[
https://issues.apache.org/jira/browse/SOLR-16827?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Joel Bernstein updated SOLR-16827:
----------------------------------
Summary: Add min/max scale and RRF capability to the reranker (was: Add
min/max scale and RRF params to reranker)
> Add min/max scale and RRF capability to the reranker
> ----------------------------------------------------
>
> Key: SOLR-16827
> URL: https://issues.apache.org/jira/browse/SOLR-16827
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Joel Bernstein
> Assignee: Joel Bernstein
> Priority: Major
> Labels: hybrid-search
>
> When combining an unbounded Lucene / Solr score with a vector score its
> difficult or impossible to balance the two components by simply adjusting the
> weight of the rerank score.
> This ticket will add two features to the ReRanker:
> Min/Max scaling: The min/max scaling will be applied to the scores of all the
> topDocs prior to combining with the rerank score.
> RRF: Reciprical rank fusion which is described here in this blog:
> https://weaviate.io/blog/hybrid-search-explained
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
