[
https://issues.apache.org/jira/browse/LUCENE-10318?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17581942#comment-17581942
]
Michael Sokolov commented on LUCENE-10318:
------------------------------------------
Another idea I played with at one point was to preserve all the graphs from the
existing segments (remapping their ordinals) and linking them together with
additional links. But a lot of links needed to be created in order to get close
to the recall for a new "from scratch" graph, and I struggled to get any
improvement. At the time I wasn't even concerning myself about deletions.
> Reuse HNSW graphs when merging segments?
> ----------------------------------------
>
> Key: LUCENE-10318
> URL: https://issues.apache.org/jira/browse/LUCENE-10318
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Julie Tibshirani
> Priority: Major
> Labels: vector-based-search
>
> Currently when merging segments, the HNSW vectors format rebuilds the entire
> graph from scratch. In general, building these graphs is very expensive, and
> it'd be nice to optimize it in any way we can. I was wondering if during
> merge, we could choose the largest segment with no deletes, and load its HNSW
> graph into heap. Then we'd add vectors from the other segments to this graph,
> through the normal build process. This could cut down on the number of
> operations we need to perform when building the graph.
> This is just an early idea, I haven't run experiments to see if it would
> help. I'd guess that whether it helps would also depend on details of the
> MergePolicy.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
