VECTOR-DATABASE-EDI

Vector database for EDI knowledge base, AI & data integration applied to EDI.

Definition

A vector database is optimised for storing and searching high-dimension vector spaces. Algorithms: HNSW (Hierarchical Navigable Small Worlds), IVF (Inverted File), DiskANN. Players: Pinecone, Weaviate, Qdrant, Milvus, pgvector (Postgres extension). Acts as back-end for EDI RAG assistants.

Origin

Category emerging from 2018 with Faiss (Meta), commercialised by Pinecone from 2021.

Example in context

Pinecone serverless with a 1536-d index, 1M EDI vectors, queries < 50 ms p99.

Related terms

• Embeddings EDI — stored data.