Fix bbq quantization algorithm but for differently distributed components #126778
+65
−4
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
benwtrent
added
>bug
auto-backport
elasticsearchmachine
added
the
Team:Search Relevance
|
Pinging @elastic/es-search-relevance (Team:Search Relevance) |
|
Hi @benwtrent, I've created a changelog YAML for you. |
john-wagster
approved these changes
Apr 14, 2025
There was a problem hiding this comment.
I believe this was the same fix you made here but applied to ES correct: apache/lucene#14374
lgtm no concerns
benwtrent
added
the
auto-merge-without-approval
benwtrent
added a commit
to benwtrent/elasticsearch
that referenced
this pull request
Apr 14, 2025
…ents (elastic#126778) We had a silly bug in quantizing vectors in bbq where we were scaling the initial quantile optimization parameters incorrectly given the vector component distribution. In distributions where this has a major impact, the recall results were abysmal and rendered the quantization technique useless. In modern, well distributed components, this change is almost a no-op.
elasticsearchmachine
pushed a commit
that referenced
this pull request
Apr 14, 2025
…ents (#126778) (#126794) We had a silly bug in quantizing vectors in bbq where we were scaling the initial quantile optimization parameters incorrectly given the vector component distribution. In distributions where this has a major impact, the recall results were abysmal and rendered the quantization technique useless. In modern, well distributed components, this change is almost a no-op.
elasticsearchmachine
pushed a commit
that referenced
this pull request
Apr 14, 2025
…ents (#126778) (#126793) We had a silly bug in quantizing vectors in bbq where we were scaling the initial quantile optimization parameters incorrectly given the vector component distribution. In distributions where this has a major impact, the recall results were abysmal and rendered the quantization technique useless. In modern, well distributed components, this change is almost a no-op.
|
Hi, @benwtrent, will this change in quantification algorithm affect existing bbq quantification data? Do old vector segments need to be repaired? |
For older segments that were getting good results, this change won't effect them. For older segments that were getting very poor results, they will need to be reindexed to take advantage of the fix. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
We had a silly bug in quantizing vectors in bbq where we were scaling the initial quantile optimization parameters incorrectly given the vector component distribution.
In distributions where this has a major impact, the recall results were abysmal and rendered the quantization technique useless.
In modern, well distributed components, this change is almost a no-op.