基于联邦学习与知识图谱的材料全程追溯方法
By using federated learning and knowledge graphs to generate unified material identifiers and time-series knowledge graphs, the problem of unified identification and association of multi-stage and multi-subject data has been solved, enabling reliable traceability and lifespan risk assessment throughout the entire lifecycle.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA RAILWAY CONSTR ENG GRP FOURTH CONSTR CO LTD
- Filing Date
- 2026-04-30
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, the lack of a unified identification and semantic mapping mechanism for data throughout the entire life cycle of materials makes it difficult to connect the traceability chain. Data barriers between multiple units and unstable communication make it difficult to centrally model and lack quantifiable basis for maintenance and replacement.
By employing a federated learning and knowledge graph-based approach, a unified identifier for materials is generated, a local temporal knowledge graph is constructed, data fields are standardized, and cross-entity privacy-preserving collaborative modeling and controllable completion of missing data are achieved through generative missing data completion and asynchronous federated weighted training.
It achieves full-chain traceability and consistent correlation across entities, improves data availability and credibility, and provides quantifiable basis for life risk assessment and maintenance/replacement.
Smart Images
Figure CN122114972B_ABST