基于Gate机制与BiLSTMCRF的汉语语义角色标注

doi:10.3969/j.issn.10062475.2018.04.001

计算机与现代化

• 人工智能 • 下一篇

基于Gate机制与BiLSTMCRF的汉语语义角色标注

(北京交通大学计算机与信息技术学院，北京100044)

出版日期:2018-04-28 发布日期:2018-05-02
作者简介:张苗苗(1992)，女，山东济宁人，北京交通大学计算机与信息技术学院硕士研究生，研究方向：自然语言处理，语义分析; 张玉洁(1961)，女，河南安阳人，教授，博士，研究方向：自然语言处理，机器翻译; 刘明童(1993)，男，四川广元人，博士研究生，研究方向：自然语言处理，复述; 徐金安(1970)，男，副教授，博士，研究方向：自然语言处理，机器翻译; 陈钰枫(1981)，女，副教授，博士，研究方向：自然语言处理，机器翻译。
基金资助:
国家自然科学基金资助项目(61370130, 61473294); 北京市自然科学基金资助项目(4172047); 北京交通大学人才基金资助项目(KKRC11001532)

#br# Chinese Semantic Role Labeling Based on Gated Mechanism and BiLSTMCRF

(School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, China)

Online:2018-04-28 Published:2018-05-02

摘要/Abstract

摘要： 目前，语义角色标注大多基于双向长短时记忆网络（BiLSTM）。但是，由于词向量表示由上下文窗口中的词嵌入拼接得到，导致其依赖于左右词嵌入的联合作用。针对该问题，引入Gate机制对词向量表示进行调整。为了获取更深层次的语义信息，对BiLSTM的深度进行扩展。此外，引入标签转移概率矩阵进行约束，并且使用条件随机场（CRF）融合全局标签信息得出最优标注序列。实验结果表明，该方法使得汉语语义角色标注的F1值提高1.71%。

关键词: 汉语语义角色标注, Gate机制, BiLSTMCRF, 标签转移概率矩阵

Abstract: With the increasing research and application of deep learning, researchers have proposed a method by applying bidirectional long shortterm memory (BiLSTM) to semantic role labeling. Among them, many methods rely on a local window in word representation. In this case, the word representation may depend on the joint effect of some fixed word embeddings and it may preserve useless information in context. Aiming at this problem, we make improvements based on BiLSTM and introduce a Gated mechanism (GM) to filter information that will be fed to the next layer. In order to gain much more semantic information, we extend the depth of the BiLSTM. Due to the strong dependencies across output labels, we model tagging decisions using a conditional random fields (CRF) and transition matrices. Experiments on CPB benchmark dataset show that the F1 score of the Chinese semantic role labeling system finally improves 1.71%.

Key words: Chinese semantic role labeling, Gated mechanism, BiLSTMCRF, transition matrix

中图分类号:

TP391

张苗苗，张玉洁，刘明童，徐金安，陈钰枫. 基于Gate机制与BiLSTMCRF的汉语语义角色标注[J]. 计算机与现代化, doi: 10.3969/j.issn.10062475.2018.04.001.

ZHANG Miaomiao, ZHANG Yujie, LIU Mingtong, XU Jinan, CHEN Yufeng. #br# Chinese Semantic Role Labeling Based on Gated Mechanism and BiLSTMCRF[J]. Computer and Modernization, doi: 10.3969/j.issn.10062475.2018.04.001.

参考文献

［1］中国中文信息学会. 中文信息处理发展报告(2016)［R］. 北京:中国中文信息学会, 2016.
［2］ Hochreiter S, Schmidhuber J. Long shortterm memory［J］. Neural Computation, 1997,9(8):17351780.
［3］〖JP3〗Sun Honglin, Jurafsky D. Shallow semantic parsing of Chinese［C］// Proceedings of the 2004 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics. 2004:249256. 
［4］〖KG-*6〗〖JP3〗Xue Nianwen. Labeling Chinese predicates with semantic roles［J］. Computational Linguistics, 2008,34(2):225255. 
［5］ Sun Weiwei, Sui Zhifang, Wang Meng, et al. Chinese semantic role labeling with shallow parsing［C］// Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing. 2009,3:14751483.
［6］ Yang Haitong, Zong Chengqing. Multipredicate semantic role labeling［C］// Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. 2014:363373.
［7］ Collobert R, Weston J. A unified architecture for natural language processing: Deep neural networks with multitask learning［C］// Proceedings of the 25th International Conference on Machine Learning. 2008:160167.
［8］ Collobert R, Weston J, Bottou L, et al. Natural language processing (almost) from scratch［J］. The Journal of Machine Learning Research, 2011,12:24932537.
［9］ Zhou Jie, Xu Wei. Endtoend learning of semantic role labeling using recurrent neural networks［C］// Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics. 2015:11271137.
［10］Roth M, Lapata M. Neural semantic role labeling with dependency path embeddings［C］// Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. 2016,1:11921202.
［11］ 王臻,常宝宝,穗志方. 基于分层输出神经网络的汉语语义角色标注［J］. 中文信息学报, 2014,28(6):5661. 
［12］Wang Zhen, Jiang Tingsong, Chang Baobao, et al. Chinese semantic role labeling with bidirectional recurrent neural networks［C］// Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. 2015:16261631.
［13］Sha Lei, Jiang Tingsong, Li Sujian, et al. Capturing argument relationships for Chinese semantic role labeling［C］// Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing. 2016:20112016.
［14］Guo Jiang, Che Wanxiang, Wang Haifeng, et al. A unified architecture for semantic role labeling and relation classification［C］// Proceedings of the 26th International Conference on Computational Linguistics. 2016:12641274.
［15］ Tu Zhaopeng, Liu Yang, Lu Zhengdong, et al. Context gates for neural machine translation［J］. Transactions of the Association for Computational Linguistics, 2017,5:8799. 
［16］Lafferty J, McCallum A, Pereira F. Conditional random fields: Probabilistic models for segmenting and labeling sequence data［C］// Proceedings of the 18th International Conference on Machine Learning. 2001:282289.
［17］Taskar B, Chatalbashev V, Koller D, et al. Learning structured prediction models: A large margin approach［C］// Proceedings of the 22nd International Conference on Machine Learning. 2005:896903.
［18］Srivastava N, Hinton G, Krizhevsky A, et al. Dropout: A simple way to prevent neural networks from overfitting［J］. The Journal of Machine Learning Research, 2014,15:19291958.
［19］王瑞波,李济洪,李国臣,等. 基于Dropout正则化的汉语框架语义角色识别［J］. 中文信息学报, 2017,31(1):147154.
［20］Glorot X, Bengio Y. Understanding the difficulty of training deep feedforward neural networks［C］// Proceedings of the 13th International Conference on Artificial Intelligence and Statistics. 2010:249256.

[1]	付鸿林, 张太红, 杨雅婷, 艾孜麦提·艾瓦尼尔, 马博. 基于生成对抗网络的维语场景文字修改网络[J]. 计算机与现代化, 2024, 0(01): 41-46.
[2]	王秋忆, 周浩, 郑婷婷. 改进RetinaNet的电力设备目标检测方法[J]. 计算机与现代化, 2024, 0(01): 47-52.
[3]	林启钊, 彭志平, 郭棉, 崔得龙. 基于双向多步预测的炉管温度场重构方法[J]. 计算机与现代化, 2024, 0(01): 53-58.
[4]	郑立瑞, 肖晓霞, 邹北骥, 刘彬, 周展. 基于BERT的电子病历命名实体识别[J]. 计算机与现代化, 2024, 0(01): 87-91.
[5]	李颖颖, 黄文培. 基于优化八叉树的场景视锥体裁剪算法[J]. 计算机与现代化, 2024, 0(01): 103-108.
[6]	夏千涵, 何胜煌, 吴元清, 赵乐乐. 基于可学习记忆特征金字塔网络的小样本目标检测[J]. 计算机与现代化, 2023, 0(12): 7-13.
[7]	周成诚, 曾庆军, 杨康, 胡家铭, 韩春伟. 基于高效通道注意力模块的运动想象脑电识别[J]. 计算机与现代化, 2023, 0(12): 19-23.
[8]	曾伟平, 陈俊洪, Muhammad ASIM, 刘文印, 杨振国. 基于多阶段分形组合的点云补全算法[J]. 计算机与现代化, 2023, 0(12): 24-29.
[9]	白晓波, 江梦茜, 王铁山, 邵景峰, 李勃, . 聚类质心与指数递减方法改进的哈里斯鹰算法[J]. 计算机与现代化, 2023, 0(12): 30-35.
[10]	邱凯星, 冯广. 基于双重特征注意力的多标签图像分类模型[J]. 计算机与现代化, 2023, 0(12): 41-47.
[11]	杜康, 郭鲁钰, 徐啟蕾, 单宝明, 张方坤. 基于模型种群分析变量选择的红外光谱建模方法[J]. 计算机与现代化, 2023, 0(12): 48-52.
[12]	刘语珵, 贺奇, 董延华, 王晓宇. 结合时间相关度与课程搭配度的课程推荐方法[J]. 计算机与现代化, 2023, 0(12): 53-58.
[13]	张浩洋, 尹梓名, 乐珺怡, 沈达聪, 束翌俊, 杨自逸, 孔祥勇, 龚伟. 3D-SPRNet: 一种基于并行解码器和双注意力机制的胆囊癌分割模型[J]. 计算机与现代化, 2023, 0(12): 59-66.
[14]	张伯泉, 麦海鹏, 陈嘉敏, 逄锦聚. 基于高灰度值注意力机制的脑白质高信号分割[J]. 计算机与现代化, 2023, 0(12): 67-75.
[15]	张在成, 李健. 改进的神经渲染方法在建筑施工场景中的应用[J]. 计算机与现代化, 2023, 0(12): 76-81.

基于Gate机制与BiLSTMCRF的汉语语义角色标注

#br# Chinese Semantic Role Labeling Based on Gated Mechanism and BiLSTMCRF

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