Vector database explained

A vector database, vector store or vector search engine is a database that can store vectors (fixed-length lists of numbers) along with other data items. Vector databases typically implement one or more Approximate Nearest Neighbor (ANN) algorithms,[1] [2] so that one can search the database with a query vector to retrieve the closest matching database records.

Vectors are mathematical representations of data in a high-dimensional space. In this space, each dimension corresponds to a feature of the data, with the number of dimensions ranging from a few hundred to tens of thousands, depending on the complexity of the data being represented. A vector's position in this space represents its characteristics. Words, phrases, or entire documents, as well as images, audio, and other types of data, can all be vectorized.[3]

These feature vectors may be computed from the raw data using machine learning methods such as feature extraction algorithms, word embeddings[4] or deep learning networks. The goal is that semantically similar data items receive feature vectors close to each other.

Vector databases can be used for similarity search, semantic search, multi-modal search, recommendations engines, large language models (LLMs), object detection, etc.[5]

Vector databases are also often used to implement retrieval-augmented generation (RAG), a method to improve domain-specific responses of large language models. The retrieval component of a RAG can be any search system, but is most often implemented as a vector database. Text documents describing the domain of interest are collected, and for each document or document section, a feature vector (known as an "embedding") is computed, typically using a deep learning network, and stored in a vector database. Given a user prompt, the feature vector of the prompt is computed, and the database is queried to retrieve the most relevant documents. These are then automatically added into the context window of the large language model, and the large language model proceeds to create a response to the prompt given this context.[6]

Techniques

The most important techniques for similarity search on high-dimensional vectors include:

and combinations of these techniques.

In recent benchmarks, HNSW-based implementations have been among the best performers.[7] Conferences such as the International Conference on Similarity Search and Applications, SISAP and the Conference on Neural Information Processing Systems (NeurIPS) host competitions on vector search in large databases.

Implementations

Name License
Aerospike[8] [9] Proprietary
AllegroGraph[10] [11] Proprietary (Managed Service)
Apache Cassandra[12] [13] Apache License 2.0
Chroma[14] [15] Apache License 2.0[16]
Azure Cosmos DB[17] Proprietary (Managed Service)
Couchbase[18] [19] BSL 1.1[20]
Elasticsearch[21] Server Side Public License, Elastic License[22]
HDF5 Query Indexing[23] BSD 3-Clause[24]
Lantern[25] BSL 1.1[26]
LlamaIndex[27] MIT License[28]
MariaDB[29] [30] GPL v2[31]
Marqo[32] Apache License 2.0
Milvus[33] [34] Apache License 2.0
MongoDB Atlas[35] Server Side Public License (Managed service)
Neo4j[36] [37] GPL v3 (Community Edition)[38]
ObjectBox[39] Apache License 2.0[40]
OpenSearch[41] [42] Apache License 2.0[43]
Oracle[44] Proprietary (Managed Service or License)
Pinecone[45] Proprietary (Managed Service)
Postgres with pgvector[46] PostgreSQL License[47]
Qdrant[48] Apache License 2.0[49]
Redis Stack[50] [51] Redis Source Available License[52]
Snowflake[53] Proprietary (Managed Service)
SurrealDB[54] BSL 1.1[55]
Typesense[56] GPL v3 (Community Edition)[57]
Vespa[58] Apache License 2.0[59]
Weaviate[60] BSD 3-Clause[61]

External links

Notes and References

  1. Web site: Roie Schwaber-Cohen . What is a Vector Database & How Does it Work . 18 November 2023 . Pinecone.
  2. Web site: What is a vector database . 18 November 2023 . Elastic.
  3. Web site: 2023-12-26 . Vector database . 2024-01-11 . learn.microsoft.com .
  4. Web site: Evan Chaki . 2023-07-31 . What is a vector database? . Microsoft . A vector database is a type of database that stores data as high-dimensional vectors, which are mathematical representations of features or attributes..
  5. Web site: 2023-12-26 . Vector database . 2024-01-11 . learn.microsoft.com .
  6. Lewis . Patrick . Perez . Ethan . Piktus . Aleksandra . Petroni . Fabio . Karpukhin . Vladimir . Goyal . Naman . Küttler . Heinrich . Retrieval-augmented generation for knowledge-intensive NLP tasks . Advances in Neural Information Processing Systems 33 . 2020 . 9459–9474. 2005.11401 .
  7. Book: Aumüller . Martin . Bernhardsson . Erik . Faithfull . Alexander . ANN-Benchmarks: A Benchmarking Tool for Approximate Nearest Neighbor Algorithms . Lecture Notes in Computer Science . 2017 . 10609 . Beecks . Christian . Borutta . Felix . Kröger . Peer . Seidl . Thomas . Similarity Search and Applications . https://link.springer.com/chapter/10.1007/978-3-319-68474-1_3 . en . Cham . Springer International Publishing . 34–49 . 10.1007/978-3-319-68474-1_3 . 978-3-319-68474-1. 1807.05614 .
  8. Web site: 2024-05-07 . Aerospike Recognized by Independent Research Firm Among Notable Vendors in Vector Databases Report. 2024-08-01 . Morningstar . en-US.
  9. Web site: 2024-04-04 . Aerospike raises $109M for its real-time database platform to capitalize on the AI boom. 2024-08-01 . TechCrunch . en-US.
  10. Web site: 2023-12-29 . AllegroGraph 8.0 Incorporates Neuro-Symbolic AI, a Pathway to AGI. 2024-06-06 . TheNewStack . en-US.
  11. Web site: 2024-01-18 . Franz Inc. Introduces AllegroGraph Cloud: A Managed Service for Neuro-Symbolic AI Knowledge Graphs. 2024-06-06 . Datanami . en-US.
  12. Web site: 2023-09-22 . 5 Hard Problems in Vector Search, and How Cassandra Solves Them. 2023-09-22 . TheNewStack . en-US.
  13. Web site: Vector Search quickstart. 2023-11-21 .
  14. Web site: Palazzolo . Stephanie . Vector database Chroma scored $18 million in seed funding at a $75 million valuation. Here's why its technology is key to helping generative AI startups.. 2023-11-16 . Business Insider . en-US.
  15. Web site: MSV . Janakiram . 2023-07-28 . Exploring Chroma: The Open Source Vector Database for LLMs. 2023-11-16 . The New Stack . en-US.
  16. Web site: chroma/LICENSE at main · chroma-core/chroma. GitHub. en.
  17. Web site: 26 December 2023. Vector database. 2024-01-10 . learn.microsoft.com.
  18. Web site: 2023-08-30 . Couchbase aims to boost developer database productivity with Capella IQ AI tool. VentureBeat. en-US.
  19. Web site: 2023-12-06 . Investor Presentation Third Quarter Fiscal 2024. Couchbase Investor Relations. en-US.
  20. Web site: Anderson . Scott . 2021-03-26 . Couchbase Adopts BSL License . 2024-02-14 . The Couchbase Blog . en-US.
  21. News: Kerner. Sean . Elasticsearch Relevance Engine brings new vectors to generative AI. 18 November 2023 . . 23 May 2023.
  22. Web site: elasticsearch/LICENSE.txt at main · elastic/elasticsearch. GitHub. en.
  23. Web site: HDF5 Query Indexing. . 3 May 2024 . 27 Sep 2019.
  24. Web site: HDFGroup/COPYING at master · HDFGroup/hdf5. 2023-10-29 . GitHub . en.
  25. Web site: 2024-04-05 . Lantern . 2024-04-05 . en-US.
  26. Web site: lantern/LICENSE at main /lanterndata/lantern . 2024-04-10 . GitHub. en.
  27. Web site: Wiggers . Kyle . 2023-06-06 . LlamaIndex adds private data to large language models. 2023-10-29 . TechCrunch . en-US.
  28. Web site: llama_index/LICENSE at main · run-llama/llama_index. 2023-10-29 . GitHub . en.
  29. Web site: MariaDB Vector . 2024-07-30 . MariaDB.org . en-US.
  30. Web site: Vector search in old and modern databases . 2024-07-30 . manticoresearch.com . en-us.
  31. Web site: Licensing FAQ . 2024-07-30 . MariaDB KnowledgeBase.
  32. Web site: Sawers . Paul . 2023-08-16 . Meet Marqo, an open source vector search engine for AI applications . 2024-08-20 . TechCrunch . en-US.
  33. Web site: Open Source Vector Database – Milvus – LFAI & DATA . 29 October 2023.
  34. Web site: Liao . Ingrid Lunden and Rita . 2022-08-24 . Zilliz raises $60M, relocates to SF. 2023-10-29 . TechCrunch . en-US.
  35. Web site: 2023-06-22 . Introducing Atlas Vector Search: Build Intelligent Applications with Semantic Search and AI Over Any Type of Data. MongoDB. en-US.
  36. Web site: 2023-08-22. Neo4j enhances its graph database with vector search. itbrief. en-AU.
  37. Web site: Vector search indexes. neo4j. en-US.
  38. Web site: Neo4j licensing.
  39. Web site: 2024-07-03. Top Fifteen Vector Databases. 2024-07-03 . db-engines.com . en-US.
  40. Web site: ObjectBox Java license. github.
  41. Web site: 2023-08-02. Using OpenSearch as a Vector Database. 2024-02-07 . OpenSearch.org . en-US.
  42. Web site: 2023-07-26 . AWS debuts new AI-powered data management and analysis tools . 2024-02-07 . SiliconANGLE . en-US.
  43. Web site: OpenSearch license. github.
  44. Web site: Hook(1) and Priyadarshi(2) . Doug(1) and Ranjan(2) . May 2, 2024 . Oracle Announces General Availability of AI Vector Search in Oracle Database 23ai . live . July 9, 2024 . oracle.
  45. Web site: 2023-07-14 . Pinecone leads 'explosion' in vector databases for generative AI. 2023-10-29 . VentureBeat . en-US.
  46. Web site: pgvector. 2023-11-27 . GitHub . en-US.
  47. Web site: pgvector/License. 2023-11-27 . GitHub . en-US.
  48. Web site: Sawers . Paul . 2023-04-19 . Qdrant, an open-source vector database startup, wants to help AI developers leverage unstructured data. 2023-10-29 . TechCrunch . en-US.
  49. Web site: qdrant/LICENSE at master · qdrant/qdrant. 2023-10-29 . GitHub . en.
  50. Web site: Using Redis as a Vector Database with OpenAI OpenAI Cookbook . 2024-02-10 . cookbook.openai.com . en.
  51. Web site: Redis as a vector database quick start guide. 2024-01-31 . Redis . en.
  52. Web site: Search and query . 2024-02-10 . Redis . en.
  53. Web site: 2024-05-17 . Vector data type and vector similarity functions — General Availability. 2024-05-17 . Snowflake . en-US.
  54. Web site: Wiggers . Kyle . 2023-01-04 . SurrealDB raises $6M for its database-as-a-service offering. 2024-01-19 . TechCrunch . en-US.
  55. Web site: SurrealDB License FAQs The ultimate multi-model database . 2024-02-14 . SurrealDB . en.
  56. Web site: Martinez . Miguel . 2024-06-20 . Typesense Homepage. 2024-06-20 . Typesense . en-US.
  57. Web site: Typesense licensing.
  58. News: Riley . Duncan . Yahoo spins off AI scaling engine Vespa as an independent company. 18 November 2023 . siliconANGLE . 4 October 2023.
  59. Web site: vespa/LICENSE at master · vespa-engine/vespa. GitHub. en.
  60. Web site: 2023-04-21 . Weaviate reels in $50M for its AI-optimized vector database. 2023-10-29 . SiliconANGLE . en-US.
  61. Web site: weaviate/LICENSE at master · weaviate/weaviate. 2023-10-29 . GitHub . en.

This article is licensed under the GNU Free Documentation License. It uses material from the Wikipedia article "Vector database".

Except where otherwise indicated, Everything.Explained.Today is © Copyright 2009-2024, A B Cryer, All Rights Reserved. Cookie policy.