《计算机应用研究》|Application Research of Computers

基于对抗学习和全局知识信息的关系检测技术研究

Research on relation detection based on adversarial learning and global knowledge information

免费全文下载 (已被下载 次)  
获取PDF全文
作者 王勇超,杨英宝,曹钰,邢卫
机构 浙江大学 a.信息技术中心;b.计算机学院,杭州 310027
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2021)05-008-1327-04
DOI 10.19734/j.issn.1001-3695.2020.07.0177
摘要 针对现有的知识库关系检测任务对于一些不可见关系无法做到准确的向量表示而出现词汇溢出的问题,提出了基于对抗学习和全局知识信息的关系检测模型。该模型使用对抗学习对知识库关系表示模型进行特征强化,使用TransH(translating on hyperplanes)模型提取全局知识信息,同时通过联合训练,将全局知识信息融合进关系表示模型中,进一步提升关系模型的表示能力。实验结果表明,提出的融合模型对于关系检测效果有一定的提升,并且缓解了词汇溢出的问题。
关键词 关系检测; 知识库; 联合训练; 全局知识信息; 对抗学习
基金项目 国家重点研发计划资助项目(2019YFC1521304,2020YFC1523101)
浙江省科技计划项目(2019C03137)
浙江省重点研发计划资助项目(2018C03051,2021C03140)
石窟寺文物数字化保护国家文物局重点科研基地资助项目
本文URL http://www.netgaindomains.com/article/01-2021-05-008.html
英文标题 Research on relation detection based on adversarial learning and global knowledge information
作者英文名 Wang Yongchao, Yang Yingbao, Cao Yu, Xing Wei
机构英文名 a.Center of Information & Technology,b.Institute of Computer Science,Zhejiang University,Hangzhou 310027,China
英文摘要 Aiming at the problem of out-of-vocabulary due to the fact that the existing knowledge base relationship detection task are unable to achieve accurate vector representation for some invisible relationships, this paper proposed a relationship detection model based on adversarial learning and global knowledge information. The model used adversarial learning to enhance the characteristics of the relational representation model of the knowledge base, and used the TransH model to extract global knowledge information. At the same time, through joint training, it integrated the global knowledge information into the relational representation model to further improve the representation ability of the relational model. The experimental results show that the proposed fusion model improves the effect of relationship detection and alleviates the problem of out-of-vocabulary.
英文关键词 relation detection; knowledge base; joint training; global knowledge information; adversarial learning
参考文献 查看稿件参考文献
 
收稿日期 2020/7/3
修回日期 2020/8/24
页码 1327-1330,1343
中图分类号 TP391.4
文献标志码 A
天下色情网