Multi-source knowledge processing method and device, computer equipment and storage medium
By constructing a multi-source knowledge graph and performing knowledge extraction, fusion, and updating, the problems of knowledge fragmentation and insufficient understanding of large language models in enterprise-level knowledge management systems are solved, enabling more accurate code generation and business analysis, and improving the efficiency and consistency of knowledge management.
CN121706910APending Publication Date: 2026-03-20YGSOFT INC
View PDF 0 Cites 0 Cited by
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-18
- Publication Date
- 2026-03-20
Smart Images

Figure CN121706910A_ABST
Abstract
The invention discloses a multi-source knowledge processing method and device, computer equipment and a storage medium, by introducing an automatic knowledge extraction, fusion and updating mechanism, the cost of manual arrangement and maintenance of a knowledge base is remarkably reduced, and the timeliness and consistency of knowledge updating are improved. And meanwhile, the fused cross-domain association knowledge graph provides multi-dimensional and structured context information for the large model, so that the large model can output a result which is more suitable for the actual situation of an enterprise in scenes such as code generation, business analysis and demand understanding. Through unified representation, mixed retrieval and incremental updating of multi-source heterogeneous knowledge, the method can flexibly adapt to technology stacks and business fields of different enterprises. According to the method, systematic knowledge extraction and unified modeling are carried out on the enterprise codes, the business data and the document data, the multi-source knowledge graph is constructed and fused, and deep association of an enterprise technology implementation layer and a business logic layer is realized, so that business requirements and technology implementation are matched more accurately.
Need to check novelty before this filing date? Find Prior Art