An ai education data governance and intelligent service method and system
By combining asymmetric spatiotemporal feature extraction networks and graph neural networks, the problem of deep fusion of multi-source heterogeneous educational data is solved, enabling real-time quality assessment and repair, generating standardized fused data representations, and outputting accurate teaching intervention strategies. This solves the data silo and real-time service problems in existing technologies and improves the decision-making accuracy of intelligent services.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XIAMEN INFORMATION SCHOOL
- Filing Date
- 2026-05-22
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies cannot deeply integrate unstructured audio and video streams with structured business data when processing multi-source heterogeneous educational data, resulting in data silos. Furthermore, they struggle to support millisecond-level instant teaching intervention services when dealing with real-time high-concurrency streaming data. Static governance rules cannot adapt to different teaching scenarios, leading to accidental data deletion or low-quality data residues, which affects the accuracy of intelligent service decisions.
By combining an asymmetric spatiotemporal feature extraction network and a graph neural network, a unified semantic space mapping of multi-source heterogeneous data is achieved. A dynamic quality assessment engine performs real-time quality assessment and cleaning. By combining an adaptive quality assessment function and the topology aggregation mechanism of the graph neural network, a standardized fusion data representation is generated, and an intervention strategy is output based on an intelligent service inference model.
It achieves deep integration of multi-source heterogeneous data, real-time quality assessment and repair, outputs highly accurate intervention strategies, reduces decision-making bias, forms a closed loop of data governance and teaching applications, and ensures millisecond-level service level agreements.
Smart Images
Figure CN122415290A_ABST