基于多层局部推理的汉语篇章关系及主次联合识别

摘要
参考文献
相关文章 (15)

全文: PDF (1312 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要篇章关系识别是篇章分析的核心组成部分。汉语中,缺少显式连接词的隐式篇章关系占比很高,篇章关系识别更具挑战性。该文给出了一个基于多层局部推理的汉语篇章关系及主次联合识别方法。该方法借助双向LSTM和多头自注意力机制进行篇章关系对应论元的表征;进一步借助软对齐方式获取论元间局部语义的推理权重,形成论元间交互语义信息的表征;再将两类信息结合进行篇章关系的局部推理,并通过堆叠多层局部推理部件构建了汉语篇章关系及主次联合识别框架,在CDTB语料库上的关系识别F₁值达到了67.0%。该文进一步将该联合识别模块嵌入一个基于转移的篇章解析器,在自动生成的篇章结构下进行篇章关系及主次的联合分析,形成了完整的汉语篇章解析器。

	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
	邢雨青
	孔芳

关键词 ：篇章分析, 篇章关系, 局部推理

Abstract：Discourse relation recognition plays a crucial part in discourse parsing. In Chinese, the task is much more challenging due to the high proportion of implicit discourse relations without explicit connectives as inference clues. This paper proposed a multi-layer local inference method for Chinese Discourse Relation Recognition. It employs bi-directional LSTM and multi-head self-attention mechanism to encode independent arguments, and then generate interactive pair representations using soft alignment between arguments achieved with soft attention. Both independent representations and interactive representations are then combined to perform local inference. By stacking the above local inference modules in our framework, we achieve 67.0% in Macro-F₁ value on CDTB corpus. Furthermore, a full automatic discourse parser is established by incorporating our trained model into an existing transition-based Chinese discourse parser, which can jointly learn the discourse relation and nuclearity.

Key words： discourse parsing discourse relation local inference

收稿日期: 2020-05-29

中图分类号:

TP391

作者简介: 邢雨青(1995—),硕士,主要研究领域为自然语言处理。E-mail:yqxing@stu.suda.edu.cn
孔芳(1977—),通信作者,博士,教授,博士生导师,主要研究领域为机器学习、自然语言处理、篇章分析。E-mail:kongfang@suda.edu.cn

链接本文:

http://jcip.cipsc.org.cn/CN/Y2022/V36/I7/42

[1] Naman G, Jacob E. A joint model of rhetorical discourse structure and summarization[C]//Proceedings of the Workshop on Structured Prediction for NLP. Austin, TX:Association for Computational Linguistics, 2016: 25-34.
[2] Choi E, Rashkin H, Zettlemoyer L, et al. Document-level sentiment inference with social, faction, and discourse context[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. Berlin, Germany, 2016: 333-343.
[3] Yang F J, Noah A S. Neural discourse structure for text categorization[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics. Vancouver, Canada, 2017: 996-1005.
[4] Prasad R, Dinesh N, Lee A, et al. The Penn discourse TreeBank 2.0[C]//Proceedings of the 6th International Conference on Language Resources and Evaluation. Marrakech, Morocco, 2008.
[5] Mann W. Rhetorical structure theory: Toward a functional theory of text organization[J]. Text & Talk, 2009, 8(3): 243-281.
[6] 李艳翠. 汉语篇章结构表示体系及资源构建研究[D].苏州: 苏州大学硕士学位论文, 2015.
[7] Marcu D, Echihabi A. An unsupervised approach to recognizing discourse relations[C]//Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics.Philadelphia, Pennsylvania, USA: Association for Computational Linguistics,2002: 368-375.
[8] Saito M, Yamamoto K, Sekine S. Using phrasal patterns to identify discourse relations[C]//Proceedings of the Human Language Technology Conference of the NAACL. New York City, USA,2006: 133-136.
[9] Pitler E, Ani N. Using syntax to disambiguate explicit discourse connectives in text[C]//Proceedings of the ACL-IJCNLP Conference Short Papers.Suntec, Singapore,2009: 13-16.
[10] Lin Z, Hwee Tou N G, Kan M Y. A PDTB-styled end-to-end discourse parser[J]. Natural Language Engineering, 2014, 20: 151-184.
[11] 徐凡,朱巧明,周国栋. 基于树核的隐式篇章关系识别[J]. 软件学报, 2013, 24(5): 1022-1035.
[12] Kong F, Zhou G. A CDT-styled end-to-end Chinesediscourse parser[J]. ACM Transactions On Asianand Low-Resource Language Information Processing, 2017, 16(4): 26.1-26.17.
[13] Bai H, Zhao H. Deep enhanced representation for implicit discourse relation recognition[C]//Proceedings of the 27th International Conference on Computational Linguistics. Santa Fe, New Mexico, USA, 2018: 571-583.
[14] Liu Y, Li S. Recognizing implicit discourse relations via repeated reading: neural networks with multi-level attention[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing. Austin, Texas, 2016: 1334-1233.
[15] Lan M, Wang J, Wu Y, et al. Multi-task attention-based neural networks for implicit discourse relationship representation and identification [C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing. Copenhagen, Denmark, 2017: 1299-1308.
[16] Guo F, He R, Di Jin,et al. Implicit discourse relation recognition using neural tensor network with interactive attention and sparse learning[C]//Proceedings of the 27th International Conference on Computational Linguistics. Santa Fe, New Mexico, USA, 2018: 547-558.
[17] Yudai K, Yugo M, Sadao K. A knowledge-augmented neural network model for implicit discourse relation classification[C]//Proceedings of the 27th International Conference on Computational Linguistics. Santa Fe, New Mexico, USA. August,2018: 584-595.
[18] RNnqvist, S, Schenk N, Chiarcos C. A recurrent neural model with attention for the recognition of Chinese implicit discourse relations[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics. Vancouver, Canada, 2017: 256-262.
[19] Peters M E, Neumann M, Iyyer M, et al. Deep contextualized word representations[C]//Proceedings of the Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. New Orleans, Louisiana, 2018: 2227-2237.
[20] Xu S, Li P, Kong F, et al. Topic tensor network for implicit discourse relation recognition in Chinese[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Florence, Italy, 2019: 608-618.
[21] Wang Y, Li S, Wang H. A two-stage parsing method for text-level discourse analysis[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics. Vancouver, Canada, 2017: 184-188.
[22] 孙成,孔芳. 基于转移的中文篇章结构解析研究[J]. 中文信息学报, 2018, 32(12): 48-56.