Power zero sample entity relationship extraction method and system based on large language model
By applying the zero-sample learning method of large language models and multi-round dialogue task decomposition in the power field, combined with syntax enhancement strategies, the ChatRE model was constructed, which solved the problem of excessive dependence on labeled data and insufficient processing capabilities in the existing technology, and achieved efficient and accurate entity relationship extraction.
Patent Information
- Application Number
- CN202411821089.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-11
- Publication Date
- 2025-03-28
AI Technical Summary
The prior art requires a large amount of labeled data in the extraction of entity relationships in the power field, and the processing capacity of complex text is limited, resulting in the impact of the accuracy of extraction.
A zero-sample learning method based on a large language model is adopted, combined with the entity relationship task decomposition and syntax enhancement strategy of multiple rounds of dialogue, a new entity relationship extraction model ChatRE is built to realize entity relationship extraction without a large amount of labeled data.
It significantly reduces the time and labor cost of data annotation, improves the accuracy and efficiency of extraction in complex texts, and can adapt to different text scenarios and changing text types.