一种用于关系抽取的双层时空图卷积神经网络
A bi-layer spatiotemporal graph convolutional neural network for relation extraction
李灵芳,陈效成,李宝山,杜永兴,杨颜博
( 内蒙古科技大学 信息工程学院,内蒙古 包头 014010)
摘 要:关系抽取是自然语言处理中一项基础的上游任务,句子的结构信息在某种意义上蕴含了实体及其关系信 息,有助于提高关系抽取的准确率,然而使用现有自然语言处理( Natural Language Processing,NLP) 语言工具进行 句法分析时会引入一定的错误传播问题,且现有的基于图结构的关系抽取模型在一定程度上忽略了句子的时序信息,通过结合双向长短时记忆网络( Bi-directional Long Short-Term Memory,Bi LSTM) 捕获句子序列的上下文关系, 同时使用传统条件随机场( Conditional Random Field,CRF) 的关系标注结果矫正 NLP 工具的错误传播问题,提出了 一种用于关系抽取的双层时空图卷积神经网络( Bilayer Spatiotemporal Graph Convolution Neural Network,Bi SpGCN) 模型 . 该模型在中文糖尿病数据集和中文人物关系数据集上的实验结果表明,相较于传统的多头注意力引导的图 卷积神经网络( Attention Guided Graph Convolutional Networks for Relation Extraction,AGGCN) 模型,BiSpGCN 模型能 够充分利用句子的有效信息,具有更好的关系抽取性能。
关键词:关系抽取; 条件随机场; 图卷积神经网络; 双向长短时记忆网络; 双层编码
DOI: 10.16559 /j.cnki.2095 - 2295.2022.03.013
基金项目:内蒙古自治区自然科学基金资助项目( 2021MS06007);内蒙古自治区科技重大专项资助项目( 2019ZD025);内蒙古科技大 学创新基金资助项目( 2019QDL-S10)
作者:李灵芳,陈效成,李宝山,杜永兴,杨颜博
参考文献:
[1] Aone C,Ramos-Santacruz M.Rees: A large-scale relation and event extraction system[C]/ / In: Proc of the 6th applied natural language processing conference. New York: ACM Press,2000: 76.
[2] Zhang T. Regularized Winnow methods[C]/ / In: Advances in Neural Inf Ormation Processing Systems ( NIPS) 13.Cambridge: MIT Press,2001: 703.
[3] Cristianini N,Shawe-Taylor J,Lodhi H. Latent semantic ker-nels[J]. Journal of Intelligent Information Systems,2002,18( 2-3) : 127.
[4] Wang H X,Qin K,Lu G M,et al. Deep neural networkbased relation extraction: an overview[J]. Neural Computing and Applications,2022,34: 1.
[5] 陈雨龙,付乾坤,张岳. 图神经网络在自然语言处理中 的应用[J]. 中文信息学报,2021,35( 03) : 1.
[6] Bugliarello E,Okazaki N. Enhancing machine translation with dependency-aware self-attention[C]/ / In proceedings of the 58th Annual meeting of the Association for Computational Linguistcs. Berlin, Germany: ACL, 2020: 1618 .
[7] Zhang Y,Gou Z,Lu W. Attention Guided Graph Convolutional Networks for Relation Extraction[J]. IEICE Transactions on Fundamentals of Electronics,Communications and Computer Sciences, Volumeabs/1906. 07510,2019.
[8] Zhou P,Tian J,Qi Z,et al. Attentionbased bidirectional long short-termmemory networks for relation classification[C]/ / In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. Berlin,Germany: ACL,2016: 207.
[9] Miwa M,Bansal M. End-to-End Relation Extraction using LSTMs on Sequences and Tree Structures[C]/ / 54th Annual Meeting of the Association-for-Computational-Linguistics. Berlin,Germany: ACL,2016: 1105.
[10] Zhang Y X,Liao X Y,ChenL,et al. Multi-BERT-wwm Model Based on Probabilistic Graph Strategy for Relation Extraction[J]. Health Information Science,2021, 13079: 95.
[11] Lai T,Cheng L,Wang D,et al. RMAN: Relational multi-head attention neural network for joint extraction of entities and relations [J]. Applied Intelligence, 2021,52( 8) : 3132.
[12] Ding K,Liu S,Zhang Y,et al.A Knowledge-enriched and span-based network for joint entity and relation extraction[J]. Computers,Materials and Continua,2021,680( 1) : 377.
[13] Zhao W,Zhao S,Chen S H,et al.Entity and relation collaborative extraction approach based multi-head attention and gated mechanism[J]. Connection Science, 2022,34( 1) : 670.
[14] Qu J,Hua W,Ou Yang D,et al.An efficient and effective approach for multi-fact extraction from text corpus [J]. World Wide Web,2021,25( 1) : 195.
[15] Wei J,Zou K. EDA: Easy data augmentation techniques for boosting performance on text classification tasks[J]. 2019,arXiv: 1901. 11196v2.
[16] 任燕春,赵瑛,王铁,等. 基于新冠肺炎知识图谱的智 能问答系统研究[J].内蒙古科技大学学报,2021,40 ( 03) : 287.
[17] Hong Y,Liu Y,Yang S,et al.Joint extraction of entities and relations using graph convolution over pruned dependency trees[J].Neurocomputing,2020,411: 302.
[18] Luo L,Yang Z,Yang P,et al. An attention-based BiLSTM-CRF approach to document-level chemical named entity recognition[J]. Bioinformatics,2018,34 ( 8 ):1381.
[19] 李灵芳,杨佳琦,李宝山,等. 基于 BERT 的中文电子 病历命名实体识别[J].内蒙古科技大学学报,2020,39( 01);71.