Azure Cosmos DB adds spherical quantization for vector search (preview)
aiazurepreviewengineerazure-cosmos-db
feature announcement
Azure Cosmos DB is introducing spherical quantization in public preview, a new vector compression technique that enhances vector indexing speed and accuracy. This feature aims to improve search recall and precision for large-scale vector operations. It is now available for users to test and adopt.
- →Public preview of spherical quantization for vector search
- →Spherical quantization as an advanced vector compression technique
Features (1) ›
- Public preview of spherical quantization for vector search
Azure Cosmos DB now offers spherical quantization in public preview. This feature is designed to achieve faster vector indexing while maintaining high recall and accuracy at scale.
Notes (1) ›
- Spherical quantization as an advanced vector compression technique
Spherical quantization is described as an advanced vector compression technique. Its primary purpose is to preserve search quality during the compression process.
Read the original announcement →
https://azure.microsoft.com/updates?id=561167
Related releases
- SQL MCP Server for Azure Updates is Generally Available Azure Updates ·
- Azure Synapse Link for Cosmos DB NoSQL is being retired Azure Updates ·
- Azure Cosmos DB agent kit now generally available Azure Updates ·
- Azure Backup for Cosmos DB now in public preview Azure Updates ·
- Azure Cosmos DB Linux emulator now generally available Azure Updates ·
- Azure Updates: AI-powered discovery and OmniVec embedding toolkit announced Azure Updates ·