Oracle AI Vector Search is a novel capability that allows users to search data based on the semantics or meaning of data, in addition to by the values of data, such as attribute values or keywords, as databases have traditionally supported.
A vector, or vector embedding, is a popular data structure used in AI applications. A vector is a list of numbers, generated by deep learning models from diverse data types (e.g. images, documents, videos, etc.), that encodes the semantics of the data.
Oracle AI Vector Search allows you to generate, store, index, and query vector embeddings along with other business data, using the full power of SQL. As an example, for searching documents, vector search is often considered more effective than keyword-based search, as vector search is based on the meaning and context behind the words and not the actual words themselves.
Oracle AI Vector Search allows you to combine semantic document search with searches on structured document properties. For example, in a database of technology articles, a question such as “find articles about fine-tuning Large Language Models (LLMs) for enterprise use cases that have been published in the last 5 years by a certain author and a certain publisher in a certain country”, requires searching both the article text as well as article attributes, that may be present in one or more tables.
Oracle AI Vector Search includes a collection of powerful capabilities to enable semantic search for business use cases. These features include
- New SQL operators to generate vector embeddings from unstructured data
- A new first-class VECTOR data type for storing vector embeddings
- New state-of-the-art Vector Indexes for fast approximate searches
- New SQL operators, and syntax, to easily express similarity search in business queries
- Support for the Full Generative AI pipeline including preprocessing and vectorizing data, and augmenting LLMs with business data
The VECTOR data type is fully integrated in SQL and PL/SQL, and is supported across multiple clients and programming languages, with native binding capabilities in python-oracledb, node-oracledb, JDBC, and ODP.NET drivers. This comprehensive support provides seamless vector search functionality across multiple development environments.
Oracle AI Vector Search is also fully integrated in popular 3rd party Generative-AI frameworks such as LangChain.
Oracle AI Vector Search includes native APIs to make REST callouts to LLM APIs for content generation (like text) or summarization and other operations used in the Generative-AI pipeline. These new capabilities allow seamless support for Retrieval Augmented Generation (RAG), a breakthrough generative AI technique that augments large language models (LLMs), with private business data, to deliver accurate responses to natural language questions on business data.
Benefits of Oracle AI Vector Search
Oracle Database is a leading repository of operational and enterprise data. Enterprise applications usually need to search a combination of business data and unstructured data. For example, a retail website could feature searches based on a natural language product description and a target product image, along with other filters, such as price, store location, manufacturer, and current availability. This search requires simultaneously searching unstructured catalog data (product description and image), structured catalog data (price, store location and manufacturer) as well as real time transactional data (such as current inventory).
The combination of the converged capabilities of Oracle Database, and Oracle AI Vector Search, provides several unique benefits.
Seamless Combination of AI Vector Data with Your Business Data
This is a key benefit of Oracle AI Vector Search, since it lets users run AI-powered vector similarity searches within their existing Oracle Databases instead of having to move business data to a separate vector database. Avoiding data movement can reduce complexity, improve security, and enable searches on current data.
Oracle AI Vector search allows far more powerful searches than most dedicated Vector databases, by combining sophisticated business data search with AI vector similarity search using simple, intuitive SQL and the full power of converged database – JSON, graph, text, relational, spatial, etc. all within a single query.
Powering Retrieval Augmented Generation with Business Data
Vector Databases improve interactions with LLMs since they provide scenario-specific private context to LLMs to obtain more accurate answers. This is a well-known workflow referred to as Retrieval Augmented Generation (RAG).
Oracle AI Vector Search also leverages the full power of business data to further refine LLM interactions, making use of business criteria such as security filters, business metrics and business rules, resulting in ultra-sophisticated RAG for the Enterprise.
Support for the Full Generative AI Pipeline for Business Data
Oracle AI Vector Search supports native database APIs to perform all aspects of the generative AI pipeline, from end to end, making it easier for your developers to build next-gen AI applications using your business data, directly within Oracle Database.
Unique Combination of AI Vector Search and Full Machine Learning Suite
Oracle Database offers both a full suite of in-database machine learning algorithms as well as similarity search on AI vectors. This combination enables Oracle Database to handle a very wide range of AI use cases involving machine learning actions (decisions, predictions, classification, forecasts, etc.) as well as the power of AI-based vector search. For instance, it is easy to combine inference and classification with AI Vector Search, within the same SQL query.
Proven, Enterprise-Class Scalability, Fault Tolerance and Security
Oracle Database is a leading repository of business data, and the combination of business data and semantic search is what enterprises need to implement artificial intelligence solutions. AI Vector Search is built into Oracle Database and leverages:
- Partitioning, RAC, Sharding and Exadata for proven, industrial-strength scalability.
- Extreme HA and DR technologies like Data Guard, Golden Gate, Flashback, RMAN, ZDLRA and more.
- Cutting-edge security with Oracle Advanced Security including features such as Transparent Data Encryption, Key Vault, Audit Vault, Virtual Private Database and more.
Example Use Cases
Oracle AI Vector Search enables the combination of search on semantic and business data resulting in more-accurate answers quickly, and securely. With the addition of AI Vector Search to Oracle Database, users can quickly, and easily get the benefits of artificial intelligence without sacrificing security, data integrity or performance.
Use cases include:
- Conversational AI, or Chatbots: Build AI-powered digital assistants
- Similarity Search: Match customers with products
- Content-Based Filtering: Enable personalized recommendations, locate retail items from pictures
- Natural Language Processing: Text classification and clustering SQL generation
- Data Analysis: Anomaly detection, pattern recognition
- Computer Vision: Face recognition, biometric identification, object detection
- Biomedical Research: Gene/DNA similarity research, molecular structure search
- Geographic Information Systems: Spatial analysis, map rendering
- Industrial Applications: Quality control, predictive maintenance, machinery malfunction
Summary
Oracle AI Vector Search, with Oracle Database, enables a new class of applications making it possible to transform traditional business processes by enabling semantic searches using LLMs augmented with existing business data.
- New SQL operators, and syntax, enable you to easily combine relational search, on business data, with semantic search on unstructured data
- AI-powered vector similarity searches using Oracle Database – no need to move business data to a separate vector database
- Easily utilize Retrieval Augmented Generation (RAG) to augment LLM responses with your enterprise specific content
- Enhance your applications with the enterprise-class security, scalability, and partitioning capabilities of Oracle Database
Source: oracle.com
Posts
0 comments:
Post a Comment