Cross-source material data semantic alignment method and device

By preprocessing, extracting semantic features, and embedding high-dimensional representations of multi-source heterogeneous material data, and combining attribute alignment with a material terminology knowledge graph, the problems of inconsistent naming and semantics among material data from different sources are solved. This enables unified management and efficient utilization of cross-source data, and improves the organization and searchability of the data.

CN122433746APending Publication Date: 2026-07-21RENMIN UNIVERSITY OF CHINA +1
0 Cites 0 Cited by

Patent Information

Application Number
CN202610750591.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-05-28
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Material data from different sources suffers from inconsistent naming, inconsistent expression, different semantic granularity, and strong context dependence due to differences in data generation mechanisms, recording habits, and application purposes. This seriously restricts the unified management and efficient utilization of material data.

Method used

By preprocessing multi-source heterogeneous material data, extracting a unified semantic expression sequence of standardized data items, performing high-dimensional semantic embedding representation and clustering, and combining it with a material terminology knowledge graph for attribute alignment and relation extraction, unified semantic modeling and standardized expression of cross-source data are achieved.

Benefits of technology

It achieves unified semantic modeling and standardized expression of cross-source material data, improves data integration, organization, retrieval and reuse capabilities, reduces the impact of terminological ambiguity and conceptual fragmentation, has incremental semantic alignment capabilities, and is suitable for material database construction, experimental data governance and intelligent analysis.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the field of material data processing, discloses a cross-source material data semantic alignment method and device, and constructs a complete technical path from multi-source material data collection, semantic feature extraction, semantic embedding representation, object entity alignment, attribute alignment to relation organization. The method can be oriented to multi-source heterogeneous material data of literature, experiment, calculation, industrial production and the like, semantic representation learning and intelligent analysis capabilities are fully combined, unified semantic modeling and alignment are realized in an attribute layer and a relation layer, then semantic barriers among cross-source data are eliminated, unified semantic modeling and standardized expression of the cross-source material data are realized, and the data processing efficiency is improved. And the integration, organization, retrieval and reuse capabilities of material data are improved.
Need to check novelty before this filing date? Find Prior Art