用于嵌套命名实体识别的边界强化分类模型

doi:10.19678/j.issn.1000-3428.0062181

摘要/Abstract

摘要： 实体嵌套是自然语言中一种常见现象，提高嵌套命名实体识别的准确性对自然语言处理各项任务具有重要作用。针对现有嵌套命名实体识别方法在识别实体边界时不够准确、未能有效利用实体边界信息等问题，提出一种嵌套命名实体识别的边界强化分类模型。采用卷积神经网络提取邻接词的特征，通过加入多头注意力的序列标注模型获取实体中的边界特征，提高实体边界检测的准确性。在此基础上，计算实体中各词语对实体类型的贡献度，将实体关键字与实体边界词相结合来表示实体，使实体表示中包含实体关键信息和边界信息，最后进行实体类型检测。实验结果表明，通过加入多头注意力机制能够有效提升对嵌套命名实体的检测和识别性能，该模型在GENIA和GermEval 2014数据集上准确率有较好表现，并且召回率和F1值较对比模型达到最优。

关键词: 嵌套命名实体识别, 实体表示, 注意力机制, 边界, 神经网络

Abstract: Entity nesting is a common phenomenon in natural language.Improving the accuracy of nested Named Entity Recognition(NER) plays an important role in various Natural Language Processing(NLP) tasks.Addressing the inaccuracy of existing nested NER methods for identifying entity boundaries and their ineffective use of entity boundary information, a boundary enhanced classification model for nested NER is proposed.A Convolution Neural Network(CNN) is used to extract features of adjacent words, and sequence annotation model with multi-heads attention is added to obtain the boundary features of entities and improve the accuracy of entity boundary detection.On this basis, the model first calculates the contribution of each word in the entity to each entity type, combines the entity keyword with each entity boundary word to represent the entity, makes the entity representation contain the entity key and boundary information, and finally determines the entity type.Experiments show that adding the multi-head attention mechanism effectively improves the performances of nested NER and recognition.The model demonstrated a good accuracy performance on the GENIA and GermEval 2014 datasets.In addition, of all models compared in the experiments, the proposed model achieved the best recall rate and F1-score.

Key words: nested Named Entity Recognition(NER), entity representation, attention mechanism, boundary, neural network

中图分类号:

TP391

连艺谋, 张英俊, 谢斌红. 用于嵌套命名实体识别的边界强化分类模型[J]. 计算机工程, 2022, 48(8): 313-320.

LIAN Yimou, ZHANG Yingjun, XIE Binhong. Boundary Enhanced Classification Model for Nested Named Entity Recognition[J]. Computer Engineering, 2022, 48(8): 313-320.

https://www.ecice06.com/CN/Y2022/V48/I8/313

图/表 10

20220825095008

20220825095011

20220825095015

20220825095018

20220825095022

20220825095026

20220825095031

20220825095034

20220825095038

20220825095041

参考文献

[1] CHINCHOR N, ROBINSON P.MUC-7 named entity task definition[C]//Proceedings of the 7th Conference on Message Understanding.Washington D.C., USA:IEEE Press, 1995:319-332.
[2] LAFFERTY J D, MCCALLUM A, PEREIRA F C N.Conditional random fields:probabilistic models for segmenting and labeling sequence data[C]//Proceedings of the 18th International Conference on Machine Learning.Berlin, Germany:Springer, 2001:282-289.
[3] GUILLAUME L, MIGUEL B, SANDEEP S, et al.Neural architectures for named entity recognition[C]//Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics.Stroudsburg, USA:Association for Computational Linguistics, 2016:260-270.
[4] HUANG Z, XU W, YU K.Bidirectional LSTM-CRF models for sequence tagging[EB/OL].[2021-05-12].https://arxiv.org/abs/1508.01991.
[5] SHEN D, ZHANG J, ZHOU G, et al.Effective adaptation of a hidden Markov model-based named entity recognizer for biomedical domain[C]//Proceedings of the ACL 2003 Workshop on Natural Language Processing.Stroudsburg, USA:Association for Computational Linguistics, 2003:49-56.
[6] ZHANG J, SHEN D, ZHOU G, et al.Enhancing HMM-based biomedical named entity recognition by studying special phenomena[J].Journal of Biomedical Informatics, 2004, 37(6):411-422.
[7] BORTHWICK A, GRISHMAN R.A maximum entropy approach to named entity recognition[D].New York, USA:New York University, 1999.
[8] LU W, ROTH D.Joint mention extraction and classification with mention hypergraphs[C]//Proceedings of 2015 Conference on Empirical Methods in Natural Language Processing.Stroudsburg, USA:Association for Computational Linguistics, 2015:857-867.
[9] WANG B, LU W.Neural segmental hypergraphs for overlapping mention recognition[C]//Proceedings of 2018 Conference on Empirical Methods in Natural Language Processing.Stroudsburg, USA:Association for Computational Linguistics, 2018:204-214.
[10] JU M, MIWA M, ANANIADOU S.A neural layered model for nested named entity recognition[C]//Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies.Stroudsburg, USA:Association for Computational Linguistics, 2018:1446-1459.
[11] LIN H, LU Y, HAN X, et al.Sequence-to-nuggets:nested entity mention detection via anchor-region networks[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics.Stroudsburg, USA:Association for Computational Linguistics, 2019:5182-5192.
[12] FISHER J, VLACHOS A.Merge and label:a novel neural network architecture for nested NER[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics.Stroudsburg, USA:Association for Computational Linguistics, 2019:5840-5850.
[13] SOHRAB M G, MIWA M.Deep exhaustive model for nested named entity recognition[C]//Proceedings of 2018 Conference on Empirical Methods in Natural Language Processing.Stroudsburg, USA:Association for Computational Linguistics, 2018:2843-2849.
[14] XU M, JIANG H, WATCHARAWITTAYAKUL S.A local detection approach for named entity recognition and mention detection[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics.Stroudsburg, USA:Association for Computational Linguistics, 2017:1237-1247.
[15] ZHENG C, CAI Y, XU J, et al.A boundary-aware neural model for nested named entity recognition[C]//Proceedings of 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing.Stroudsburg, USA:Association for Computational Linguistics, 2019:357-366.
[16] TAN C Q, QIU W, CHEN M S, et al.Boundary enhanced neural span classification for nested named entity recognition[C]//Proceedings of AAAI Conference on Artificial Intelligence.Palo Alto, USA:AAAI Press, 2020:9016-9023.
[17] MA X Z, HOVY E.End-to-end sequence labeling via Bi-directional LSTM-CNNs-CRF[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics.Stroudsburg, USA:Association for Computational Linguistics, 2016:1064-1074.
[18] WANG Y R, TIAN F.Recurrent residual learning for sequence classification[C]//Proceedings of 2016 Conference on Empirical Methods in Natural Language Processing.Stroudsburg, USA:Association for Computational Linguistics, 2016:938-943.
[19] 杨姗姗, 姜丽芬, 孙华志, 等.基于时间卷积网络的多项选择机器阅读理解[J].计算机工程, 2020, 46(11):97-103. YANG S S, JIANG L F, SUN H Z, et al.Multiple choice machine reading comprehension based on temporal convolutional network[J].Computer Engineering, 2020, 46(11):97-103.(in Chinese)
[20] JENNY R F, CHRISTOPHER D M.Nested named entity recognition[C]//Proceedings of 2009 Conference on Empirical Methods in Natural Language Processing.Stroudsburg, USA:Association for Computational Linguistics, 2009:141-150.
[21] MUIS A O, LU W.Labeling gaps between words:recognizing overlapping mentions with mention separators[C]//Proceedings of 2017 Conference on Empirical Methods in Natural Language Processing.Stroudsburg, USA:Association for Computational Linguistics, 2017:2608-2618.

选择文件类型/文献管理软件名称

选择包含的内容