Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[FEATURE] Lucene Based k-NN search support #380

Closed
vamshin opened this issue May 2, 2022 · 2 comments
Closed

[FEATURE] Lucene Based k-NN search support #380

vamshin opened this issue May 2, 2022 · 2 comments

Comments

@vamshin
Copy link
Member

vamshin commented May 2, 2022

Is your feature request related to a problem?
Recently Lucene started building in-house Approximate k-NN search in Java based on HNSW algorithm. k-NN plugin already has support for HNSW based out of C library namely NMSlib. By having the Lucene based HNSW support, we will be able to provide AKNN functionality independently outside of k-NN plugin with out having to rely on the NMSlib or FAISS. This will unlock the platform dependency issues because of C based libraries.

As a prerequisite we need to introduce new data type dense_vector in OpenSearch core and build Field Mappers to index data from this field to Lucene graphs.

@elfisher
Copy link

I added the roadmap label since this is on the project roadmap.

@alexahorgan
Copy link

Demo feedback (8/3/22):

Outcome:
Approved, ship it.

Action Items/Follow up:

  • Connect with documentation team to document limits and approaches.
  • Follow up on performance and scale.
  • Begin exploring opportunities to leverage UI to drive insights/run searches.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

6 participants