Fix handling of abandoned arms in ExperimentData #4155
Closed
+155
−29
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
) Summary: `Acquisition._update_objective_thresholds` or `infer_objective_thresholds` call within can lead to errors during `Acquisition` initialization, even if the corresponding acquisition function doesn't need objective thresholds. We have some conditional logic to protect against this but it is not comprehensive enough. The particular example that triggered this diff is the use of P_Feasible acquisition function when there are no feasible points. In this case, we can't infer objective thresholds since we need feasible points. This diff adds a `try/except` block to catch errors raised in `infer_objective_thresholds` and logs a warning if any error is caught. This should make it possible to easily debug any related errors (thanks to the log) while letting the acquisition functions that don't need the thresholds run without issues. Reviewed By: sunnyshen321 Differential Revision: D80259199
Summary: This diff makes two changes to `extract_experiment_data` to correctly filter out abandoned (and more generally other invalid status) trials / arms. - Updates `extract_observation_data` to check for `(Batch)Trial.abandoned_arms` and discard those if abandoned is not a valid status. - Updates filtering of `arm_data` to keep only the rows that have corresponding rows in `observation_data`. Previously this only checked if there was corresponding data in `Data` object, but this didn't account for invalid trial statuses. Keeping it in sync with `observation_data` should help eliminate some downstream issues. Reviewed By: sunnyshen321 Differential Revision: D80292268
meta-cla
bot
added
the
CLA Signed
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D80292268 |
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #4155 +/- ##
=======================================
Coverage 96.04% 96.05%
=======================================
Files 566 566
Lines 57168 57214 +46
=======================================
+ Hits 54908 54955 +47
+ Misses 2260 2259 -1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Contributor
|
This pull request has been merged in faf4c6e. |
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:
This diff makes two changes to
extract_experiment_datato correctly filter out abandoned (and more generally other invalid status) trials / arms.extract_observation_datato check for(Batch)Trial.abandoned_armsand discard those if abandoned is not a valid status.arm_datato keep only the rows that have corresponding rows inobservation_data. Previously this only checked if there was corresponding data inDataobject, but this didn't account for invalid trial statuses. Keeping it in sync withobservation_datashould help eliminate some downstream issues.Reviewed By: sunnyshen321
Differential Revision: D80292268