一种面向法律文书的命名实体识别模型

doi:10.3969/j.issn.1671-1122.2024.11.017

摘要/Abstract

摘要：

准确识别法律文书中的实体是构建智慧司法的基础，但通用的命名实体识别模型不能很好地识别法律文书中实体边界，识别结果不能与法律业务紧密结合。为有效提高法律文书中各实体的识别效果，文章提出一种面向法律文书的命名实体识别模型BBAG-NER。该模型首先利用BERT对字符序列进行编码，然后运用双向长短记忆神经网络和Attention分配不同权重以提高对实体边界的划分能力，最后采用全局指针识别备选司法实体片段，并通过实体分类器得到最终的实体类别。实验结果表明，在法律文书语料数据集上，BBAG-NER模型的F1值达到了89.18%，较BERT-CRF模型提高了2.1%，验证了模型整体的有效性。

关键词: 法律文书, 命名实体识别, 全局指针网络, 双向长短时记忆

Abstract:

Accurate identification of entities in legal documents is fundamental for building an intelligent judicial system. However, generic Named Entity Recognition models often struggle with accurately recognizing entity boundaries in legal documents and integrating recognition results closely with legal practices. To improve the accuracy of entity recognition in legal documents, this paper proposed the BBAG-NER model for Named Entity Recognition in legal documents. The model first encoded character sequences using BERT, then employed Bidirectional Long Short-Term Memory and Attention mechanisms to assign different weights and enhance the ability to delineate entity boundaries. Finally, it used a global pointer network to identify potential judicial entity segments and obtained the final entity categories through an entity classifier. Experimental results on a legal document corpus dataset show that our proposed model achieves an F1 score of 89.18%, representing a 2.1% improvement compared to the BERT-CRF model, demonstrating the overall effectiveness of our proposed model.

Key words: legal documents, named entity recognition, global pointer network, BiLSTM

中图分类号:

TP309

卢睿, 李林瑛. 一种面向法律文书的命名实体识别模型[J]. 信息网络安全, 2024, 24(11): 1783-1792.

LU Rui, LI Linying. A Named Entity Recognition Model for Legal Documents[J]. Netinfo Security, 2024, 24(11): 1783-1792.

图/表 10

图1

图2

图3

图4

表1

表2

表3

表4

表5

表6

参考文献 36

[1]	HUANG Zhiheng, XU Wei, YU Kai. Bidirectional LSTM-CRF Models for Sequence Tagging[EB/OL]. [2024-05-08]. https://arxiv.org/pdf/1508.01991.
[2]	MA Xuezhe, HOVY E. End-to-End Sequence Labeling via Bi-Directional LSTM-CNNS-CRF[C]// ACL. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1:Long Papers). Stroudsburg: ACL, 2016: 1064-1074.
[3]	LAMPLE G, BALLESTEROS M, SUBRAMANIAN S, et al. Neural Architectures for Named Entity Recognition[C]// NAACL. Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies. Stroudsburg: NAACL, 2016: 260-270.
[4]	AKBIK A, BLYTHE D, VOLLGRAF R. Contextual String Embeddings for Sequence Labeling[C]// ACL. Proceedings of the 27th International Conference on Computational Linguistics. Stroudsburg: ACL, 2018: 1638-1649.
[5]	CONTRACTOR D, PATRA B, MAUSAM, et al. Constrained BERT BiLSTM CRF for Understanding Multi-Sentence Entity-Seeking Questions[J]. Natural Language Engineering, 2021, 27(1): 65-87.
[6]	XU Lei, LI Shuang, WANG Yuchen, et al. Named Entity Recognition of BERT-BiLSTM-CRF Combined with Self-Attention[C]// Springer. Lecture Notes in Computer Science. Heidelberg: Springer, 2021: 556-564.
[7]	ZHANG Qiang, SUN Yong, ZHANG Linlin, et al. Named Entity Recognition Method in Health Preserving Field Based on BERT[J]. Procedia Computer Science, 2021, 183: 212-220.
[8]	ZHANG Xinrui, LUO Xudong. A Machine-Reading-Comprehension Method for Named Entity Recognition in Legal Documents[C]// Springer. Communications in Computer and Information Science. Heidelberg: Springer, 2023: 224-236.
[9]	YUAN Zheng, TAN Chuanqi, HUANG Songfang, et al. Fusing Heterogeneous Factors with Triaffine Mechanism for Nested Named Entity Recognition[C]// ACL. Findings of the Association for Computational Linguistics:ACL 2022. Stroudsburg: ACL, 2022: 3174-3186.
[10]	SU Jianlin, MURTADHA A, PAN Shengfeng, et al. Global Pointer: Novel Efficient Span-Based Approach for Named Entity Recognition[EB/OL]. [2024-05-08]. https://arxiv.org/abs/2208.03054.
[11]	SUÁREZ-PANIAGUA V, RIVERA ZAVALA R M, SEGURA-BEDMAR I, et al. A Two-Stage Deep Learning Approach for Extracting Entities and Relationships from Medical Texts[EB/OL]. (2019-09-20)[2024-05-08]. https://www.sciencedirect.com/science/article/pii/S1532046419302047?via%3Dihub.
[12]	WU Jun, CHENG Yao, HAO Han, et al. Automatic Extraction of Chinese Terminology Based on BERT Embedding and BiLSTM-CRF Model[J]. Journal of the China Society for Scientific and Technical Information, 2020, 39(4): 409-418.
	吴俊, 程垚, 郝瀚, 等. 基于BERT嵌入BiLSTM-CRF模型的中文专业术语抽取研究[J]. 情报学报, 2020, 39(4): 409-418.
[13]	WANG Yue, WANG Mengxuan, ZHANG Sheng, et al. Alarm Text Named Entity Recognition Based on BERT[J]. Journal of Computer Applications, 2020, 40(2): 535-540. doi: 10.11772/j.issn.1001-9081.2019101717
	王月, 王孟轩, 张胜, 等. 基于BERT的警情文本命名实体识别[J]. 计算机应用, 2020, 40(2): 535-540. doi: 10.11772/j.issn.1001-9081.2019101717
[14]	FENG Luanluan, LI Junhui, LI Peifen, et al. Technology and Terminology Detection Oriented National Defense Science[J]. Computer Science, 2019, 46(12): 231-236. doi: 10.11896/jsjkx.190300069
	冯鸾鸾, 李军辉, 李培峰, 等. 面向国防科技领域的技术和术语识别方法研究[J]. 计算机科学, 2019, 46(12): 231-236. doi: 10.11896/jsjkx.190300069
[15]	LI Jinchen, LI Yanling, GE Fengpei, et al. Survey of Research on Intelligent System for Legal Domain[J]. Computer Engineering and Applications, 2023, 59(7): 31-50. doi: 10.3778/j.issn.1002-8331.2208-0383
	李瑾晨, 李艳玲, 葛凤培, 等. 面向法律领域的智能系统研究综述[J]. 计算机工程与应用, 2023, 59(7): 31-50. doi: 10.3778/j.issn.1002-8331.2208-0383
[16]	WANG Zhili, WU Yufan, LEI Pengbin, et al. Named Entity Recognition Method of Brazilian Legal Text Based on Pre-Training Model[J]. Journal of Physics: Conference Series, 2020, 1550(3): 1-5.
[17]	LEITNER E, REHM G, MORENO-SCHNEIDER J. Fine-Grained Named Entity Recognition in Legal Documents[C]// Springer. Lecture Notes in Computer Science. Heidelberg: Springer, 2019: 272-287.
[18]	WANG Xiao, WAN Yuqing. A Named Entity Identification Method for Legal Documents[J]. Computer Applications and Software, 2023, 40(8): 180-186.
	王霄, 万玉晴. 一种面向法律文书的命名实体识别方法[J]. 计算机应用与软件, 2023, 40(8): 180-186.
[19]	JI Donghong, TAO Peng, FEI Hao, et al. An End-to-End Joint Model for Evidence Information Extraction from Court Record Document[EB/OL]. (2020-05-29)[2024-05-08]. https://www.sciencedirect.com/science/article/abs/pii/S0306457320308001?via%3Dihub.
[20]	LI Chunnan, WANG Lei, SUN Yuanyuan, et al. BERT Based Named Entity Recognition for Legal Texts on Theft Cases[J]. Journal of Chinese Information Processing, 2021, 35(8): 73-81.
	李春楠, 王雷, 孙媛媛, 等. 基于BERT的盗窃罪法律文书命名实体识别方法[J]. 中文信息学报, 2021, 35(8): 73-81.
[21]	WANG Dexian, WANG Suge, PEI Wensheng, et al. Named Entity Recognition Based on JCWA-DLSTM for Legal Instruments[J]. Journal of Chinese Information Processing, 2020, 34(10): 51-58.
	王得贤, 王素格, 裴文生, 等. 基于JCWA-DLSTM的法律文书命名实体识别方法[J]. 中文信息学报, 2020, 34(10): 51-58.
[22]	GUO Lihua, LI Yang, WAGN Suge, et al. Name Entity Recognition in Legal Instruments Based on Matching Strategy and Community Attention Mechanism[J]. Journal of Chinese Information Processing, 2022, 36(2): 85-92.
	郭力华, 李旸, 王素格, 等. 基于匹配策略和社区注意力机制的法律文书命名实体识别[J]. 中文信息学报, 2022, 36(2): 85-92.
[23]	ZENG Lanlan, WANG Yisong, CHEN Panfeng. Named Entity Recognition Based on BERT and Joint Learning for Judgment Documents[J]. Journal of Computer Applications, 2022, 42(10): 3011-3017. doi: 10.11772/j.issn.1001-9081.2021091565
	曾兰兰, 王以松, 陈攀峰. 基于BERT和联合学习的裁判文书命名实体识别[J]. 计算机应用, 2022, 42(10): 3011-3017. doi: 10.11772/j.issn.1001-9081.2021091565
[24]	MAO Xingliang, CHEN Xiaohong, NING Ken, et al. Global and Local Information Integration for Recognizing Key Case Elements[J]. Journal of Software, 2023, 34(12): 5724-5736.
	毛星亮, 陈晓红, 宁肯, 等. 全局和局部信息融合的案情关键要素识别[J]. 软件学报, 2023, 34(12): 5724-5736.
[25]	NING Ken. A Study of Chinese Named Entity Recognition for Judicial Field[D]. Changsha: Central South University, 2022.
	宁肯. 面向司法领域的中文命名实体识别研究[D]. 长沙: 中南大学, 2022.
[26]	YANG Shuxin, LIU Tianyang, HUANG Weidong. Chinese Legal Document Named Entity Recognition Based on MVBCN-FLW[EB/OL]. [2024-01-02]. http://kns.cnki.net/kcms/detail/11.2127.tp.20231228.1537.016.html.
	杨书新, 刘天扬, 黄伟东. 基于MVBCN-FLW的中文法律文书命名实体识别[EB/OL]. [2024-01-02]. http://kns.cnki.net/kcms/detail/11.2127.tp.20231228.1537.016.html
[27]	CAI Li, WANG Shuting, LIU Junhui, et al. Survey of Data Annotation[J]. Journal of Software, 2020, 31(2): 302-320.
	蔡莉, 王淑婷, 刘俊晖, 等. 数据标注研究综述[J]. 软件学报, 2020, 31(2): 302-320.
[28]	LU Yaojie, LIU Qing, DAI Dai, et al. Unified Structure Generation for Universal Information Extraction[C]// ACL. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1:Long Papers). Stroudsburg: ACL, 2022: 5755-5772.
[29]	LI Jingye, FEI Hao, LIU Jiang, et al. Unified Named Entity Recognition as Word-Word Relation Classification[C]// AAAI. Proceedings of the AAAI Conference on Artificial Intelligence. MenloPark: AAAI, 2022: 10965-10973.
[30]	FU Jinlan, HUANG Xuanjing, LIU Pengfei. SpanNER: Named Entity Re-/Recognition as Span Prediction[C]// ACL. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1:Long Papers). Stroudsburg: ACL, 2021: 7183-7195.
[31]	HU Shulin, ZHANG Huajun, HU Xuesong, et al. Chinese Named Entity Recognition Based on BERT-CRF Model[C]// IEEE. 2022 IEEE/ACIS 22nd International Conference on Computer and Information Science (ICIS). New York: IEEE, 2022: 105-108.
[32]	LI Xiaoya, FENG Jingrong, MENG Yuxian, et al. A Unified MRC Framework for Named Entity Recognition[C]// ACL. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2020: 5849-5859.
[33]	GUO Zhixin, DENG Xiaolong. Intelligent Identification Method of Legal Case Entity Based on BERT-BiLSTM-CRF[J]. Journal of Bejing University of Posts and Telecommunications, 2021, 44(4): 129-134.
	郭知鑫, 邓小龙. 基于BERT-BiLSTM-CRF的法律案件智能识别方法[J]. 北京邮电大学学报, 2021, 44(4): 129-134. doi: 10.13190/j.jbupt.2020-241
[34]	DENG Liang, QI Panhu, LIU Zhenlong, et al. BGPNRE: A BERT-Based Global Pointer Network for Named Entity-Relation Joint Extraction Method[J]. Computer Science, 2023, 50(3): 42-48. doi: 10.11896/jsjkx.220600239
	邓亮, 齐攀虎, 刘振龙, 等. BGPNRE: 一种基于BERT的全局指针网络实体关系联合抽取方法[J]. 计算机科学, 2023, 50(3): 42-48. doi: 10.11896/jsjkx.220600239
[35]	LU Rui, HUANG Junbo, LI Linying. Research on Entity Identification Model of Global Terrorism Research Database Based on BERT-BiLSTM-CRF[J]. Mathematics in Practice and Theory, 2022, 52(8): 128-136.
	卢睿, 黄俊博, 李林瑛. 基于BERT-BiLSTM-CRF的涉恐实体识别模型研究[J]. 数学的实践与认识, 2022, 52(8): 128-136.
[36]	HUANG Junbo. Research on Event Extraction Method in Terrorism News Text[D]. Dalian: Dalian University of Foreign Languages, 2022.
	黄俊博. 涉恐新闻文本的事件抽取方法研究[D]. 大连: 大连外国语大学, 2022.

实体类型	标识	数量/个
涉案时间	SASJ	674
涉案地点	SADD	946
涉案机构	SAJG	190
犯罪嫌疑人	FZXYR	1544
受害人	SHZ	665
作案工具	ZAGJ	190
被盗物品	BDWP	1424
被盗金额	BDJE	216
物品价值	WPJZ	524
盗窃获利	DQHL	90

参数	参数值
操作系统	Microsoft Windows 10
CPU	Intel Core i5-8265U CPU@3.9 GHz
GPU	NVIDIA GeForce RTX3060（12 GB）
内存	32 GB
Python	3.7.15
PyTorch	1.13.0

模型	P	R	F1
W2NER^[29]	91.83%	69.26%	78.98%
BERT+SPAN^[30]	87.75%	90.11%	88.91%
BERT+CRF^[31]	88.68%	89.45%	89.06%
BERT+MRC^[32]	88.13%	87.95%	88.03%
BERT+BiLSTM+CRF^[33]	84.28%	90.08%	87.08%
本文模型	89.39%	88.98%	89.18%

类别	P	R	F1
涉案时间	92.53%	91.84%	92.18%
涉案地点	78.87%	80.87%	79.85%
涉案机构	91.16%	86.84%	88.95%
犯罪嫌疑人	96.60%	97.54%	97.07%
受害人	94.44%	92.03%	93.22%
作案工具	86.59%	81.58%	84.01%
被盗物品	82.94%	80.90%	81.91%
被盗金额	89.27%	84.72%	86.94%
物品价值	96.61%	97.90%	97.25%
盗窃获利	78.64%	90.00%	83.94%

模型	P	R	F1
w/o bilstm+att	86.78%	90.46%	88.58%
w/o att	89.43%	88.50%	88.97%
本文模型	89.39%	88.98%	89.18%