Platform Engineering
Governance
Snowflake-technologies-quantyca
Scopri

Overview

Information manifests in various forms, some structured, such as tables and application logs, others unstructured like text documents and multimedia content. Thanks to innovations in AI and machine learning, integration models have been developed to convert various types of data into vectors, providing them with meaning and context.

Vector databases are commonly employed for various vector search cases, including visual, semantic, and multimodal search. The process begins with text conversion into embeddings, numerical vectors encoding the semantics and meaning of words. These vectors are then placed in a multidimensional space, where each dimension represents a specific semantic meaning. This approach enables the search for similar information based on the proximity of points in the multidimensional space: vector search methods are paving the way for numerous revolutionary use cases and unique experiences for users.

The versatility of vector databases manifests through various use cases, including:

Semantic search enhances search results by considering user intent and query context, providing a more relevant and meaningful set of results compared to traditional keyword matching. This technique is implemented by semantic search engines, which analyze not only the search terms but also the context surrounding the query.

Visual search allows users to retrieve information by utilizing images as queries instead of relying on text or basic keywords. This technique revolutionizes the search process, offering a more intuitive and visually-driven way for users to find relevant information.

Multimodal search emerges as a new approach to information retrieval, specifically designed to simultaneously handle keywords, semantic context, and data (structured and unstructured). Its aim is to provide a holistic, comprehensive, and entirely personalized search experience.

Vector databases can enhance the capabilities of Large Language Models (LLMs) by serving as an external memory. For instance, they can function as technological components providing external knowledge to LLMs, aiming to mitigate.

Capabilities

OpenSearch is a highly scalable and flexible open-source service that provides low-latency search capabilities. On the other hand, Amazon OpenSearch Service is AWS’s fully managed service based on OpenSearch, designed to operate entirely in the cloud with high scalability and operational efficiency. Notably, Amazon OpenSearch offers vector database functionality, enabling the use cases described earlier.

Semantic search enhances the relevance of search results, even when using natural language queries like “I would like a sparkling red dress for an ’80s party.” Amazon OpenSearch leverages the vector database to retrieve the most relevant documents for the request, considering both the context and semantics of the query. Additionally, it can seamlessly integrate with other AWS services for embedding construction, such as deploying a Text-to-Embedding model on Amazon SageMaker or utilizing Amazon Titan services.

Visual search enables users to perform search queries using rich multimedia content, such as images. Its implementation is similar to semantic search: vector embeddings are created from images in the knowledge base, and the OpenSearch service is queried with a vector (encoding of the search query). The distinction from semantic search lies in the use of an external embedding model (for example, a Convolutional Neural Network (CNN) hosted on Amazon SageMaker or an Amazon Titan model) to convert images into vectors.

Amazon OpenSearch has introduced multimedia support for Neural Search, reducing the need for integration with external embedding models and favoring the use of multimodal APIs for text and images provided by Amazon Bedrock. With multimodal support, queries can be made on OpenSearch using images, text, or both. This functionality allows searching for images by describing their visual characteristics, discovering other similar images using a visual knowledge base, or combining hybrid text and image search to find matches from both semantic and visual perspectives.

Retrieval Augmented Generation (RAG) is a method of generative AI that leverages Large Language Models (LLMs) to develop conversational experiences for users. However, generative language models, when used in isolation, may lead to hallucinations, where the model generates credible yet factually incorrect responses. The vector database service provided by Amazon OpenSearch emerges as the ideal solution to address this challenge, acting as an external knowledge base. It provides LLMs with a knowledge source derived from transforming textual documents into embeddings, aiming to significantly enhance the accuracy and coherence of responses.

Partnership

As an AWS Partner, we bring our expertise in cloud data management processes, leveraging the flexibility, scalability, and reliability of the AWS Glue service.

Our consulting services:

  • Consulting to start with new Cloud Native projects 
  • Assessment of existing solutions and Data Platform migrations 
  • Data Management landing zone design and implementation for multi-account and multi-region management 
  • Support on developing data integration pipelines 
  • Maintenance of Cloud environments 

Use Cases

Need personalised advice? Contact us to find the best solution!

This field is for validation purposes and should be left unchanged.

Join the Quantyca team, let's be a team!

We are always looking for talented people to join the team, discover all our open positions.

SEE ALL VACANCIES