计算机科学

计算机科学 ›› 2024, Vol. 51 ›› Issue (6A): 230700071-5.doi: 10.11896/jsjkx.230700071

• 人工智能 • 上一篇    下一篇

一种基于异构图神经网络和文本语义增强的实体关系抽取方法

彭勃1, 李耀东1, 龚贤夫1, 李浩2   

  1. 1 广东电网有限责任公司电网规划研究中心 广州 510080
    2 四川大学计算机学院 成都 610065
  • 发布日期:2024-06-06
  • 通讯作者: 彭勃(1339512151@qq.com)
  • 基金资助:
    中国南方电网有限责任公司科技项037700KK52220042(GDKJXM20220906)

Method for Entity Relation Extraction Based on Heterogeneous Graph Neural Networks and TextSemantic Enhancement

PENG Bo1, LI Yaodong1, GONG Xianfu1, LI Hao2   

  1. 1 Grid Planning and Research Center of Guangdong Power Grid Co,Guangzhou 510080,China
    2 College of Computer Science,Sichuan University,Chengdu 610065,China
  • Published:2024-06-06
  • About author:PENG Bo,born in 1991,postgraduate,engineer.His main research interest is power grid planning.
  • Supported by:
    science and Technology Proiect of China Southern PowerGrid Co. Ltd 037700KK5220042(GDKIXM20220906).

PDF (PC)

315

摘要： 信息化时代,如何从海量自然语言文本中提取结构化信息已经成为研究热点。电力系统中繁杂的知识信息需要通过构建知识图谱来解决,而实体关系抽取是其上游的信息抽取任务,其完成度直接关系到知识图谱的有效性。而随着深度学习的不断发展,利用深度学习技术来完成实体关系抽取任务的研究逐渐展开并取得了良好的效果。然而目前依然存在文本语义应用不完全等问题。针对这些问题本文尝试提出了一种基于异构图神经网络和文本语义增强的实体关系抽取方法,该方法使用词节点与关系节点学习语义特征,并通过BRET与预训练任务分别获得两种节点的初始特征,使用多层图网络结构迭代更新,并在每一层中使用基于多头注意力机制的信息传递实现两种节点的交互。通过该模型与其他实体关系抽取在两个公开数据集上实验对比,所提模型取得了预期效果,在多种情境下普遍优于对比模型。

关键词: 深度学习, 自然语言处理, 知识图谱, 实体关系抽取, 异构图神经网络, 文本语义增强

Abstract: In the era of information technology,extracting structured information from massive natural language texts has become a research hotspot.The complex knowledge information in the power system needs to be solved by constructing a knowledge graph,and entity relation extraction is the upstream information extraction task,whose completeness directly affects the effectiveness of the knowledge graph.With the continuous development of deep learning,research on using deep learning techniques to solve entity relation extraction tasks has gradually been carried out and achieved good results.However,there are still problems such as incomplete application of text semantics.This paper attempts to propose an entity relation extraction method based on heterogeneous graph neural network and text semantic enhancement to address these issues.This method uses word nodes and relationship nodes to learn semantic features and obtains initial features of the two types of nodes through BRET and pre-training tasks respectively.It uses a multi-layer graph network structure for iteration and implements the interaction between the two types of nodes by using multi-head attention mechanism for information transmission in each layer.Through experimental comparison with other models on two public datasets,this model achieves the expected effect and generally outperforms other entity relationship extraction models in various scenarios.

Key words: Deep learning, Natural language processing, Knowledge graph, Entity relation extraction, Heterogeneous graph neural networks, Text semantic enhancement

中图分类号: 

  • TP391
[1]LIU K.A survey on neural relation extraction[J].Science China Technological Sciences,2020,63(10):1971-1989.
[2]KAMBHATLA N.Combining lexical,syntactic,and semanticfeatures with maximum entropy models or information extraction[C]//Proceedings of the ACL Interactive Poster and De-monstration Sessions.2004:178-181.
[3]CULOTTA A,MCCALLUM A,BETZ J.Integrating probabilistic extraction models and data mining to iscover relations and patterns in text[C]//Proceedings of the Human Language Technology Conference of the NAACL.2006:296-303.
[4]ZHANG M,ZHANG J,SU J.Exploring syntactic features for relation extraction using a convolution tree kernel[C]//Proceedings of the Human Language Technology Conference of the NAACL.2006:288-295.
[5]CHAN Y S,ROTH D.Exploiting syntactico-semantic structures for relation extraction[C]//Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics:Human Language Technologies.2011:551-560.
[6]RINK B,HARABAGIU S.Utd:Classifying semantic relationsby combining lexical and semantic resources[C]//Proceedings of the 5th International Workshop on Semantic Evaluation.2010:256-259.
[7]ZENG D,LIU K,LAI S,et al.Relation classification via convolutional deep neural network[C]//25th International Confe-rence on Computational Linguistics:Technicalpapers(COLING 2014).2014:2335-2344.
[8]SHEN Y L,ZHANG P,LI X Y,et al.Grid robuststate estimation based on multi-source data fusion[J].Guangdong Electric Power,2019,32(9):146-153.
[9]YANG S L,FU W Z,DONG J.Research on the incrementalknowledge base system in intelligent decision support system[J].Journal of Hefei University of technology,2004,27(4):339-343.
[10]XU B,ZHANG Y M.System condition-based maintenance decision-making model considering severe weather conditions[J].Guangdong Electric Power,2019,32(6):61-69.
[11]HUANG C L,WANG F,WANG J J.Intelligent online hydropower dispatching based on knowledge reasoning[J].Automation of Power Systems,2010,34(21):50-54.
[12]DEVLIN J,CHANG M W,LEE K,et al.Bert:Pre-training of deep bidirectional transformers for language understanding[C]//Proceedings of NAACL-HLT.2019.
[13]YU B,ZHANG Z,SHU X,et al.Joint extraction of entities and relations based on a novel decomposition strategy[C]//Euro-pean Conference on Artificial Intelligence.2020:2282-2289
[14]ZENG X,HE S,ZENG D,et al.Learning the extraction order of multiple relational facts in a sentence with reinforcement lear-ning[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing(EMNLPIJCNLP).2019:367-377.
[15]GUPTA P,SCHÜTZE H,ANDRASSY B.Table filling multi-task recurrent neural network for joint entity and relation extraction[C]//26th International Conference on Computational Linguistics:Technical Papers(COLING 2016).2016:2537-2547.
[16]ZHANG Y,QI P,MANNING C D.Graph convolution overpruned dependency trees improves relation extraction[C]//Conference on Empirical Methods in Natural Language Processing.2018.
[17]HONG Y,LIU Y,YANG S,et al.Improving graph convolu-tional networks based on relation-aware attention for end-to-end relation extraction[J].IEEE Access,2020,8:51315-51323.
[18]ZENG X,ZENG D,HE S,et al.Extracting relational facts by an end-to-end neural model with copy mechanism[C]//Proceedings of the 56th Annual Meeting of the Association for Compu-tational Linguistics(Volume 1:Long Papers).2018:506-514.
[19]FU T J,LI P H,MA W Y.GraphRel:Modeling text as rela-tional graphs for joint entity and relation extraction[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics.Florence,Italy:Association for Computational Linguistics,2019:1409-1418.
[20]WEI Z,SU J,WANG Y,et al.A novel cascade binary tagging framework for relational triple extraction[C]//Annual Meeting of the Association for Computa-tional Linguistics.2019.
[21]ZHAO K,XU H,CHENG Y,et al.Representation iterative fusion based on heterogeneous graph neural network for joint entity and relation extraction[J].Knowledge-Based Systems,2021,219:106888.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发