Discover the recent enhancements and optimizations that notably improve vector search performance in Elasticsearch & Lucene vector database.

Elasticsearch and Lucene report card: noteworthy speed and efficiency investments

Harnessing Lucene's architecture for multi-threaded search

Accelerating multi-graph vector search

Harnessing Java's latest advancements for ludicrous speed

Maximizing memory efficiency with scalar quantization

Achieving seamless compression with minimal impact on recall

Paving the way for binary quantization

Multi-vector integration in Lucene and Elasticsearch

Wrapping up (for now)

Making Elasticsearch and Lucene the best vector database

Making Elasticsearch and Lucene the best vector database: up to 8x faster and 32x efficient

Explore the cost, performance and benchmarking for running large-scale vector search in Elasticsearch, with a focus on high-fidelity dense vector search.

Part 1: High-fidelity dense vector search

Introduction

Notes on embeddings

Matryoshka Representation Learning (MRL)

Embedding quantization learning

Adaptive embedding quantization

Large scale benchmark

Back-of-the-envelope estimation

Benchmark configuration

Indexing performance

Search evaluation

Conclusion

Elasticsearch vector search: Designing for large scale

Designing for large scale vector search with Elasticsearch

Author

Jim Ferenczi

Articles

Making Elasticsearch and Lucene the best vector database: up to 8x faster and 32x efficient

Designing for large scale vector search with Elasticsearch