A Knowledge Graph and Credibility Assessment-Based Intelligent Question Answering System for Power Grid Applications
Depin Feng, Shanzhi Yu, Xiang Song, Qing Xu, Zheng Wang, Wenlei Xu
Problems Identified (2)
Power grid knowledge management inadequacy: Traditional knowledge management tools are inadequate for power grid tasks such as dispatch support, fault diagnosis, and regulatory procedure interpretation amid increasing grid complexity and diversified operational data.
Power grid knowledge management inadequacy: Traditional knowledge management tools are inadequate for power grid tasks such as dispatch support, fault diagnosis, and regulatory procedure interpretation amid increasing grid complexity and diversified operational data.
Proposed Solutions (5)
KG credibility-based QA system: The paper proposes KG-CA-QA, an intelligent question answering system for the power grid domain integrating knowledge graphs and credibility assessment.
Multi-source power grid knowledge graph: The system constructs a multi-source knowledge graph for unified modeling and querying of power grid entities and relationships.
RAG with credibility scoring: The system combines retrieval-augmented generation with credibility scoring, semantic understanding, knowledge graph reasoning, and source authority to produce fact-based, interpretable, credibility-scored answers.
Dynamic knowledge updating: The system includes a dynamic knowledge updating module to maintain the timeliness of the knowledge base.
KG credibility-based QA system: The paper proposes KG-CA-QA, an intelligent question answering system for the power grid domain integrating knowledge graphs and credibility assessment.
Results (3)
Practical engineering value:
Improved QA performance:
Practical engineering value:
Research Domain
Power grid intelligent question answering / knowledge management