·526· 智能系统学报 第16卷 4结束语 ic representations from tree-structured long short-Term memory networks[C]//Proceedings of the 53rd Annual 本文提出了一种新型的注意力图长短时记忆 Meeting of the Association for Computational Linguistics 神经网络模型(AGLSTM)。该模型将注意力机制 and the 7th International Joint Conference on Natural Lan 与句子结构树结合，实现了模型自主学习句子结 guage Processing.Beijing,China:Association for Compu 构信息的功能。所提模型不仅拥有很好的捕捉复 tational Linguistics,2015:1556-1566 杂语义关系和依赖结构的能力，并且弥补了图卷 [8]ZHANG Yuhao,QI Peng,MANNING C D.Graph convo- lution over pruned dependency trees improves relation ex- 积网络对时序信息捕捉能力差的不足。与10种 traction[C]//Proceedings of the 2018 Conference on Empir- 关系提取模型或方法进行对比，实验结果表明， ical Methods in Natural Language Processing.Brussels, 所提模型在关系抽取上具有较佳的性能，其准确 Belgium:Association for Computational Linguistics,2018: 率要远高于其他对比模型。在未来工作中，将深 2205-2215. 入研究句内结构以及句间结构的信息提取，将模 [9]甘丽新，万常选，刘德喜，等.基于句法语义特征的中文 型进行完善并应用到句间关系抽取任务。 实体关系抽取[J].计算机研究与发展，2016,53(2)： 284302. 参考文献： GAN Lixin,WANG Changxuan,LIU Dexi,et al.Chinese named entity relation extraction based on syntactic and se- [1]杨志豪洪莉，林鸿飞，等.基于支持向量机的生物医学文 mantic features[J].Journal of computer research and devel- 献蛋白质关系抽取).智能系统学报，2008(4)：361-369. 0 oment,.2016,53(2284-302. Yang Zhihao,Hong Li,Lin Hongfei,et al.Extraction of in- [10]GUO Zhijiang,ZHANG Yan,LU Wei.Attention guided formation on protein-protein interaction from biomedical graph convolutional networks for relation extraction literatures using an SVM[J].CAAI transactions on intelli- [Cl//Proceedings of the 57th Annual Meeting of the Asso- gent systems,2008(4):361-369 ciation for Computational Linguistics.Florence,Italy: [2]李智超.图文知识图谱中的关系抽取算法研究D],北 ACL,241-251. 京：北京邮电大学，2018 [11]FU T J,LI P H,MA Weiyun.GraphRel:modeling text as LI Zhichao.A relation extraction algorithm in multi-modal relational graphs for joint entity and relation extraction knowledge graph[D].Beijing:Beijing University of Posts [Cl//Proceedings of the 57th Annual Meeting of the Asso- and Telecommunications,2018. ciation for Computational Linguistics.Florence,Italy:As- [3]张涛贾真，李天瑞，等.基于知识库的开放领域问答系统 sociation for Computational Linguistics,2019: [J.智能系统学报，2018,13(4)：557-563 1409-1418. ZHANG Tao,JIA Zhen,LI Tianrui,et al.Open-domain [12]PENG Nanyun,POON H.QUIRK C,et al.Cross-sen- question-answering system based on large-scale know- tence N-ary relation extraction with graph LSTMs[J]. ledge base[J].CAAI transactions on intelligent systems, Transactions of the association for computational linguist- 2018.13(4):557-563. ics.2017,5:101-115. [4]ZHOU Peng,SHI Wei,TIAN Jun,et al.Attention-based [13]SONG Linfeng,ZHANG Yue,WANG Zhiguo,et al.N- bidirectional long short-term memory networks for rela- ary relation extraction using graph state LSTM[C]//Pro- tion classification[C]//Proceedings of the 54th Annual ceedings of the 2018 Conference on Empirical Methods in Meeting of the Association for Computational Linguistics. Natural Language Processing.Brussels,Belgium:Associ- Berlin,Germany:Association for Computational Linguist- ation for Computational Linguistics,2018:2226-2235. ics.2016:207-212 [14]ZHOU Peng,XU Jiaming,QI Zhenyu,et al.Distant su- [5]ZHANG Lei,XIANG Fusheng.Relation classification via pervision for relation extraction with hierarchical select- BiLSTM-CNN[C]//Proceedings of the 3rd International ive attention[J].Neural networks.2018.108:240-247. Conference on Data Mining and Big Data.Shanghai, [15]JI Guoliang,LIU Kang,HE Shizhu,et al.Distant supervi- China:Springer,2018:373-382. sion for relation extraction with sentence-level attention [6]XU Yan,MOU Lili,LI Ge,et al.Classifying relations via and entity descriptions[C]//Proceedings of the 31st AAAl long short term memory networks along shortest depend- Conference on Artificial Intelligence.San Francisco. ency paths[C]//Proceedings of the 2015 Conference on USA:AAAI Press,2017. Empirical Methods in Natural Language Processing.Lis- [16]ZHANG Shu,ZHENG Dequan,HU Xinchen,et al.Bid- bon,Portugal:Association for Computational Linguistics, irectional long short-term memory networks for relation 2015:1785-1794. classification[C]//Proceedings of the 29th Pacific Asia [7]TAI K S.SOCHER R,MANNING C D.Improved semant- Conference on Language,Information and Computation.4 结束语 本文提出了一种新型的注意力图长短时记忆 神经网络模型 (AGLSTM)。该模型将注意力机制 与句子结构树结合，实现了模型自主学习句子结 构信息的功能。所提模型不仅拥有很好的捕捉复 杂语义关系和依赖结构的能力，并且弥补了图卷 积网络对时序信息捕捉能力差的不足。与 10 种 关系提取模型或方法进行对比，实验结果表明， 所提模型在关系抽取上具有较佳的性能，其准确 率要远高于其他对比模型。在未来工作中，将深 入研究句内结构以及句间结构的信息提取，将模 型进行完善并应用到句间关系抽取任务。 参考文献： 杨志豪,洪莉,林鸿飞,等. 基于支持向量机的生物医学文 献蛋白质关系抽取 [J]. 智能系统学报, 2008(4): 361–369. Yang Zhihao, Hong Li, Lin Hongfei, et al. Extraction of in￾formation on protein-protein interaction from biomedical literatures using an SVM[J]. CAAI transactions on intelli￾gent systems, 2008(4): 361–369. [1] 李智超. 图文知识图谱中的关系抽取算法研究 [D]. 北 京: 北京邮电大学, 2018. LI Zhichao. A relation extraction algorithm in multi-modal knowledge graph[D]. Beijing: Beijing University of Posts and Telecommunications, 2018. [2] 张涛,贾真,李天瑞,等. 基于知识库的开放领域问答系统 [J]. 智能系统学报, 2018, 13(4): 557–563. ZHANG Tao, JIA Zhen, LI Tianrui, et al. Open-domain question-answering system based on large-scale know￾ledge base[J]. CAAI transactions on intelligent systems, 2018, 13(4): 557–563. [3] ZHOU Peng, SHI Wei, TIAN Jun, et al. Attention-based bidirectional long short-term memory networks for rela￾tion classification[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. Berlin, Germany: Association for Computational Linguist￾ics, 2016: 207−212. [4] ZHANG Lei, XIANG Fusheng. Relation classification via BiLSTM-CNN[C]//Proceedings of the 3rd International Conference on Data Mining and Big Data. Shanghai, China: Springer, 2018: 373−382. [5] XU Yan, MOU Lili, LI Ge, et al. Classifying relations via long short term memory networks along shortest depend￾ency paths[C]//Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Lis￾bon, Portugal: Association for Computational Linguistics, 2015: 1785−1794. [6] [7] TAI K S, SOCHER R, MANNING C D. Improved semant￾ic representations from tree-structured long short-Term memory networks[C]//Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Lan￾guage Processing. Beijing, China: Association for Compu￾tational Linguistics, 2015: 1556−1566. ZHANG Yuhao, QI Peng, MANNING C D. Graph convo￾lution over pruned dependency trees improves relation ex￾traction[C]//Proceedings of the 2018 Conference on Empir￾ical Methods in Natural Language Processing. Brussels, Belgium: Association for Computational Linguistics, 2018: 2205−2215. [8] 甘丽新, 万常选, 刘德喜, 等. 基于句法语义特征的中文 实体关系抽取 [J]. 计算机研究与发展, 2016, 53(2): 284–302. GAN Lixin, WANG Changxuan, LIU Dexi, et al. Chinese named entity relation extraction based on syntactic and se￾mantic features[J]. Journal of computer research and devel￾opment, 2016, 53(2): 284–302. [9] GUO Zhijiang, ZHANG Yan, LU Wei. Attention guided graph convolutional networks for relation extraction [C]//Proceedings of the 57th Annual Meeting of the Asso￾ciation for Computational Linguistics. Florence, Italy: ACL, 241−251. [10] FU T J, LI P H, MA Weiyun. GraphRel: modeling text as relational graphs for joint entity and relation extraction [C]//Proceedings of the 57th Annual Meeting of the Asso￾ciation for Computational Linguistics. Florence, Italy: As￾sociation for Computational Linguistics, 2019: 1409−1418. [11] PENG Nanyun, POON H, QUIRK C, et al. Cross-sen￾tence N-ary relation extraction with graph LSTMs[J]. Transactions of the association for computational linguist￾ics, 2017, 5: 101–115. [12] SONG Linfeng, ZHANG Yue, WANG Zhiguo, et al. N￾ary relation extraction using graph state LSTM[C]//Pro￾ceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Brussels, Belgium: Associ￾ation for Computational Linguistics, 2018: 2226−2235. [13] ZHOU Peng, XU Jiaming, QI Zhenyu, et al. Distant su￾pervision for relation extraction with hierarchical select￾ive attention[J]. Neural networks, 2018, 108: 240–247. [14] JI Guoliang, LIU Kang, HE Shizhu, et al. Distant supervi￾sion for relation extraction with sentence-level attention and entity descriptions[C]//Proceedings of the 31st AAAI Conference on Artificial Intelligence. San Francisco, USA: AAAI Press, 2017. [15] ZHANG Shu, ZHENG Dequan, HU Xinchen, et al. Bid￾irectional long short-term memory networks for relation classification[C]//Proceedings of the 29th Pacific Asia Conference on Language, Information and Computation. [16] ·526· 智 能 系 统 学 报 第 16 卷