parameter Compatibility (`ax/core/parameter.py`) Added parameter compatibility checking: - New `is_compatible_with()` method on the base `Parameter` class - New abstract `_is_domain_compatible()` method with implementations for: - **`RangeParameter`**: Compatible if both are RangeParameters (bounds don't need to overlap) - **`ChoiceParameter`**: Compatible if they have the same set of values - **`FixedParameter`**: Compatible if they have the same value - **`DerivedParameter`**: Always returns `False` ## Search Space Overlap (`ax/core/search_space.py`) - Added `get_overlapping_parameters()` method to identify compatible overlapping parameters between two search spaces ## Experiment Validation (`ax/core/experiment.py`) - Added `validate_auxiliary_experiment()` method that validates a source experiment against a target experiment for a given purpose (currently supports `TRANSFERABLE_EXPERIMENT`) ## Storage Layer Changes (`ax/storage/sqa_store/load.py`) Moved and refactored transfer learning identification logic: - Added `_query_historical_experiments_given_parameters()` - queries DB for historical experiments with specific parameters - Added `identify_transferable_experiments()` - finds transferable experiments based on parameter overlap - Added `load_candidate_source_auxiliary_experiments()` - loads candidate source experiments for a target experiment Reviewed By: lena-kashtelyan Differential Revision: D84102485
Cesar-Cardoso
force-pushed
the
export-D84102485
branch
2 times, most recently
from
0d7cfff to
975db3b
Compare
Cesar-Cardoso
added a commit
to Cesar-Cardoso/Ax
that referenced
this pull request
Dec 17, 2025
…rces through AxService (facebook#4615) Summary: ## Core Data Structures (`ax/core/auxiliary.py`) Added new dataclasses for auxiliary experiment handling: - **`AuxiliaryExperimentMetadata`** - Abstract base class for metadata - **`PreferenceExplorationMetadata`** - For PE experiments (stores `overlap_metrics`) - **`TransferLearningMetadata`** - For TL experiments (stores `overlap_parameters`) - **`AuxiliaryExperimentValidation`** - Validation result containing `is_valid`, `invalid_reason`, and `metadata` ## Parameter Compatibility (`ax/core/parameter.py`) Added parameter compatibility checking: - New `is_compatible_with()` method on the base `Parameter` class - New abstract `_is_domain_compatible()` method with implementations for: - **`RangeParameter`**: Compatible if both are RangeParameters (bounds don't need to overlap) - **`ChoiceParameter`**: Compatible if they have the same set of values - **`FixedParameter`**: Compatible if they have the same value - **`DerivedParameter`**: Always returns `False` ## Search Space Overlap (`ax/core/search_space.py`) - Added `get_overlapping_parameters()` method to identify compatible overlapping parameters between two search spaces ## Experiment Validation (`ax/core/experiment.py`) - Added `validate_auxiliary_experiment()` method that validates a source experiment against a target experiment for a given purpose (currently supports `TRANSFERABLE_EXPERIMENT`) ## Storage Layer Changes (`ax/storage/sqa_store/load.py`) Moved and refactored transfer learning identification logic: - Added `_query_historical_experiments_given_parameters()` - queries DB for historical experiments with specific parameters - Added `identify_transferable_experiments()` - finds transferable experiments based on parameter overlap - Added `load_candidate_source_auxiliary_experiments()` - loads candidate source experiments for a target experiment Reviewed By: lena-kashtelyan Differential Revision: D84102485
…rces through AxService (facebook#4615) Summary: ## Core Data Structures (`ax/core/auxiliary.py`) Added new dataclasses for auxiliary experiment handling: - **`AuxiliaryExperimentMetadata`** - Abstract base class for metadata - **`PreferenceExplorationMetadata`** - For PE experiments (stores `overlap_metrics`) - **`TransferLearningMetadata`** - For TL experiments (stores `overlap_parameters`) - **`AuxiliaryExperimentValidation`** - Validation result containing `is_valid`, `invalid_reason`, and `metadata` ## Parameter Compatibility (`ax/core/parameter.py`) Added parameter compatibility checking: - New `is_compatible_with()` method on the base `Parameter` class - New abstract `_is_domain_compatible()` method with implementations for: - **`RangeParameter`**: Compatible if both are RangeParameters (bounds don't need to overlap) - **`ChoiceParameter`**: Compatible if they have the same set of values - **`FixedParameter`**: Compatible if they have the same value - **`DerivedParameter`**: Always returns `False` ## Search Space Overlap (`ax/core/search_space.py`) - Added `get_overlapping_parameters()` method to identify compatible overlapping parameters between two search spaces ## Experiment Validation (`ax/core/experiment.py`) - Added `validate_auxiliary_experiment()` method that validates a source experiment against a target experiment for a given purpose (currently supports `TRANSFERABLE_EXPERIMENT`) ## Storage Layer Changes (`ax/storage/sqa_store/load.py`) Moved and refactored transfer learning identification logic: - Added `_query_historical_experiments_given_parameters()` - queries DB for historical experiments with specific parameters - Added `identify_transferable_experiments()` - finds transferable experiments based on parameter overlap - Added `load_candidate_source_auxiliary_experiments()` - loads candidate source experiments for a target experiment Reviewed By: lena-kashtelyan Differential Revision: D84102485
Cesar-Cardoso
force-pushed
the
export-D84102485
branch
from
975db3b to
70af965
Compare
|
This pull request has been merged in 441f000. |
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:
Core Data Structures (
ax/core/auxiliary.py)Added new dataclasses for auxiliary experiment handling:
AuxiliaryExperimentMetadata- Abstract base class for metadataPreferenceExplorationMetadata- For PE experiments (storesoverlap_metrics)TransferLearningMetadata- For TL experiments (storesoverlap_parameters)AuxiliaryExperimentValidation- Validation result containingis_valid,invalid_reason, andmetadataParameter Compatibility (
ax/core/parameter.py)Added parameter compatibility checking:
is_compatible_with()method on the baseParameterclass_is_domain_compatible()method with implementations for:RangeParameter: Compatible if both are RangeParameters (bounds don't need to overlap)ChoiceParameter: Compatible if they have the same set of valuesFixedParameter: Compatible if they have the same valueDerivedParameter: Always returnsFalseSearch Space Overlap (
ax/core/search_space.py)get_overlapping_parameters()method to identify compatible overlapping parameters between two search spacesExperiment Validation (
ax/core/experiment.py)validate_auxiliary_experiment()method that validates a source experiment against a target experiment for a given purpose (currently supportsTRANSFERABLE_EXPERIMENT)Storage Layer Changes (
ax/storage/sqa_store/load.py)Moved and refactored transfer learning identification logic:
_query_historical_experiments_given_parameters()- queries DB for historical experiments with specific parametersidentify_transferable_experiments()- finds transferable experiments based on parameter overlapload_candidate_source_auxiliary_experiments()- loads candidate source experiments for a target experimentReviewed By: lena-kashtelyan
Differential Revision: D84102485