使用有序词语移动距离特征进行中文文本蕴含识别

doi:10.13190/j.jbupt.2016-221

北京邮电大学学报 ›› 2017, Vol. 40 ›› Issue (5): 123-128.doi: 10.13190/j.jbupt.2016-221

使用有序词语移动距离特征进行中文文本蕴含识别

谭咏梅, 王敏达, 牛少彰

北京邮电大学计算机学院, 北京 100876

收稿日期:2016-09-07 出版日期:2017-10-28 发布日期:2017-11-21
作者简介:谭咏梅(1975-),女,副教授,硕士生导师,E-mail:ymtan@bupt.edu.cn.
基金资助:
国家自然科学基金项目（U1536121，61370195）

Chinese Textual Entailment Recognition Via Ordered Word Mover Distance

TAN Yong-mei, WANG Min-da, NIU Shao-zhang

School of Computer Science, Beijing University of Posts and Telecommunications, Beijing 100876, China

Received:2016-09-07 Online:2017-10-28 Published:2017-11-21

摘要/Abstract

摘要： 提出了一种基于有序词语移动距离的中文文本蕴含识别方法，该方法基于word2vec词向量计算有序词语移动距离特征，进而利用有序词语移动距离特征和传统语言学特征通过支持向量机生成分类模型，然后使用分类模型进行蕴含识别，最终得到蕴含结果.该方法在RITE-VAL评测任务的CS数据上的MacroF1为0.629，超过RITE-VAL的最优评测结果（BUPTTeam，0.615）.实验结果表明，该方法可以提升中文文本蕴含识别系统的性能.

关键词: 文本蕴含, word2vec, 有序词语移动距离, SVM

Abstract: Chinese textual entailment recognition method based on ordered word mover distance was proposed. The ordered word mover distance was computed based on word2vec. The ordered word mover distance feature, grammar feature, and semantic feature were used to generate classification module based support vector machine (SVM). With use of classification module, the entailment result was obtained. An experiment was conducted in the CS data of RITE-VAL evaluation task in 2014, the MacroF1 of the experiment is 0.629, outperforming optimal value (BUPTTeam,0.615), which illustrates the effectiveness of the method to lifting the performance of Chinese textual entailment.

Key words: textual entailment, word2vec, ordered word mover distance, support vector machine

中图分类号:

TN911.22

谭咏梅, 王敏达, 牛少彰. 使用有序词语移动距离特征进行中文文本蕴含识别[J]. 北京邮电大学学报, 2017, 40(5): 123-128.

TAN Yong-mei, WANG Min-da, NIU Shao-zhang. Chinese Textual Entailment Recognition Via Ordered Word Mover Distance[J]. JOURNAL OF BEIJING UNIVERSITY OF POSTS AND TELECOM, 2017, 40(5): 123-128.

参考文献

[1] 袁毓林, 王明华. 文本蕴含的推理模型与识别模型[J]. 中文信息学报, 2010, 24(2):3-13. Yuan Yulin, Wang Minghua. The inference and identification models for textual entailment[J]. Journal of Chinese Information Processing, 2010, 24(2):3-13.
[2] Matsuyoshi S, Miyao Y, Shibata T, et al. Overview of the NTCIR-11 recognizing inference in text and validation (RITE-VAL) task[C]//Proceedings of the 11^th NTCIR Conference. Tokyo:[s.n.], 2014:223-232.
[3] 王宝鑫, 郑德权, 王晓雪, 等. 基于文本蕴含的选择类问题解答技术研究[J]. 北京大学学报:自然科学版, 2016, 52(1):134-140. Wang Baoxin, Zheng Dequan, Wang Xiaoxue, et al. Multiple-choice question answering based on textual entailment[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 2016, 52(1):134-140.
[4] Liu Maofu, Zhang Luming, Hu Huijun, et al. A classification model for semantic entailment recognition with feature combination[J]. Neurocomputing, 2016, 208(C):127-135.
[5] 张晗, 盛雅琦, 吕晨, 等. 基于短文本隐含语义特征的文本蕴涵识别[J]. 中文信息学报, 2016, 30(3):163-171. Zhang Han, Sheng Yaqi, Lü Chen, et al. Recognizing textual entailment based on short text latent semantic feature[J]. Journal of Chinese Information Processing, 2016, 30(3):163-171.
[6] Siblini R, Kosseim L. Using ontology alignment for the TAC RTE challenge[C]//Text Analysis Conference (TAC 2008) Workshop-RTE-4 Track. Gaithersburg:[s.n.], 2008.
[7] Mikolov T, Sutskever I, Chen K, et al. Distributed representations of words and phrases and their compositionality[C]//Advances in Neural Information Processing Systems, 2013:3111-3119.
[8] Kusner M J, Sun Y, Kolkin N I, et al. From word embeddings to document distances[C]//International Conference on Machine Learning. Lille:[s.n.], 2015:957-966.
[9] Vapnik V. The Nature of statistical learning theory[M]. Berlin:Springer, 1995.
[10] 张学工. 关于统计学习理论与支持向量机[J]. 自动化学报, 2000, 26(1):3243. Zhang Xuegong. Introduction to statistical learning theory and support vector machines[J]. Acta Automatica Sinica, 2000, 26(1):3243.
[11] Mikolov T, Chen K, Corrado G, et al. Efficient estimation of word representations in vector space[EB/OL]. 2013. https://arxiv.org/abs/1301.3781.
[12] 李妍, 刘茂福, 姬东鸿. 基于支持向量机的中文文本蕴涵识别研究[J]. 计算机应用与软件, 2014, 31(4):51-55. Li Yan, Liu Maofu, Ji Donghong. On SVM-based Chinese textual entailment recognition[J]. Computer Application and Software, 2014, 31(4):51-55.

使用有序词语移动距离特征进行中文文本蕴含识别

Chinese Textual Entailment Recognition Via Ordered Word Mover Distance

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

编辑推荐

Metrics

本文评价