基于高速多核网络的远监督关系抽取方法

doi:10.13190/j.jbupt.2020-071

北京邮电大学学报 ›› 2020, Vol. 43 ›› Issue (5): 71-76.doi: 10.13190/j.jbupt.2020-071

基于高速多核网络的远监督关系抽取方法

李威¹, 陈曙东^1,2, 欧阳小叶², 杜蓉², 王荣³

1. 中国科学院大学微电子学院, 北京 100049;
2. 中国科学院微电子研究所, 北京 100029;
3. 北京跟踪与通信技术研究所空间目标测量重点实验室, 北京 100094

收稿日期:2020-06-20 发布日期:2021-03-11
通讯作者: 陈曙东(1977-),女,研究员,博士生导师,E-mail:chenshudong@ime.ac.cn. E-mail:chenshudong@ime.ac.cn
作者简介:李威(1995-),男,硕士生.
基金资助:
中国科学院战略性先导科技专项（C类）（XDC02070600）

Distant Supervision Relation Extraction Method Based on Highway Multi-Kernel Network

LI Wei¹, CHEN Shu-dong^1,2, OUYANG Xiao-ye², DU Rong², WANG Rong³

1. School of Microelectronics, University of Chinese Academy of Sciences, Beijing 100049, China;
2. Institute of Microelectronics, Chinese Academy of Sciences, Beijing 100029, China;
3. Key Laboratory of Space Object Measurement, Department, Beijing Institute of Tracking and Telecommunications Technology, Beijing 100094, China

Received:2020-06-20 Published:2021-03-11

摘要/Abstract

摘要： 远监督作为一种能够快速大量产生标注数据的技术，在关系抽取任务中的应用愈加广泛，但仍存在文本特征提取不足、包内噪声过多等问题.对此，提出了一种基于高速多核网络的远监督关系抽取方法.首先通过高速网络和多核卷积对句子特征进行深层提取；然后采用包内注意力机制提高包内正确标注的句子权重，降低包内噪声，实现包级向量化；使用包间注意力机制降低包间噪声，得到组级向量化；最后，将组作为训练样本训练分类器，实现关系抽取.实验结果表明，该方法比现有方法具有更好的关系抽取性能.

关键词: 关系抽取, 远监督, 注意力机制, 神经网络, 高速多核网络模型

Abstract: As a technology that can quickly generate large amounts of labeled data, the distant supervision is increasingly used in relation extraction. However, there are still problems such as insufficient text feature extraction and noise in the bag. A distant supervision relation extraction method based on highway multi-kernel network is proposed to solve these questions. Firstly, the feature of sentences are deeply extracted by highway network and multi-kernel convolution; and then the intra-bag attention mechanism is used to improve the sentence weight of the correct annotation in bag and reduce the intra-bag noise to obtain the bag‘s embedding. Subsequently, the inter-bag attention mechanism is used to reduce the inter-bag noise for each group of bags with the same relation to obtain the group‘s embedding. Finally, groups are used as training samples to train the classifier to achieve relation extraction. Experiment shows that this method has better relation extraction performance than existing methods.

Key words: relation extraction, distant supervision, attention mechanism, neural network, highway multi-kernel network

中图分类号:

TP391

李威, 陈曙东, 欧阳小叶, 杜蓉, 王荣. 基于高速多核网络的远监督关系抽取方法[J]. 北京邮电大学学报, 2020, 43(5): 71-76.

LI Wei, CHEN Shu-dong, OUYANG Xiao-ye, DU Rong, WANG Rong. Distant Supervision Relation Extraction Method Based on Highway Multi-Kernel Network[J]. Journal of Beijing University of Posts and Telecommunications, 2020, 43(5): 71-76.

参考文献

[1] Riedel S, Yao Limin, McCallum A. Modeling relations and their mentions without labeled text[J]. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 2010(part 1):148-163.
[2] Mintz M, Bills S, Snow R, et al. Distant supervision for relation extraction without labeled data[C]//International Joint Conference on Association for Computational Linguistics. Singapore:ACL and AFNLP, 2009:1003-1011.
[3] Ding Xiaohan, Guo Yuchen, Ding Guiguang, et al. ACNet:strengthening the kernel skeletons for powerful CNN via asymmetric convolution blocks[J]. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2019(3):1911-1920.
[4] 冀振燕, 宋晓军, 皮怀雨, 等. 基于深度学习的融合多源异构数据的推荐模型[J]. 北京邮电大学学报, 2019, 42(6):35-42. Ji Zhenyan, Song Xiaojun, Pi Huaiyu, et al. Recommended model for fusing multi-source heterogeneous data based on deep learning[J]. Journal of Beijing University of Posts and Telecommunications, 2019, 42(6):35-42.
[5] Zeng Daojian, Liu Kang, Chen Yubo, et al. Distant supervision for relation extraction via piecewise convolutional neural networks[C]//Proceedings of the 2015 Conference in Empirical Methods in Natural Language Processing. Lisbon:ACL, 2015:1753-1762.
[6] Lin Yankai, Shen Shiqi, Liu Zhiyuan, et al. Neural relation extraction with selective attention over instances[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. Berlin:ACL, 2016:2124-2133.
[7] Ye Zhixiu, Ling Zhenhua. Distant supervision relation extraction with intra-bag and inter-bag attentions[J]. Proceedings of NAACL-HLT, 2019(6):2810-2819.
[8] Liu Tianyu, Wang Kexiang, Chang Baobao, et al. A soft-label method for noise-tolerant distantly supervised relation extraction[C]//Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing. Copenhagen:ACL, 2017:1790-1795.
[9] 李浩, 刘永坚, 解庆, 等. 基于多层次注意力机制的远程监督关系抽取模型[J]. 计算机科学, 2019, 46(10):252-257. Li Hao, Liu Yongjian, Xie Qing, et al. Distant supervision relation extraction model based on multi-level attention mechanism[J]. Computer Science, 2019, 46(10):252-257.
[10] Raphael Hoffmann, Zhang Congle, Ling Xiao, et al. Knowledge-based weak supervision for information extraction of overlapping relations[C]//Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics. Portland:ACL, 2011:541-550.
[11] Surdeanu M, Tibshirani J, Nallapati R, et al. Multi-instance multi-label learning for relation extraction[C]//Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning. Jeju Island:ACL, 2012:455-465.

[1]	包乾辉文娟石淑珍董萌萍刘雪. 融合句法依存与 BERT-Att-BiLSTM 的属性分类[J]. 北京邮电大学学报, 2023, 46(4): 123-128.
[2]	杨丽花, 聂倩, 呼博, 江婷. 联合BP神经网络与基扩展模型的信道预测算法 [J]. 北京邮电大学学报, 2023, 46(3): 13-18.
[3]	吴进, 赵博, 汶恒, 王宇. 一种异构架构的图神经网络加速器优化研究一种异构架构的图神经网络加速器优化研究[J]. 北京邮电大学学报, 2023, 46(3): 19-24.
[4]	王延凯龙华邵玉斌杜庆治王瑶. 基于融合特征MGCC的语种识别方法[J]. 北京邮电大学学报, 2023, 46(2): 116-121.
[5]	胡敏熊思黄宏程张光华王春芳. 基于RAPNet的脑肿瘤MRI图像三维分割[J]. 北京邮电大学学报, 2023, 46(2): 91-97.
[6]	张昊阁邵玉斌龙华杜庆治周大春. 基于Gammatone尺度功率规整系数谱图的语种识别[J]. 北京邮电大学学报, 2023, 46(2): 122-128.
[7]	张昊阁邵玉斌龙华彭艺周大春. 基于对数Gammatone滤波器能量谱图的语种识别[J]. 北京邮电大学学报, 2023, 46(1): 38-43.
[8]	张天骐汪锐安泽亮方竹王雪怡. 基于多任务学习的MIMO-OFDM信号联合信噪比估计与调制识别[J]. 北京邮电大学学报, 2022, 45(6): 98-104.
[9]	余佳润王亚峰. 一种低信噪比下基于深度学习的DoA估计方法[J]. 北京邮电大学学报, 2022, 45(6): 119-125.
[10]	叶康张淑军郭淇李辉崔雪红. 基于CM-Transformer的连续手语识别[J]. 北京邮电大学学报, 2022, 45(5): 49-53,78.
[11]	武兴坤, 罗涛, 刘爱军, 杨明, 张文静. 面向主动脉缩窄辅助诊断的神经网络模型设计[J]. 北京邮电大学学报, 2022, 45(4): 1-6.
[12]	李星原, 汪鹏, 申牧, 李蕾, 张琳. 癫痫病相关论文多模态知识图谱的构建初探[J]. 北京邮电大学学报, 2022, 45(4): 19-24.
[13]	刘琳岚, 宋修洋, 陈宇斌. 基于网络表示学习的机会网络链路预测[J]. 北京邮电大学学报, 2022, 45(4): 64-69,103.
[14]	武兴坤罗涛刘爱军杨明张文静李剑峰李文秀. 面向主动脉缩窄辅助诊断的神经网络模型设计[J]. 北京邮电大学学报, 2022, 45(4): 1-6.
[15]	李星原汪鹏申牧李蕾张琳. 癫痫领域多模态知识图谱的构建初探[J]. 北京邮电大学学报, 2022, 45(4): 21-27.

基于高速多核网络的远监督关系抽取方法

Distant Supervision Relation Extraction Method Based on Highway Multi-Kernel Network

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价