Vamshi Vijay Nakkirtha

Hi @luyuncheng, We have few concerns with the non optimized index for Hamming distance in nmslib. In Elasticsearch we would store serialized graph per segment which means one additional file...

Hi @juliusbachnick, This is not a bug and expected behavior from k-NN design. Plugin extracts k neighbors from each segment inside the shard and that accounts for atleast k hits....

Hi @juliusbachnick, Good point. Seems like a limitation on usage of aggregations on the k-NN search. Possible work around i could see is having shards with segments force merged to...

Sorry for getting late on this thread. Elasticsearch performs aggregations on all the documents that are "hit". Since 'k' nearest neighbors are evaluated at segment level, each segment return at...

This Pr is put on hold. Please refer https://github.com/opendistro-for-elasticsearch/k-NN/issues/283#issuecomment-755598088

https://github.com/opendistro-for-elasticsearch/k-NN/issues/133

Hi @RamanRajarathinam, We are in plans to support FAISS. Nothing concrete yet. Will keep this thread open as a Feature request. Based on the community feedback we could prioritize the...

Hi @JamesIsHuang, At this moment we do not have the numbers documented. We are planning to come up with dedicated blog on this. Meanwhile this blog could give some insights...

Hi @JamesIsHuang , We have done performance analysys for different vector dimensions and collection. We need t formalize and put in the consumable manner. We are prioritizing the effort to...

Hi @JamesIsHuang, we made k dynamic and its between 50 and 1500. Please note, we also did warmup to have graphs loaded to memory and our experiment do not account...