基于实体森林的实体语义关系联合抽取方法及系统

By using an entity forest-based approach, nested entity trees are identified and generated. Combined with a multi-head attention mechanism to extract interaction information, the complexity and accuracy issues of joint extraction of nested entity relationships are resolved, achieving efficient nested entity identification and relationship extraction.

CN115934953BActive Publication Date: 2026-07-17INST OF COMPUTING TECH CHINESE ACAD OF SCI
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
INST OF COMPUTING TECH CHINESE ACAD OF SCI
Filing Date
2022-09-29
Publication Date
2026-07-17

Smart Images

  • Figure CN115934953B_ABST
    Figure CN115934953B_ABST
Patent Text Reader

Abstract

本发明提出一种基于实体森林的实体语义关系联合抽取方法和系统，包括：获取待识别实体语义关系的语料；得到句子及其对应的词序列，对句子的词序列进行编码，得到训练语料中句子的分布式表示；句子的分布式表示进行序列标注，得到实体头部，作为实体树的根节点，以根节点为循环神经网络模型的初始状态，依次输入句子中子词至循环神经网络模型，以森林的形式识别嵌套实体，得到多棵嵌套实体树；将嵌套实体树的实体表示输入Transformer Decoder模块，通过多头注意力机制，得到嵌套实体树中包含实体树间交互信息、实体和输入文本之间的交互信息的隐层向量；将隐层向量和实体表示输入由多棵嵌套实体树构成的分层三元组森林，获得语料的实体语义关系三元组。
Need to check novelty before this filing date? Find Prior Art