基于错误驱动的语义文法自动扩展学习方法研究

doi:10.12263/DZXB.20190159

电子学报 ›› 2021, Vol. 49 ›› Issue (2): 248-259.DOI: 10.12263/DZXB.20190159

基于错误驱动的语义文法自动扩展学习方法研究

王东升¹, 王卫民¹, 祁云松¹, 王石², 曹存根²

1. 江苏科技大学计算机学院, 江苏镇江 212003;
2. 中国科学院计算技术研究所中科院智能信息处理重点实验室, 北京 100190

收稿日期:2019-01-28 修回日期:2020-11-03 出版日期:2021-02-25
- 作者简介:
- 王东升 男,1982年生,江苏盐城人,现为江苏科技大学计算机学院副教授.主要研究方向为问答系统、知识图谱和自然语言处理.E-mail:jsjxy_wds@just.edu.cn;王卫民 男,1977年生,浙江绍兴人,现为江苏科技大学计算机学院讲师、博士.主要研究方向为问答系统和自然语言处理.E-mail:wangweimin@just.edu.cn;祁云嵩 男,1967年生,江苏如皋人,博士,现为江苏科技大学计算机学院教授.主要研究方向机器学习理论与应用,装备综合保障.E-mail:mailqys@163.com;王石男,1981年生,山东博兴人,现为中国科学院计算技术研究所智能信息处理重点实验室副研究员.主要研究方向为问答系统、知识图谱和自然语言处理.E-mail:wangshi@ict.ac.cn;曹存根 男,1964年生,江苏东台人,现为中国科学院计算技术研究所智能信息处理重点实验室研究员.主要研究方向为大规模知识工程.E-mail:cgcao@ict.ac.cn
- 基金资助:
- 国家自然科学基金 (No.61702234）; 科技部重点研发项目 (No.2017YFC1700302）

Automatic Error-driven Learning Method of Semantic Grammar

WANG Dong-sheng¹, WANG Wei-min¹, QI Yun-song¹, WANG Shi², CAO Cun-gen²

1. School of Computer Science, Jiangsu University of Science of Technology, Zhenjiang, Jiangsu 212003, China;
2. Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China

Received:2019-01-28 Revised:2020-11-03 Online:2021-02-25 Published:2021-02-25

摘要/Abstract

摘要： 面向领域的自然语言理解技术是垂直搜索引擎、领域相关问答系统等应用的核心技术之一.本文在已构建的基于本体和语义文法的自然语言理解系统的基础上，提出一种基于错误驱动的语义文法自动扩展学习方法，对于解析错误的句子，利用核心文法生成部分解析树，按照打分函数选择一组最佳的部分解析树，利用预测模型预测部分解析树的上层节点并试图构建完整的解析树，从而学习得到新的文法规则，对于学习得到的不同类型的规则进行验证并更新核心文法库，通过对句子的可学习性度量来筛选学习对象，从而提高文法扩展学习的整体质量和效率.分别在两个不同规模的领域数据集进行了测试，在交互式学习范式下，测试对比了学习算法在不同规模领域的学习效率，在批量学习范式下，测试对比了更新后的文法和核心文法在两个领域数据集上的准确率和识别率等性能指标.实验结果表明，本文所提出的方法是有效的.

关键词: 语义文法, 文法扩展, 自然语言理解, 领域, 本体

Abstract: Domain-specific natural language understanding technology is one of the core technology of vertical search engines,domain-specific question answering system and other applications.This research focus on a novel constrained semantic grammar and its automatic learning methods based on an existing domain-specific question answering system.An error-driven learning method of semantic grammar is proposed.The method first partially parses the ungrammatical sentence based on the core semantic grammar,then it attempts to build a complete parse tree,including predicting the top-level node of the partial parsing tree,generating and verifying hypotheses of new grammar rules.Learnability metrics is used to filter sentences in the training corpus to improve the overall quality and efficiency of grammar extending algorithm.The proposed algorithm is applied to two domains of different scales.In the interactive learning paradigm,learning efficiency are compared in different domains.In the batch learning paradigm,the paper compares the accuracy,MRR and recognition rate of the extended grammar and core grammar on twodatasets.The test results show that the proposed method is effective.

Key words: semantic grammar, grammar extending, natural language understanding, domain specific, ontology

中图分类号:

TP311

王东升, 王卫民, 祁云松, 王石, 曹存根. 基于错误驱动的语义文法自动扩展学习方法研究[J]. 电子学报, 2021, 49(2): 248-259.

WANG Dong-sheng, WANG Wei-min, QI Yun-song, WANG Shi, CAO Cun-gen. Automatic Error-driven Learning Method of Semantic Grammar[J]. Acta Electronica Sinica, 2021, 49(2): 248-259.

参考文献

[1] Hua W,Wang Z,Wang H,et al.Short text understanding through lexical-semantic analysis[A].2015 IEEE 31st International Conference on Data Engineering (ICDE).Piscataway[C].NJ:IEEE,2015.495-506.
[2] Khanam S A,Liu F,Chen Y P P.Comprehensive structured knowledge base system construction with natural language presentation[J].Human-centric Computing and Information Sciences,2019,9(1):23-30.
[3] 周国栋,李军辉.中文信息处理发展报告(2016)[R].北京:中国中文信息学会,2016,15-16.
[4] Kunneman F,Ferreira T C,Krahmer E,et al.Question similarity in community question answering:A systematic exploration of preprocessing methods and models[A].Proceedings of the International Conference on Recent Advances in Natural Language Processing[C].Varna,Bulgaria:INCOMA Ltd.,2019.593-601.
[5] Choi E,Bahadori M T,Song L,et al.GRAM:Graph-based attention model for healthcare representation learning[J].arXiv Preprint,2016,arXiv:1611.07012.
[6] Cui W,Xiao Y,Wang H,et al.KBQA:learning question answering over QA corpora and knowledge bases[J].arXiv Preprint,2019,arXiv:1903.02419.
[7] Ferrone L,Zanzotto F M.Symbolic,Distributed and distributional representations for natural language processing in the era of deep learning:a survey[J].arXiv Preprint,2017,arXiv:1702.00764.
[8] Fernández A M.Closed-domain Natural Language Approaches:Methods and Applications[M].España,ES:Editorial de la Universidad de Granada,2014,19-23.
[9] R.R.Burton.Semantic Grammar:An Engineering Technique for Constructing Natural Language Understanding Systems[R].Cambridge,Mass,UK:BBN Report #3453,Bolt,Beranek,and Newman,1976.
[10] 陈肇雄.SC文法功能体系[J].计算机学报,1992,15(11):801-808. Chen Zhaoxiong.A new context-sensitive subcategory (SC) grammar for machine translation[J].Chinese Journal of Computers,1992,15(11):801-808.(in Chinese)
[11] Steedman M.14 Combinatory categorial grammar[J].Current Approaches to Syntax:A Comparative Handbook,2019,3:389-416.
[12] Liang P,Jordan M I,Klein D.Learning dependency-based compositionalsemantics[A].Proceedings of the 49th An-nual Meeting of the Association for Computational Lin-guistics[C].Portland,Oregon,USA:Association for Computational Linguistics,2011.590-599.
[13] Titov I,Klementiev A.A Bayesian model for unsupervised semantic parsing[A].Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics[C].Portland,Oregon,USA:Association for Computational Linguistics,2011.1445-1455.
[14] P.Liang,Learning executable semantic parsers for natural languageunderstanding[J].Commun ACM,2016,59(9):68-76.
[15] Kate R J,Wong Y W,Mooney R J.Learning to transform natural to formal languages[A].Proceedings of the Na-tional Conference on Artificial Intelligence[C].Pittsburgh,Pennsylvania,USA:AAAI Press,2005.1062-1068.
[16] Karlsson F.Constraint grammar as a framework for parsing running text[A].Proceedings of the 13th Conference on Computational Linguistics[C].Finland:Association for Computational Linguistics,1990.168-173.
[17] Bick E,Didriksen T.CG-3-beyond classical constraint grammar[A].Proceedings of the 20th Nordic Conference of Computational Linguistics (NODALIDA 2015)[C].Vilnius,Lithuania:Linköping University Electronic Press,2015.31-39.
[18] L.Antonsen,S.Huhmarniemi,T.Trosterud.Constraint grammar in dialogue systems[A].Proceedings of the NODALIDA 2009 Workshop Constraint Grammar and Robust Parsing[C].Odense,Denmark:NEALT,2009.13-14.
[19] Mitchell T,Cohen W,Hruschka E,et al.Never-ending learning[J].Communications of the ACM,2018,61(5):103-115.
[20] Ibañez R,Soria Á,Teyseyre A,et al.Approximate string matching:A lightweight approach to recognize gestures with Kinect[J].Pattern Recognition,2017,62:73-86.
[21] 郑黎晓,王成.XML模式推断研究综述[J].电子学报,2016,44(2):461-471. Zheng Xiao-li,Wang Cheng.Schema inference from XML data:a review[J].Acta Electronica Sinica,2016,44(2):461-471.(in Chinese)
[22] Dahl V,Tessaris S,Bispo M D S.Parsing as semantically guided constraint solving:the role of ontologies[J].Annals of Mathematics and Artificial Intelligence,2018,82(1-3):1-25.
[23] Rosé C P.A framework for robust semantic interpretation[A].Proceedings of the 1st North American Chapter of the Association for Computational Linguistics Conference[C].Minneapolis,Minnesota:Association for Computational Linguistics,2000.311-318.
[24] Portele T.Data-driven classification of linguistic styles in spoken dialogues[A].Proceedings of the 19th International Conference on Computational Linguistics[C].Taipei,China:Association for Computational Linguistics,2002.1-7.
[25] Wang Y Y,Acero A,Chelba C,et al.Combination of statistical and rule-based approaches for spoken language understanding[A].Proceedings of Seventh International Conference on Spoken Language Processing[C].Valletta,Malta:European Language Resources Association (ELRA),2002.609-612.
[26] Gil Y.Human tutorial instruction in the raw[J].ACM Transactions on Interactive Intelligent Systems (TiiS),2015,5(1):2-8.
[27] Wang Y Y,Deng L,Acero A.Semantic frame-based spoken language understanding[J].Spoken Language Understanding:Systems for Extracting Semantic Information from Speech,2011,22(5):41-91.
[28] Lake B M,Ullman T D,Tenenbaum J B,et al.Building machines that learn and think like people[J].Behavioral and Brain Sciences,2017,40:1-70.
[29] B?k K,Zayan D,Czarnecki K,et al.Example-driven modeling:model=abstractions+examples[A].Proceedings of the 2013 International Conference on Software Engineering[C].New York,USA:IEEE Press,2013.1273-1276.
[30] López-Fernández J J,Cuadrado J S,Guerra E,et al.Example-driven meta-model development[J].Software & Systems Modeling,2015,14(4):1323-1347.
[31] Dongsheng Wang.Answering contextual questions based on ontology and question templates[J].Frontier of Computer Science in China,2011,5(4):405-418.
[32] M.Gavaldà and A.Waibel.Growing semantic grammars[A].Proc 36th Ann Meeting of the Assoc[C].Computational Linguistics,Montreal,Quebec,Canada:Association for Computational Linguistics,1998.451-456.
[33] Liang P,Potts C.Bringing machine learning and compositional semantics together[J].Annu Rev Linguist,2015,1(1):355-376.

基于错误驱动的语义文法自动扩展学习方法研究

Automatic Error-driven Learning Method of Semantic Grammar

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

[1]	张丽佳, 彭玺, 宋朝晖, 何杰, 刘克. 国家自然科学基金信息领域重点国际（地区）合作研究项目资助情况分析[J]. 电子学报, 2022, 50(6): 1487-1491.
[2]	王格格, 郭涛, 余游, 苏菡. 基于生成对抗网络的无监督域适应分类模型[J]. 电子学报, 2020, 48(6): 1190-1197.
[3]	黄子杰, 陈军华, 高建华. Code Smell视角下分层Web应用失血及充血现象的量化分析[J]. 电子学报, 2020, 48(4): 772-780.
[4]	刘真, 田靖玉, 苑宝鑫, 孙永奇. 基于知识聚合和迁移的跨领域推荐算法[J]. 电子学报, 2020, 48(10): 1928-1932.
[5]	李林峰, 刘真, 魏港明, 任爽, 葛梦凡. 基于共享知识模型的跨领域推荐算法[J]. 电子学报, 2018, 46(8): 1947-1953.
[6]	付雪峰, 漆桂林, 张勇. 一种基于图的DL-Lite本体最小不可满足保持子集的计算方法[J]. 电子学报, 2016, 44(9): 2040-2045.
[7]	黄颖, 何克清, 冯在文, 黄贻望. 基于本体的业务流程适应性配置方法研究[J]. 电子学报, 2016, 44(3): 699-708.
[8]	田刚, 何克清, 王健, 孙承爱, 徐建建. 面向领域标签辅助的服务聚类方法[J]. 电子学报, 2015, 43(7): 1266-1274.
[9]	苏雪阳, 左万利, 王俊华. 基于本体与模式的网络用户兴趣挖掘[J]. 电子学报, 2014, 42(8): 1556-1563.
[10]	薛涵, 秦兵, 刘挺. 基于Folksonomy的本体构建综述[J]. 电子学报, 2014, 42(4): 791-797.
[11]	黄霞. 基于本体和奇异值分解的图像数据索引查询算法[J]. 电子学报, 2014, 42(2): 288-291.
[12]	欧阳柳波, 邹北骥, 刘丽杰. 一种基于混合判定模型的复合概念抽取方法[J]. 电子学报, 2013, 41(3): 488-495.
[13]	张波, 向阳, 黄震华. 基于本体的决策问题语义理解及精炼方法[J]. 电子学报, 2012, 40(8): 1603-1608.
[14]	刘秀磊, 廖建新, 朱晓民, 杨迪, 徐童. 本体匹配中基于词义组合的词法分析算法[J]. 电子学报, 2012, 40(8): 1624-1630.
[15]	杨汝涛, 张绍谦, 窦万春. 一种基于QoS剪枝的Top-k自动服务组合方法[J]. 电子学报, 2012, 40(7): 1489-1491.