基于多层误差归因的知识图谱自适应修复方法
By adopting a knowledge graph adaptive repair method with multi-layer error attribution, this method solves the problems of decoupling evaluation and reasoning tasks, single dimension of defect simulation, and difficulty in decoupling error sources in knowledge graph evaluation and repair. It achieves fine-grained evaluation and automated closed loop, improving evaluation accuracy and operation and maintenance efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SICHUAN UNIV JINCHENG INST
- Filing Date
- 2026-05-14
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies in knowledge graph quality assessment and graph retrieval enhancement generation systems suffer from problems such as decoupling of assessment and reasoning tasks, single dimension of defect simulation, difficulty in decoupling error sources, and lack of assessment and repair closed loop. As a result, the assessment results are difficult to reflect the actual support capabilities and there is a lack of clear direction for optimization.
An adaptive repair method based on knowledge graphs with multi-layer error attribution is adopted. By designing a multi-dimensional controlled defect injection system, a three-layer progressive evaluation and attribution mechanism is constructed to identify the parameterized knowledge masking effect and realize an automatic repair closed loop guided by evaluation, generating structured repair scripts and control instructions.
It achieves fine-grained evaluation and decoupling for reasoning tasks, improves evaluation accuracy and system reliability, provides quantitative decision-making basis for knowledge graph maintenance, realizes an automated closed loop from evaluation to governance, and significantly improves the efficiency and intelligence level of knowledge graph operation and maintenance.
Smart Images
Figure CN122242692B_ABST