一种层级注意力图卷积网络关系抽取方法
By using a hierarchical attention graph convolutional network, combining sentence-level and layer-level attention mechanisms, the problem of low accuracy in entity relation extraction in existing technologies is solved, achieving effective understanding of complex sentence structures and high-precision extraction of entity relations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CIVIL AVIATION UNIV OF CHINA
- Filing Date
- 2023-09-27
- Publication Date
- 2026-07-17
AI Technical Summary
Existing entity relation extraction methods based on attention and graph neural networks suffer from problems such as uncertain entity positions and information loss when dealing with complex sentence structures, resulting in low accuracy of entity relation extraction.
A hierarchical attention graph convolutional network is adopted. By introducing a hierarchical structure and combining graph convolutional network relation extraction methods, word embedding representation matrices and adjacency matrices are used to extract vector features through sentence-level and layer-level attention mechanisms, thereby enhancing the understanding of complex sentence structures.
By introducing a hierarchical attention mechanism, the accuracy of entity relation extraction is improved. In particular, when dealing with complex sentence structures, it can better capture the dependencies between entities, thereby improving the accuracy and generalization ability of entity relation extraction.
Smart Images
Figure CN117252187B_ABST