Generative AI applications, where accuracy and reliability are critical, often use an approach called RAG that leverages external data sources to improve the accuracy of the generated AI responses and make use of expertise and up-to-date information. Fujitsu Enterprise Postgres provides knowledge data management capabilities for using Fujitsu Enterprise Postgres as an external data source in the RAG approach.

Knowledge in the RAG approach is expressed in text and graph form, and vector data is also leveraged for fast retrieval. Knowledge data management allows you to centralize the management of knowledge data such as vectors and graphs in Fujitsu Enterprise Postgres. In addition to eliminating the need for a dedicated database for each data type, you can protect your knowledge data with the reliable and secure features of Fujitsu Enterprise Postgres.

In addition, the knowledge data management supports:

• Vector data management feature

  This feature stores and retrieves vector data.

• Semantic text search and automatic vector conversion

  Text Semantic Search is a feature that retrieves semantically similar text data from Fujitsu Enterprise Postgres without requiring the application to process vector data.

  Automated vectorization is a feature to automatically generate and manage semantic vectors used for semantic search of text, based on declarative specifications.

• Graph management feature

  This feature stores and retrieves graph data.

• LangChain linkage

  Knowledge data stored in Fujitsu Enterprise Postgres is available to RAG-based applications using LangChain.