incremental qLogNEI #2760
Closed
incremental qLogNEI #2760
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
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D70288526 |
sdaulton
force-pushed
the
export-D70288526
branch
from
2033d27 to
d1d4c5c
Compare
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #2760 +/- ##
=======================================
Coverage 99.99% 99.99%
=======================================
Files 203 203
Lines 18685 18705 +20
=======================================
+ Hits 18684 18704 +20
Misses 1 1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
sdaulton
added a commit
to sdaulton/botorch
that referenced
this pull request
Mar 5, 2025
Summary: This diff adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely). The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized. TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used. Reviewed By: esantorella Differential Revision: D70288526
sdaulton
force-pushed
the
export-D70288526
branch
from
d1d4c5c to
5c4f885
Compare
sdaulton
added a commit
to sdaulton/botorch
that referenced
this pull request
Mar 5, 2025
Summary: This diff adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely). The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized. TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used. Reviewed By: esantorella Differential Revision: D70288526
sdaulton
force-pushed
the
export-D70288526
branch
from
5c4f885 to
9b6cb72
Compare
|
This pull request was exported from Phabricator. Differential Revision: D70288526 |
sdaulton
added a commit
to sdaulton/botorch
that referenced
this pull request
Mar 5, 2025
Summary: Pull Request resolved: meta-pytorch#2760 This diff adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely). The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized. TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used. Reviewed By: esantorella Differential Revision: D70288526
sdaulton
force-pushed
the
export-D70288526
branch
from
9b6cb72 to
99d4093
Compare
|
This pull request was exported from Phabricator. Differential Revision: D70288526 |
sdaulton
added a commit
to sdaulton/botorch
that referenced
this pull request
Mar 5, 2025
Summary: Pull Request resolved: meta-pytorch#2760 This diff adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely). The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized. TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used. Reviewed By: esantorella Differential Revision: D70288526
sdaulton
force-pushed
the
export-D70288526
branch
from
99d4093 to
d101274
Compare
sdaulton
added a commit
to sdaulton/botorch
that referenced
this pull request
Mar 5, 2025
Summary: This diff adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely). The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized. TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used. Reviewed By: esantorella Differential Revision: D70288526
sdaulton
force-pushed
the
export-D70288526
branch
from
d101274 to
7c5117a
Compare
sdaulton
added a commit
to sdaulton/botorch
that referenced
this pull request
Mar 5, 2025
Summary: This diff adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely). The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized. TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used. Reviewed By: esantorella Differential Revision: D70288526
sdaulton
force-pushed
the
export-D70288526
branch
2 times, most recently
from
0500e05 to
7c5117a
Compare
|
This pull request was exported from Phabricator. Differential Revision: D70288526 |
sdaulton
force-pushed
the
export-D70288526
branch
from
7c5117a to
fc6757f
Compare
sdaulton
added a commit
to sdaulton/botorch
that referenced
this pull request
Mar 5, 2025
Summary: Pull Request resolved: meta-pytorch#2760 This diff adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely). The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized. TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used. Reviewed By: esantorella Differential Revision: D70288526
Summary: Pull Request resolved: meta-pytorch#2760 This diff adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely). The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized. TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used. Reviewed By: esantorella Differential Revision: D70288526
|
This pull request was exported from Phabricator. Differential Revision: D70288526 |
sdaulton
force-pushed
the
export-D70288526
branch
from
fc6757f to
e8c88cb
Compare
|
This pull request has been merged in 290c0ba. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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 adds an incremental qLogNEI, that addresses many cases where the first candidate in the batch has positive EI (and satisfies the constraints) and subsequent arms violate the constraints (often severely).
The issue appears to stem from optimizing the joint EI of the new candidate and the pending points w.r.t the current incumbent(s). My hypothesis is that this makes the initialization strategy perform worse and choose bad starting points. Using sequential batch optimization and optimizing the incremental EI of the new arm relative to the pending points (and the current incumbent) avoids the issue by only quanitifying the improvment of the current arm being optimized.
TODO: add this for qNEI in a later diff, but that seems low pri since qLogNEI is widely used.
Reviewed By: esantorella
Differential Revision: D70288526