# multi-vector

> Browse 6 items tagged with multi-vector.

## Items

- [MUVERA](https://vectordb.ever.works/ja/items/muvera) — Multi-Vector Retrieval Algorithm that reduces multi-vector similarity search to single-vector similarity search via Fixed Dimensional Encodings. Achieves 10% improved recall with 90% lower latency compared to existing approaches.
- [ConstBERT](https://vectordb.ever.works/ja/items/constbert) — Novel approach to reduce storage footprint of multi-vector retrieval by encoding each document with a fixed, smaller set of learned embeddings. Reduces index sizes by over 50% compared to ColBERT while retaining most effectiveness.
- [FastPLAID](https://vectordb.ever.works/ja/items/fastplaid) — Optimized implementation of PLAID index for fast ColBERT retrieval, providing 10x storage compression and sub-200ms latency. Default index backend for PyLate library, enabling efficient multi-vector late interaction retrieval.
- [XTR](https://vectordb.ever.works/ja/items/xtr) — ConteXtualized Token Retriever that introduces a novel objective function encouraging the model to retrieve the most important document tokens first. Enables ranking 4,000x cheaper than ColBERT's refinement stage with state-of-the-art performance.
- [ColBERT and Late Interaction](https://vectordb.ever.works/ja/items/colbert-and-late-interaction) — Multi-vector retrieval architecture where queries and documents are represented by multiple vectors enabling fine-grained matching and improved retrieval quality through late interaction scoring.
- [ColBERT](https://vectordb.ever.works/ja/items/colbert) — State-of-the-art late interaction retrieval model that produces multi-vector token-level representations, enabling efficient and effective passage search with rich contextual understanding.

---

_Canonical page: https://vectordb.ever.works/ja/tags/multi-vector_