一种基于深度语义匹配的知识库内容精准生成方法及系统
By combining deep semantic matching and knowledge graph association with graph neural networks for multi-hop reasoning and fine-grained verification, the problems of insufficient semantic matching and inconsistent generated content in existing knowledge base construction are solved, and a knowledge base with high accuracy and reliability is generated.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TUGUAN (TIANJIN) DIGITAL TECH CO LTD
- Filing Date
- 2026-04-07
- Publication Date
- 2026-07-17
AI Technical Summary
Existing knowledge base construction technologies suffer from insufficient semantic matching accuracy, poor knowledge fusion and coordination, lack of structured knowledge guidance, and illusion problems when generating content using large language models, resulting in insufficient accuracy and consistency of the generated content.
We employ a deep semantic matching approach, constructing a multimodal vector index through semantic block processing and knowledge graph association. We combine this with graph neural networks for multi-hop reasoning and refined semantic verification and fusion, filtering out high-confidence candidate knowledge fragments and constructing structured generation prompts. The input is then used to generate accurate content from a large language model.
It significantly improves the accuracy and consistency of knowledge base generation, reduces the illusion rate of generated content, and enhances the reliability and controllability of generated content.
Smart Images
Figure CN121979905B_ABST