ResNet结合BiGRU的关系抽取混合模型

摘要
参考文献
相关文章 (15)

全文: PDF (2671 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要关系抽取主要目的是将非结构化或半结构化描述的自然语言文本转化成结构化数据,其主要负责从文本中识别出实体,抽取实体间的语义关系。就关系抽取任务而言,当前流行的网络结构是仅使用CNN作为编码器,经过多层卷积操作后,对池化的结果进行softmax分类。还有部分工作则使用RNN并结合Attention机制对最后的结果做分类。这些网络结构在远程监督带噪声的关系抽取任务中表现并不理想。该文主要根据ResNet残差块的特性,提出了一种混合模型,它有效融合,ResNet和BiGRU,将带有残差特性的CNN和双向RNN结合起来,最后融入注意力机制来完成基于远程监督的关系抽取任务。实验验证了该混合模型在远程监督的噪声过滤方面的有效性。在NYT-Freebase 数据集上,P@N值相比使用单一ResNet提高了2.9%。另外,该文所建混合模型可以很轻易地移植应用到其他NLP任务中。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	唐朝
	诺明花
	胡岩

关键词 ：关系抽取, 卷积神经网络, 递归神经网络, 注意力机制

Abstract：The main purpose of relational extraction is to transform unstructured or semi-structured text into structured data, focusing on identifying entities from text and especially extracting semantic relationships between entities. This paper explores a hybrid model of ResNet and BiGRU. Based on the characteristics of the ResNet, we combine residual learning CNN with RNN on the extraction of entity relation tasks. The residual block, RNN and attention mechanism are simultaneously used for the weakly-supervised relation extraction. Experimental results indicate that, on NYT-Freebase dataset, the P@N results are improved by 2.9% compared with the single ResNet.

Key words： relation extraction CNN RNN attention mechanism

收稿日期: 2019-07-12

中图分类号:

TP183

基金资助:国家自然科学基金(61966025);内蒙古自然科学基金(2019MS06010);内蒙古自治区高等学校科学研究项目(NJZY19011)

作者简介: 唐朝(1994—),硕士研究生,主要研究领域为信息抽取、关系抽取、机器学习。E-mail:282056240@qq.com
诺明花(1981—),通信作者,博士,副教授,硕士生导师,主要研究领域为知识图谱、机器学习、中文(包括少数民族)信息处理。E-mail:nuominghua@163.com
胡岩(1994—),硕士研究生,主要研究领域为信息抽取、关系抽取、机器学习。E-mail:yanhuhy@163.com

链接本文:

http://jcip.cipsc.org.cn/CN/Y2020/V34/I2/38

[1] Banko M,Etzioni O. The tradeoffs between open and traditional relation extraction[C]//Proceedings of the ACL08,2008:28-36.
[2] Pascanu R,Mikolov T,Bengio Y. On the difficulty of training recurrent neural networks[C]//Proceedings of the International Conference on Machine Learning. 2013:1310-1318.
[3] He K,Zhang X,Ren S,et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. 2016:770-778.
[4] Huang Y Y,Wang W Y. Deep residual learning for Weakly-supervised relation extraction[C]//Proceed-ings of the 2017 Conference on Empirical Methods in Natural Language Processing,Copenhagen,Denmark,September 7-11,2017:1803-1807.
[5] Zeng D J,Liu K,Chen Y B,et al. Distant supervision for relation extraction via piecewise convolutional neural networks[C]//Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing,2015 EMNLP. Lisbon,Portugal,17-21 September. 2015:1753-1762.
[6] 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.
[7] 姚春华,刘潇,高弘毅,等. 基于句法语义特征的实体关系抽取技术[J]. 通信技术,2018,51(8):1828-1835.
[8] 欧阳丹彤,肖君,叶育鑫.基于实体对弱约束的远监督关系抽取[J].吉林大学学报(工学版),2019,49(3):912-919.
[9] Mintz M,Bills S,Snow R,et al. Distant supervision for relation extraction without labeled data[C]//Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP:Volume 2,ACL09,Association for Computational Linguistics,Stroudsburg,2009:1003-1011.
[10] 鄂海红,张文静,肖思琪,等. 深度学习实体关系抽取研究综述[J]. 软件学报,2019,(6):1973-1818.
[11] Zheng S,Wang F,Bao H,et al. Joint extraction of entities and relations based on a novel tagging scheme[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics,ACL 2017,2017:1227-1236.
[12] Bekoulis G,Deleu J,Demeester T,et al. Joint entity recognition and relation extraction as a multi-head selection problem[J]. Expert Systems with Applica-tions,2018,114:34-45.
[13] Peng N,Poon H,Quirk C,et al. Cross-sentence n-ary relation extraction with graph lstms[J]. Transactions of the Association for Computational Linguistics,2017,5(1):101-115.
[14] Jiang X T,Wang Q,Li P,et al. Relation extraction with multi-instance multi-label convolutional neural net-works[C]//Proceedings of the 26th International Conference on Computational Linguistics(COLING 2016),2016:1471-1480.
[15] Miwa M,Bansal M. End-to-end relation extraction using LSTMs on sequences and tree structures[C]//Proceedings of the ACL 2016,2016:1105-1116.
[16] 张兰霞,胡文心. 基于双向神经网络和双层注意力机制的中文文本中人物关系抽取研究[J]. 计算机应用与软件,2018,35(11):136-141.
[17] Pascanu R,Mikolov T,Bengio Y. On the difficulty of training recurrent neural networks[C]//Proceedings of the International Conference on Machine Learning. 2013:1310-1318.
[18] 王红,史金钏,张志伟. 基于注意力机制的LSTM的语义关系抽取[J]. 计算机应用研究,2018,5(5):1417-1420.
[19] 张晓斌,陈福才,黄瑞阳. 基于双向融合的实体关系抽取[J]. 网络与信息安全学报,2018,4(09):48-55.
[20] Hihi S E,Bengio Y. Hierarchical recurrent neural networks for long-term dependencies[C]//Proceed-ings of Advances in Neural Information Processing Systems. 1996:493-499.
[21] Hoffmann R,Zhang C,Ling X,et al. Knowledge-based weak supervision for information extraction of overlapping relations[C]//Proceedings of Meeting of the Association for Computational Linguistics:Human Language Technologies. Association for Computational Linguistics,2011:541-550.
[22] Scott M,Fox H,Ramshaw L,et al. A novel use of statistical parsing to extract information from text[C]//Proceedings of the NAACL,2000:226-233.
[23] Zeng D,Liu K,Lai S,et al. Relation classification via convolutional deep neural network[C]//Proceedings of 25th International Conference on Computational Linguistics,COLING 2014,Technical Papers,2014,2335-2344.