Knowledge Bases for Amazon Bedrock now supports hybrid search
Hybrid Search is now generally available in Knowledge Bases for Bedrock for OpenSearch Serverless. Retrieval augmented generation (RAG) applications typically use semantic search, which relies on semantic vectors to search unstructured text. These vectors are created from machine learning models to capture contextual and linguistic meaning within the data to answer human-like questions. Hybrid search combines this semantic search with a text based search to improve the relevance of retrieved results, especially for keyword searches.
Hybrid Search is now generally available in Knowledge Bases for Bedrock for OpenSearch Serverless. Retrieval augmented generation (RAG) applications typically use semantic search, which relies on semantic vectors to search unstructured text. These vectors are created from machine learning models to capture contextual and linguistic meaning within the data to answer human-like questions. Hybrid search combines this semantic search with a text based search to improve the relevance of retrieved results, especially for keyword searches.
Hybrid search works by making two search queries to the semantic and text retrieval systems, and then combines the results through intelligent ranking. This increases the relevancy of results by casting a wider search net and retrieving relevant documents which may not necessarily contain the appropriate semantic structure of the document. You can enable hybrid search through the Knowledge Bases SDK or in the console. Select hybrid search as your preferred search option within Knowledge Bases, or choose the default search option where AWS will intelligently determine the best search methods to use for your data. Hybrid search is available in US East (N. Virginia) and US West (Oregon) AWS Regions. To learn more, refer to Knowledge Bases for Amazon Bedrock documentation. To get started, visit the Amazon Bedrock console.