基于CNN与双向LSTM的中文文本蕴含识别方法

摘要
参考文献
相关文章 (14)

全文: PDF (2526 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要为了避免基于传统机器学习的中文文本蕴含识别方法需要人工筛选大量特征以及使用多种自然语言处理工具造成的错误累计问题,该文提出了基于CNN与双向LSTM的中文文本蕴含识别方法。该方法使用CNN与双向LSTM分别对句子进行编码,自动提取相关特征,然后使用全连接层进行分类得到初步的识别结果,最后使用语义规则对网络识别结果进行修正,得到最终的蕴含识别结果。在2014年RITE-VAL评测任务的数据集上MacroF1结果为61.74%,超过评测第一名的结果61.51%。实验结果表明,该方法对于中文文本蕴含识别是有效的。

	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
	谭咏梅
	刘姝雯
	吕学强

关键词 ：中文文本蕴含, 卷积神经网络, 双向长短时记忆网络

Abstract：A CNN and BiLSTM based Chinese textual entailment recognition method is presented. By using CNN and BiLSTM, the method can automatically extract relevant features, and then generate the initial result by a fully connected layer. The final result is further processed by semantic rules. Evaluated on the dataset of RITE-VAL in 2014, the method obtains 61.74%, outperforming the top-ranked 61.51% in that evaluation campaign.

Key words： Chinese textual entailment convolutional neural network bidirectional long short-term memory network

收稿日期: 2017-07-18

中图分类号:

TP391

基金资助:网络文化与数字传播北京市重点实验室开放课题(ICDD201703);国家自然科学基金(61671070)

作者简介: 谭咏梅(1975—),通信作者,博士,副教授,主要研究领域为计算语言学、自然语言处理、机器学习、知识工程。E-mail:ymtan@bupt.edu.cn
刘姝雯(1993—),硕士,主要研究领域为自然语言处理、机器学习。E-mail:swanliu@bupt.edu.cn
吕学强(1970—),博士,教授,主要研究领域为中文信息处理、多媒体信息处理。E-mail:lxq@bistu.edu.cn

链接本文:

http://jcip.cipsc.org.cn/CN/Y2018/V32/I7/11

[1] Dagan Ido, Glickman Oren, Magnini Bernardo. The PASCAL recognising textual entailment challenge[C] //Proceedings of the International Conference on Machine Learning Challenges: Evaluating Predictive Uncertainty Visual Object Classification, and Recognizing Textual Entailment. Springer-Verlag, 2006: 177-190.
[2] 郭茂盛, 张宇, 刘挺. 文本蕴含关系识别与知识获取研究进展及展望[J]. 计算机学报, 2017, 40(4): 889-910.
[3] Jagannatha Abhyuday N, Yu Hong. Structured prediction models for RNN based sequence labeling in clinical text[C]//Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, 2016: 856.
[4] Lample Guillaume, Ballesteros Miguel, Subramanian Sandeep, et al. Neural architectures for named entity recognition[C]//Proceedings of the Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2016: 260-270.
[5] Tan Yongmei, Wang Minda, Wang Xiaohui, et al. Recognizing textual entailment using multiple features and filters[C]//Proceedings of the 11th NTCIR Conference, 2014.
[6] 刘茂福, 李妍, 姬东鸿. 基于事件语义特征的中文文本蕴含识别[J]. 中文信息学报, 2013, 27(5): 129-136.
[7] Shima Hideki, Kanayama Hiroshi, Lee Cheng-wei, et al. Overview of NTCIR-9 RITE: recognizing inference in text[C]//Proceedings of the NTCIR, 2011.
[8] Watanabe Yotaro, Miyao Yusuke, Mizuno Junta, et al. Overview of the recognizing inference in text(RITE-2) at NTCIR-10[C]//Proceedings of the 10^th NTCIR Conference, 2013.
[9] Matsuyoshi Suguru, Miyao Yusuke, Shibata Tomohide, et al. Overview of the NTCIR-11 recognizing inference in text and validation (RITE-VAL) Task[C]//Proceedings of the 11^th NTCIR Conference, 2014.
[10] Bar-Haim Roy, Berant Jonathan, Dagan Ido. A compact forest for scalable inference over entailment and paraphrase rules[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, EMNLP 2009, 2009: 1056-1065.
[11] Alabbas Maytham, Ramsay Allan. Natural language inference for Arabic using extended tree edit distance with subtrees[J]. Journal of Artificial Intelligence Research, 2014, 48(48): 1-22.
[12] Wang Xiaolin, Zhao Hai, Lu Baoliang. BCMI-NLP labeled-alignment-based entailment system for NTCIR-10 RITE-2 Task[C]//Proceedings of the 10^th NTCIR Conference, 2013.
[13] 王宝鑫. 基于深度神经网络的文本蕴含识别及应用研究[D]. 哈尔滨: 哈尔滨工业大学硕士学位论文, 2016.
[14] 王志浩. 文本蕴含识别研究[D]. 北京: 北京邮电大学硕士学位论文, 2015.
[15] Hubel David Hunter, Wiesel Torsten Nils. Receptive fields, binocular interaction and functional architecture in the cat's visual cortex[J]. Journal of Physiology, 1962, 160(1): 106-154.
[16] LeCun Yann, Bottou Léon, Bengio Yoshua, et al. Gradient-based learning applied to document Recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
[17] Yoon Kim. Convolutional neural networks for sentence classification[C]//Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2014.
[18] Collobert Ronan, Weston Jason, Bottou Léon, et al. Natural language processing (almost) from scratch[J]. Journal of Machine Learning Research, 2011, 12(1): 2493-2537.
[19] Hochreiter Sepp, Schmidhuber Jürgen. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780.
[20] Kolen John, Kremer Stefan. Gradient Flow in Recurrent Nets: the Difficulty of Learning Long-Term Dependencies[M]. Wiley-IEEE Press, 2007.
[21] Graves Alex, Schmidhuber Jürgen. 2005 Special Issue: Framewise phoneme classification with bidirectional LSTM and other neural network architectures[J]. Neural Networks, 2005, 18(5-6): 602-610.
[22] Zhang Zhichang, Yao Dongren, Mao Longlong, et al. NWNU minimum information recognizing entailment system for NTCIR-11 RITE-3 Task[C]//Proceedings of the 11^th NTCIR Conference, 2014.
[23] Wu Shihung, Hsu Lijen, Lin Huawei, et al. III & CYUT Chinese textual entailment recognition system for NTCIR-11 RITE-VAL[C]//Proceedings of the 11^th NTCIR Conference, 2014.