Combine keyword and semantic search for text and images using Amazon Bedrock and Amazon OpenSearch Service

This article discusses how to combine keyword and semantic search for text and images using Amazon Bedrock and Amazon OpenSearch Service, enabling more precise and contextual product search experiences.

• Semantic search uses embeddings to capture query meaning, allowing more flexible searching across text and images
• Keyword search provides precise matching of specific product names and attributes
• Hybrid search combines keyword and semantic search, improving result quality by 8-12%
• Uses Amazon Titan Multimodal Embeddings G1 to generate embeddings for text and images
• Demonstrates a solution that allows users to search using both text and image inputs

The solution provides a comprehensive approach to search functionality, enabling more intuitive and accurate product discovery for e-commerce platforms.