首页> 外国专利> DEPENDENCY PARSING METHOD BASED ON NEURAL NETWORK AND DEPENDENCY PARSING APPARATUS USING THEREOF

DEPENDENCY PARSING METHOD BASED ON NEURAL NETWORK AND DEPENDENCY PARSING APPARATUS USING THEREOF

机译:基于神经网络的依存度解析方法及其使用的依存度解析装置

摘要

The present invention relates to a method of using a bi-directional neural network to perform dependent parsing. More specifically, the present invention, in a control method for dependency parsing (dependency parsing), calculating a first vector quantifying the dictionary meaning of a predetermined word, at least one component included in the predetermined word Control comprising the step of calculating a second vector quantifying subword information, and designating a dependency word having a dependency relationship with the predetermined word based on the first and second vectors. It's about how.
机译:本发明涉及一种使用双向神经网络来执行相关解析的方法。更具体地,本发明在用于依赖性解析(dependency parsing)的控制方法中,计算量化预定单词的词典含义的第一矢量,包括在预定单词Control中的至少一个分量包括计算第二矢量的步骤。量化子词信息,并基于第一向量和第二向量来指定与预定词具有依赖关系的依赖词。关于如何。

著录项

  • 公开/公告号KR1020200072593A

    专利类型

  • 公开/公告日2020-06-23

    原文格式PDF

  • 申请/专利权人 동국대학교 산학협력단;

    申请/专利号KR1020180154384

  • 发明设计人 양기주;

    申请日2018-12-04

  • 分类号

  • 国家 KR

  • 入库时间 2022-08-21 10:57:05

相似文献

  • 专利
  • 外文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号