Even better(er) binary quantization (#117994)

This measurably improves BBQ by adjusting the underlying algorithm to an
optimized per vector scalar quantization.

This is a brand new way to quantize vectors. Instead of there being a
global set of upper and lower quantile bands, these are optimized and
calculated per individual vector. Additionally, vectors are centered on
a common centroid. 

This allows for an almost 32x reduction in memory, and even better
recall than before at the cost of slightly increasing indexing time.

Additionally, this new approach is easily generalizable to various other
bit sizes (e.g. 2 bits, etc.). While not taken advantage of yet, we may
update our scalar quantized indices in the future to use this new
algorithm, giving significant boosts in recall.

The recall gains spread from 2% to almost 10% for certain datasets with
an additional 5-10% indexing cost when indexing with HNSW when compared
with current BBQ.
This commit is contained in:
Benjamin Trent 2024-12-09 11:06:27 -05:00 committed by GitHub
parent 0586cbfb34
commit 5e859d9301
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
32 changed files with 3501 additions and 137 deletions

View file

@ -459,7 +459,9 @@ module org.elasticsearch.server {
org.elasticsearch.index.codec.vectors.ES815HnswBitVectorsFormat,
org.elasticsearch.index.codec.vectors.ES815BitFlatVectorFormat,
org.elasticsearch.index.codec.vectors.es816.ES816BinaryQuantizedVectorsFormat,
org.elasticsearch.index.codec.vectors.es816.ES816HnswBinaryQuantizedVectorsFormat;
org.elasticsearch.index.codec.vectors.es816.ES816HnswBinaryQuantizedVectorsFormat,
org.elasticsearch.index.codec.vectors.es818.ES818BinaryQuantizedVectorsFormat,
org.elasticsearch.index.codec.vectors.es818.ES818HnswBinaryQuantizedVectorsFormat;
provides org.apache.lucene.codecs.Codec
with