Normalize negative scores for text_similarity_reranker retriever #120930
Conversation
This file contains 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
|
Documentation preview: |
pmpailis
added
>bug
:Search Relevance/Ranking
|
Hi @pmpailis, I've created a changelog YAML for you. |
…b.com:pmpailis/elasticsearch into properly_handle_negative_scores_in_retrievers
elasticsearchmachine
added
the
Team:Search Relevance
|
Pinging @elastic/es-search-relevance (Team:Search Relevance) |
benwtrent
reviewed
Jan 28, 2025
server/src/main/java/org/elasticsearch/search/retriever/rankdoc/RankDocsQuery.java
Show resolved
Hide resolved
...pack/inference/rank/textsimilarity/TextSimilarityRankFeaturePhaseRankCoordinatorContext.java
Outdated
Show resolved
Hide resolved
…b.com:pmpailis/elasticsearch into properly_handle_negative_scores_in_retrievers
💔 Backport failed
You can use sqren/backport to manually backport by running |
This was referenced Jan 28, 2025
elasticsearchmachine
pushed a commit
that referenced
this pull request
Jan 28, 2025
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Outdated suggestions cannot be applied.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.
It is valid for some rerankers to return negative scores, so in this PR we add a normalization step for the final scores, to ensure that all scores are >=0, as Lucene assumes that all scores should be positive.
The normalization function that we'll apply is the following:
which will ensure that all positive scores lie in the
[1, inf)while negative values will be projected to(0, 1)Closes #120201