Rediger

Del via

Vectors in Azure AI Search

  • Artikkel

Vector search is an approach in information retrieval that supports indexing and query execution over numeric representations of content. Because the content is numeric rather than plain text, matching is based on vectors that are most similar to the query vector, which enables matching across:

  • semantic or conceptual likeness ("dog" and "canine", conceptually similar yet linguistically distinct)
  • multilingual content ("dog" in English and "hund" in German)
  • multiple content types ("dog" in plain text and a photograph of a dog in an image file)

This article provides a high-level introduction to vectors in Azure AI Search. It also explains integration with other Azure services and covers terminology and concepts related to vector search development.

We recommend this article for background, but if you'd rather get started, follow these steps:

You could also begin with the vector quickstart or the code samples on GitHub.

What scenarios can vector search support?

Scenarios for vector search include:

  • Similarity search. Encode text using embedding models such as OpenAI embeddings or open source models such as SBERT, and retrieve documents with queries that are also encoded as vectors.

  • Search across different content types (multimodal). Encode images and text using multimodal embeddings (for example, with OpenAI CLIP or GPT-4 Turbo with Vision in Azure OpenAI) and query an embedding space composed of vectors from both content types.

  • Hybrid search. In Azure AI Search, hybrid search refers to vector and keyword query execution in the same request. Vector support is implemented at the field level, with an index containing both vector fields and searchable text fields. The queries execute in parallel and the results are merged into a single response. Optionally, add semantic ranking for more accuracy with L2 reranking using the same language models that power Bing.

  • Multilingual search. Providing a search experience in the users own language is possible through embedding models and chat models trained in multiple languages. If you need more control over translation, you can supplement with the multi-language capabilities that Azure AI Search supports for nonvector content, in hybrid search scenarios.

  • Filtered vector search. A query request can include a vector query and a filter expression. Filters apply to text and numeric fields, and are useful for metadata filters, and including or excluding search results based on filter criteria. Although a vector field isn't filterable itself, you can set up a filterable text or numeric field. The search engine can process the filter before or after the vector query executes.

  • Vector database. Azure AI Search stores the data that you query over. Use it as a pure vector store any time you need long-term memory or a knowledge base, or grounding data for Retrieval Augmented Generation (RAG) architecture, or any app that uses vectors.

Vector support includes indexing, storing, and querying of vector embeddings from a search index.

The following diagram shows the indexing and query workflows for vector search.

Architecture of vector search workflow.

On the indexing side, Azure AI Search takes vector embeddings and uses a nearest neighbors algorithm to place similar vectors close together in an index. Internally, it creates vector indexes for each vector field.

How you get embeddings from your source content into Azure AI Search depends on whether you want to perform the work within an Azure AI Search indexing pipeline, or externally. Azure AI Search offers integrated data chunking and vectorization in an indexer pipeline. You still provide the resources (endpoints and connection information to Azure OpenAI), but Azure AI Search makes all of the calls and handles the transitions. This approach requires an indexer, a supported data source, and a skillset that drives chunking and embedding. Otherwise, you can handle all vectorization separately, and then push prevectorized content to vector fields in a vector store.

On the query side, in your client application, you collect the query input from a user, usually through a prompt workflow. You can then add an encoding step that converts the input into a vector, and then send the vector query to your index on Azure AI Search for a similarity search. As with indexing, you can deploy the integrated vectorization to convert the question into a vector. For either approach, Azure AI Search returns documents with the requested k nearest neighbors (kNN) in the results.

Azure AI Search supports hybrid scenarios that run vector and keyword search in parallel, returning a unified result set that often provides better results than just vector or keyword search alone. For hybrid, vector and nonvector content is ingested into the same index, for queries that run side by side.

Availability and pricing

Vector search is available as part of all Azure AI Search tiers in all regions at no extra charge.

Newer services created after April 3, 2024 support higher quotas for vector indexes.

Vector search is available in:

Note

Some older search services created before January 1, 2019 are deployed on infrastructure that doesn't support vector workloads. If you try to add a vector field to a schema and get an error, it's a result of outdated services. In this situation, you must create a new search service to try out the vector feature.

Azure AI Search is deeply integrated across the Azure AI platform. The following table lists several that are useful in vector workloads.

Product Integration
Azure AI Studio In the chat with your data playground, Add your own data uses Azure AI Search for grounding data and conversational search. This is the easiest and fastest approach for chatting with your data.
Azure OpenAI Azure OpenAI provides embedding models and chat models. Demos and samples target the text-embedding-ada-002. We recommend Azure OpenAI for generating embeddings for text.
Azure AI Services Image Retrieval Vectorize Image API(Preview) supports vectorization of image content. We recommend this API for generating embeddings for images.
Azure data platforms: Azure Blob Storage, Azure Cosmos DB You can use indexers to automate data ingestion, and then use integrated vectorization to generate embeddings. Azure AI Search can automatically index vector data from two data sources: Azure blob indexers and Azure Cosmos DB for NoSQL indexers. For more information, see Add vector fields to a search index..

It's also commonly used in open-source frameworks like LangChain.

Vector search concepts

If you're new to vectors, this section explains some core concepts.

Vector search is a method of information retrieval where documents and queries are represented as vectors instead of plain text. In vector search, machine learning models generate the vector representations of source inputs, which can be text, images, or other content. Having a mathematic representation of content provides a common basis for search scenarios. If everything is a vector, a query can find a match in vector space, even if the associated original content is in different media or language than the query.

When searchable content is represented as vectors, a query can find close matches in similar content. The embedding model used for vector generation knows which words and concepts are similar, and it places the resulting vectors close together in the embedding space. For example, vectorized source documents about "clouds" and "fog" are more likely to show up in a query about "mist" because they're semantically similar, even if they aren't a lexical match.

Embeddings and vectorization

Embeddings are a specific type of vector representation of content or a query, created by machine learning models that capture the semantic meaning of text or representations of other content such as images. Natural language machine learning models are trained on large amounts of data to identify patterns and relationships between words. During training, they learn to represent any input as a vector of real numbers in an intermediary step called the encoder. After training is complete, these language models can be modified so the intermediary vector representation becomes the model's output. The resulting embeddings are high-dimensional vectors, where words with similar meanings are closer together in the vector space, as explained in Understand embeddings (Azure OpenAI).

The effectiveness of vector search in retrieving relevant information depends on the effectiveness of the embedding model in distilling the meaning of documents and queries into the resulting vector. The best models are well-trained on the types of data they're representing. You can evaluate existing models such as Azure OpenAI text-embedding-ada-002, bring your own model that's trained directly on the problem space, or fine-tune a general-purpose model. Azure AI Search doesn't impose constraints on which model you choose, so pick the best one for your data.

In order to create effective embeddings for vector search, it's important to take input size limitations into account. We recommend following the guidelines for chunking data before generating embeddings. This best practice ensures that the embeddings accurately capture the relevant information and enable more efficient vector search.

What is the embedding space?

Embedding space is the corpus for vector queries. Within a search index, an embedding space is all of the vector fields populated with embeddings from the same embedding model. Machine learning models create the embedding space by mapping individual words, phrases, or documents (for natural language processing), images, or other forms of data into a representation comprised of a vector of real numbers representing a coordinate in a high-dimensional space. In this embedding space, similar items are located close together, and dissimilar items are located farther apart.

For example, documents that talk about different species of dogs would be clustered close together in the embedding space. Documents about cats would be close together, but farther from the dogs cluster while still being in the neighborhood for animals. Dissimilar concepts such as cloud computing would be much farther away. In practice, these embedding spaces are abstract and don't have well-defined, human-interpretable meanings, but the core idea stays the same.

In vector search, the search engine scans vectors within the embedding space to identify vectors that are closest to the query vector. This technique is called nearest neighbor search. Nearest neighbors help quantify the similarity between items. A high degree of vector similarity indicates that the original data was similar too. To facilitate fast nearest neighbor search, the search engine performs optimizations, or employs data structures and data partitioning, to reduce the search space. Each vector search algorithm solves the nearest neighbor problems in different ways as they optimize for minimum latency, maximum throughput, recall, and memory. To compute similarity, similarity metrics provide the mechanism for computing distance.

Azure AI Search currently supports the following algorithms:

  • Hierarchical Navigable Small World (HNSW): HNSW is a leading ANN algorithm optimized for high-recall, low-latency applications where data distribution is unknown or can change frequently. It organizes high-dimensional data points into a hierarchical graph structure that enables fast and scalable similarity search while allowing a tunable a trade-off between search accuracy and computational cost. Because the algorithm requires all data points to reside in memory for fast random access, this algorithm consumes vector index size quota.

  • Exhaustive K-nearest neighbors (KNN): Calculates the distances between the query vector and all data points. It's computationally intensive, so it works best for smaller datasets. Because the algorithm doesn't require fast random access of data points, this algorithm doesn't consume vector index size quota. However, this algorithm provides the global set of nearest neighbors.

To use these algorithms, see Create a vector field for instructions on specifying the algorithm, vector profiles, and profile assignment.

Algorithm parameters that are used to initialize the index during index creation are immutable and can't be changed after the index is built. However, parameters that affect the query-time characteristics (efSearch) can be modified.

In addition, fields that specify HNSW algorithm also support exhaustive KNN search using the query request parameter "exhaustive": true. The opposite isn't true however. If a field is indexed for exhaustiveKnn, you can't use HNSW in the query because the extra data structures that enable efficient search don’t exist.

Approximate Nearest Neighbors

Approximate Nearest Neighbor (ANN) search is a class of algorithms for finding matches in vector space. This class of algorithms employs different data structures or data partitioning methods to significantly reduce the search space to accelerate query processing.

ANN algorithms sacrifice some accuracy, but offer scalable and faster retrieval of approximate nearest neighbors, which makes them ideal for balancing accuracy against efficiency in modern information retrieval applications. You can adjust the parameters of your algorithm to fine-tune the recall, latency, memory, and disk footprint requirements of your search application.

Azure AI Search uses HNSW for its ANN algorithm.

Next steps