opendistro-for-elasticsearch / k-NN

🆕 A machine learning plugin which supports an approximate k-NN search algorithm for Open Distro.
https://opendistro.github.io/
Apache License 2.0
277 stars 55 forks source link

Perform clustering #359

Open gauthier-schweitzer opened 3 years ago

gauthier-schweitzer commented 3 years ago

Hi ! Thanks for the amazing job you're doing. I want to use elasticsearch to perform on the fly clustering of hundreds of thousands of dense vectors. Is there a way I can use some sort of aggregation based on similarity ? My intuition is that HNSW graphs could be useful for that : is there a way to export them ? Thanks

jmazanec15 commented 3 years ago

Hi @gauthier-schweitzer , could you elaborate on your use case more a little bit? Maybe give an example?

In terms of exporting them, what do you mean by that? Could you not just use nmslib directly?