Elastic, the Search AI Company, has unveiled performance enhancements in Elasticsearch 9.1 with the introduction of the ACORN-1 algorithm and Better Binary Quantization (BBQ) as the default for high-dimensional dense vectors. These advancements accelerate filtered vector search, improve ranking accuracy, and reduce infrastructure costs, enabling developers to build scalable AI applications efficiently.
The ACORN-1 algorithm enhances filtered k-Nearest Neighbor (kNN) search by integrating filtering into the Hierarchical Navigable Small World (HNSW) graph traversal in Elasticsearch. Unlike traditional methods requiring post-search filtering or pre-indexing, ACORN-1 allows dynamic filter definitions at query time, even post-ingestion. Real-world benchmarks show up to 5X faster query performance while maintaining result accuracy, making it ideal for complex AI-driven searches.
Better Binary Quantization (BBQ), now the default for dense vectors with 384+ dimensions in Elasticsearch 9.1, improves ranking quality and reduces latency. With ~32X compression, BBQ minimizes resource usage while evaluating more candidates during search. Across 10 BEIR datasets, BBQ surpassed traditional float32-based search in 9 cases, based on the NDCG@10 metric, ensuring superior top-10 ranking accuracy for AI applications.
“We’re committed to giving developers the best tools to build and iterate AI applications at scale,” said Ajay Nair, general manager, Platform at Elastic. ACORN-1 and BBQ enable developers to execute high-speed, low-latency queries with reduced memory demands, supporting cost-efficient deployment of AI solutions like recommendation systems and large language models without compromising performance.
Available in Elasticsearch 9.1, these enhancements strengthen Elastic’s position in vector search technology, addressing the demand for efficient, scalable AI solutions. By optimizing query performance and resource usage, ACORN-1 and BBQ empower organizations to develop high-performance AI applications while minimizing operational costs.
Elastic’s advancements in Elasticsearch 9.1 with ACORN-1 and BBQ provide developers with powerful tools to enhance AI application performance. These innovations deliver faster, more accurate searches and significant cost savings, driving the next generation of scalable AI solutions.
Elastic (NYSE: ESTC), the Search AI Company, integrates its deep expertise in search technology with artificial intelligence to help everyone transform all of their data into answers, actions, and outcomes. Elastic's Search AI Platform — the foundation for its search, observability, and security solutions — is used by thousands of companies, including more than 50% of the Fortune 500.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.