AI engineering glossary · RAG vector search
What is multi-vector retrieval?
Multi-vector retrieval: Representing each document by several vectors, such as one per token, sentence or aspect, rather than a single one. It captures more detail and can improve matching, but needs more storage and a scoring rule that combines vectors.
Multi-vector retrieval: an example
A long article is stored as 6 section vectors, and its score is the best match among them.
Related terms
More RAG vector search terms
- Multimodal retrieval
- Nearest neighbour search
- nprobe
- Parent-document retrieval
- pgvector
- Post-filtering
- Pre-filtering
- Product quantisation
Where you learn it
Multi-vector retrieval is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.