计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (5): 330-336.doi: 10.11896/jsjkx.240300162

• 信息安全 • 上一篇    下一篇

基于多视角IB的安全风险关系抽取研究

李希望1, 曹培松1, 吴俞颖1, 郭淑明2,3, 佘维1,2   

  1. 1 郑州大学网络空间安全学院 郑州 450000
    2 嵩山实验室 郑州 450000
    3 国家数字交换系统工程技术研究中心 郑州 450000
  • 收稿日期:2024-03-25 修回日期:2024-08-16 出版日期:2025-05-15 发布日期:2025-05-12
  • 通讯作者: 佘维(wshe@zzu.edu.cn)
  • 作者简介:(350071532@qq.com)
  • 基金资助:
    国家重点研发计划(31703-3);嵩山实验室预研项目(YYYY022022003)

Study on Security Risk Relation Extraction Based on Multi-view IB

LI Xiwang1, CAO Peisong1, WU Yuying1, GUO Shuming2,3, SHE Wei1,2   

  1. 1 School of Cyber Science and Engineering,Zhengzhou University,Zhengzhou 450000,China
    2 Songshan Laboratory,Zhengzhou 450000,China
    3 National Digital Switching System Engineering & Technological R&D Center,Zhengzhou 450000,China
  • Received:2024-03-25 Revised:2024-08-16 Online:2025-05-15 Published:2025-05-12
  • About author:LI Xiwang,born in 1999,postgraduate.His main research interests include relation extraction and data mining.
    SHE Wei,born in 1977,Ph.D,professor.His main research interests include complex system modeling and simulation,machine learning and intelligent systems,information security and blockchain technology.
  • Supported by:
    National Key Research and Development Program of China(31703-3) and Songshan Laboratory Pre-research Project(YYYY022022003).

PDF (PC)

43

摘要： 安全风险管理是保障安全的核心任务,传统识别安全风险的方法已经不能满足智能化发展的需求。关系抽取是安全风险识别的方法之一,研究关系抽取对安全风险管理具有重要意义。尽管现有的模型已经取得了较好的性能,但是大多数现有的关系抽取模型忽略了领域实体表征不足的问题,并且数据中存在较多不相关信息。针对该问题,提出了一个基于多视角IB(Information Bottleneck)的安全风险关系抽取模型MIBRE(Multi-view Information Bottleneck for Relation Extraction),它通过融合多视角语义信息来达到增强领域实体语义的目的。这两个视角分别是文本视角和图像视角。为了最大化获取两个视角之间的相关信息,基于信息瓶颈方法构造了一个目标函数,在压缩两个视角信息的同时最大化地保留了相关信息。在两个真实的铁路领域数据集上的实验表明,MIBRE识别的F1值分别达到了64.28%和74.34%,相较于基于异构图的LGGCN模型F1值分别提升了4.41%和2.98%,相较于基于注意力机制的TDGAT模型F1值分别提升了1.89%和1.53%。实验结果验证了所提模型在安全风险识别上的有效性。

关键词: 关系抽取, 信息瓶颈, 多视角, 安全风险, 特征融合

Abstract: Safety risk management is the core assignment to ensure safety,and the traditional methods of identifying safety risks can no longer meet the needs of intelligent development.Research on relation extraction is of significant importance for security risk management,as it serves as one of the methods for identifying security risks.However,most existing relation extraction models ignore the problem of insufficient representation of domain entity and contain more noise in the data.To address the above problems,a multi-view IB-based safety risk relation extraction model(MIBRE) is proposed.Specifically,it achieves enhanced domain entity semantics by fusing semantic information from multi-view.In order to obtain the maximum relevant information between the two views,an objective function is constructed using the information bottleneck approach.The relevant information is maximally preserved and restored while compressing the information between the two views.Experiments on two real domain datasets show that the F1 value recognized by MIBRE reaches 64.28% and 74.34% respectively,which is 4.41% and 2.98% higher than that of LGGCN based on heterogeneous graph model.Compared with TDGAT based on attention mechanism model,F1 value increased by 1.89% and 1.53% respectively.The effectiveness of the proposed model in security risk identification is verified by experiments.

Key words: Relation extraction, Information bottleneck, Multi-view, Safety risk, Feature fusion

中图分类号: 

  • TP391
[1]SHANG B,ZHAO Y,LIU J.Learnable convolutional attention network for knowledge graph completion[J].Knowledge-Based Systems,2024,285:111360.
[2]SOUSA R T,SILVA S,PESQUITA C.Explaining protein-protein interactions with knowledge graph-based semantic similarity[J].Computers in Biology and Medicine,2024,170:108076.
[3]CHEN J,HU J,LI T,et al.An effective relation-first detection model for relational triple extraction[J].Expert Systems with Applications,2024,238:122007.
[4]LAI T,JI H,ZHAI C,et al.Tran,Joint biomedical entity and relation extraction with knowledge-enhanced collective inference[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics.2021:6248-6260.
[5]CHEN J,HU B,PENG W,et al.Biomedical relation extraction via knowledge-enhanced reading comprehension[J].BMC Bioinformtics,2022,23:20.
[6]PÉREZ-PÉREZ M,FERREIRA T,IGREJAS G,et al.A deep learning relation extraction approach to support a biomedical semi-automatic curation task:The case of the gluten bibliome[J].Expert Systems with Applications,2022,195:116616.
[7]ROY S,PACHECO M,GOLDWASSER D.Identifying morality frames in political tweets using relational learning[C]//Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing.2021:9939-9958.
[8]ZHAO D,WANG J,LIN H,et al.Biomedical cross-sentence relation extraction via multihead attention and graph convolutional networks[J].Applied Soft Computing,2021,104:107230.
[9]HILLEBRAND L,DEUßER T,KHAMENEH T,et al.KPI-BERT:A joint named entity recognition and relation extraction model for financial reports [C]//Proceedings of the 26th International Conference on Pattern Recognition(ICPR).2022:606-612.
[10]ZENG D,LIU K,LAI S,et al.Relation classification via convolutional deep neural network[C]//Proceedings of the Confe-rence 25th International Conference on Computational Linguistics.Dublin,Ireland,2014:2335-2344.
[11]YANG C,XIAO D,LUO Y,et al.A hybrid method based on semi-supervised learning for relation extraction in Chinese emrs[J].BMC Medical Informatics Decis Mak,2022,22:169.
[12]WU R,YAO Y,HAN X,et al.Open relation extraction:Relational knowledge transfer from supervised data to unsupervised data[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing.2019:219-228.
[13]MINTZ M,BILLS S,SNOW R,et al.Distant supervision for relation extraction without labeled data[C]//Proceedings of the 47th Annual Meeting of the Association for Computational Linguistics.Singapore,2009:1003-1011.
[14]SOCHER R,HUVAL B,MANNING C,et al.,Semantic compositionality through recursive matrix-vector spaces[C]//Procee-dings of the Joint Conference on Empirical Methods in Natural Language Processing & Computational Natural Language Learning.2012:1201-1211.
[15]ZENG D,LIU K,LAI S,et al.Relation classification via convolutional deep neural network[C]//Proceedings of International Conference on Computational Linguistics.2014:2335-2344.
[16]SUN H B,LI S X,TONG W Y,et al.Construction of Knowledge Graph of Power Communication Planning based on Deep Learning[C]//Proceedings of the 6th Information Technology and Mechatronics Engineering Conference.2022:843-851.
[17]XU Y,MOU L,GE L,et al.Classifying relations via long short term memory networks along shortest dependency paths[C]//Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing.2015:1785-1794.
[18]WU T,YOU X,XIAN X,et al.Towards deep understanding of graph convolutional networks for relation extraction [J].Data &Knowledge Engineering,2024,149:102265.
[19]ZHUANG L,FEI H,HU P.Knowledge-enhanced event relation extraction via event ontology prompt[J].Information Fusion,2023,100:101919.
[20]TISHBY N,PEREIRA F,BIALEK W.The information bottleneck method[C]//Proceedings of the 37th Annual Allerton Conference on Communnication Control Computing.1999:368-377.
[21]CUI S,CAO J,CONG X,et al.Enhancing multimodal entity and relation extraction with variational information bottleneck[J].IEEE/ACM Transactions on Audio,Speech,and Language Processing,2024,32,1274-1285.
[22]LI X,EISNER J.Specializing word embeddings(for parsing) by information bottleneck[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing.2019:2744-2754.
[23]AMJAD R,GEIGER B.Learning representations for neural net-work-based classification using the information bottleneck principle[J].IEEE Trans.Pattern Anal.Mach.Intell.,2020,42:2225-2239.
[24]HUANG W,MAO Y,YANG L,et al.Local-to-global GCN withknowledge-aware representation for distantly supervised relation extraction[J].Knowledge-Based Systems,2021,234:107565.
[25]SUN Q,ZHANG K,HUANG K,et al.Document-level relation extraction with two-stage dynamic graph attention networks[J].Knowledge-Based Systems,2023,267:110428.
[26]XU S,SUN S,ZHANG Z,et al.BERT gatedmulti-window attention network for relation extraction[J].Neurocomputing,2022,492:516-529.
[27]LIN Y,SHEN S,LIU Z,et al.Neural relation extraction with selective attention over instances[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics(Volume 1:Long Papers).2016:2124-2133.
[28]ZHAO Q,GAO T,GUO N.A novel chinese relation extraction method using polysemy rethinking mechanism[J].Applied Intelligence,2023,53(7):7665-7676.
[29]BUSST M M A,ANBANANTHEN K S M,KANNAN S,et al.Ensemble BiLSTM:A Novel Approach for Aspect Extraction From Online Text[J].IEEE Access,2024,12:3528-3539.
[30]YU M,CHEN Y,ZHAO M,et al.Semantic piecewise convolutional neural network with adaptive negative training for distantly supervisedrelation extraction[J].Neurocomputing,2023,537:12-21.
[31]WU M,ZHANG Q,WU C,et al.End-to-end multi-granulation causality extraction model[J].Digital Communications and Networks,2023,10(6):1864-1873.
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