Business information pushing method and system based on knowledge graph and large language model
By injecting business knowledge graphs into a large language model and performing graph retrieval, the problem of inaccurate responses from the large language model in complex business processes is solved, enabling efficient business information delivery and rapid decision support.
CN121810015BActive Publication Date: 2026-06-05GUANGZHOU BAINA SOFTWARE TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GUANGZHOU BAINA SOFTWARE TECH CO LTD
- Filing Date
- 2026-03-12
- Publication Date
- 2026-06-05
AI Technical Summary
Technical Problem
Existing large language models lack deep domain knowledge in complex business processes, resulting in inaccurate responses and difficulty in quickly adapting to the needs of non-expert users, thus affecting business execution efficiency.
Method used
Construct a collaborative system based on knowledge graphs and large language models. By injecting business knowledge bases and performing graph retrieval and multi-hop graph reasoning, accurate business information is generated, and knowledge is updated through a closed-loop mechanism.
Benefits of technology
It improves the accuracy and reliability of responses from large language models, lowers the learning threshold for non-expert users, and enhances business response speed and overall operational efficiency.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN121810015B_ABST
Abstract
The application provides a business information pushing method and system based on a knowledge graph and a large language model, wherein the method comprises the following steps: acquiring a business knowledge graph of a target business and a corresponding business knowledge base; injecting the business knowledge base into a preset large language model to acquire a corresponding business guidance large language model; receiving question information of a client, acquiring corresponding graph query statements through the business guidance large language model; acquiring evidence subgraphs associated with the graph query statements based on the graph query statements; inputting the question information, the evidence subgraphs and text summary information corresponding to the evidence subgraphs as prompt contexts into the business guidance large language model to generate at least one piece of business information corresponding to the question information and push the business information to the client. Therefore, the application can effectively improve the reply generation accuracy of the large language model and improve the operation processing efficiency of the target business.
Need to check novelty before this filing date? Find Prior Art