一种基于核心论元的篇章级事件抽取方法
A document-level event extraction method based on core arguments
孙承杰 1李宗蔚 1单丽莉 1林磊1
作者信息
- 1. 哈尔滨工业大学计算学部,黑龙江 哈尔滨 150001
- 折叠
摘要
提出一种基于核心论元的篇章级事件抽取选取方法(core arguments-based document level event extraction,CA-DocEE),该方法根据论元在篇章级事件中的分布特点定义核心论元的选取标准,采用异质图卷积神经网络将篇章上下文信息用于增强论元实体编码,基于机器阅读理解方法捕捉句子中的深层次语义信息来进行论元角色分类.在篇章级事件抽取公开数据集上,本文提出的方法的微平均F1 值达到了 80.1%,取得了与目前已知最好方法相当的效果.
Abstract
A document-level event extraction method based on core arguments(CA-DocEE)is proposed,which defines criteria for selecting core arguments based on their distributions in document-level events,uses heterogeneous graph convolutional neural net-works to augment document contextual information for encoding argument entities,and captures deep semantic information in sen-tences based on machine reading comprehension methods for classifying the role of arguments.On the document-level event extrac-tion dataset,the method proposed in this paper achieves a micro-average F1 value of 80.1%,which is comparable with the state-of-the-art methods.
关键词
事件抽取/篇章级事件抽取/机器阅读理解/图卷积神经网络Key words
event extraction/document-level event extraction/machine reading comprehension/graph convolutional neural network引用本文复制引用
基金项目
国家重点研发计划资助项目(2021YFF0901600)
国家自然科学基金资助项目(62176074)
哈尔滨工业大学新兴交叉融拓计划(SYL-JC-202203)
出版年
2024
NETL
NSTL
万方数据