基于BERT-BiLSTM-CRF的法律案件实体智能识别方法

doi:10.13190/j.jbupt.2020-241

北京邮电大学学报 ›› 2021, Vol. 44 ›› Issue (4): 129-134.doi: 10.13190/j.jbupt.2020-241

基于BERT-BiLSTM-CRF的法律案件实体智能识别方法

郭知鑫, 邓小龙

北京邮电大学网络空间安全学院, 北京 100876

收稿日期:2020-11-30 发布日期:2021-07-13
通讯作者: 邓小龙(1977-),男,副教授,E-mail:shannondeng@bupt.edu.cn. E-mail:shannondeng@bupt.edu.cn
作者简介:郭知鑫(1994-),男,硕士生.
基金资助:
国家重点研发项目子课题（2017YFC0820603）

Intelligent Identification Method of Legal Case Entity Based on BERT-BiLSTM-CRF

GUO Zhi-xin, DENG Xiao-long

School of Cyberspace Security, Beijing University of Posts and Telecommunications, Beijing 100876, China

Received:2020-11-30 Published:2021-07-13

摘要/Abstract

摘要： 在智能法务系统应用中，人工智能自然语言处理相关技术常采用静态特征向量模型，算法效率低，精度偏差较大.为了对法律文本中的案件实体进行智能识别，提高案件的处理效率，针对动态字向量模型提出以基于转换器的双向编码表征模型作为输入层的识别方法.在其基础上通过融合双向长短期记忆网络和条件随机场模型，构建了高精度的法律案件实体智能识别方法，并通过实验验证了模型的性能.

关键词: 自然语言处理, 智能法务, 基于转换器的双向编码表征模型

Abstract: In the past, artificial intelligence natural language processing related technologies often used static feature vector models in the application of intelligent legal systems, which had problems such as low algorithm efficiency and large accuracy deviations. To intelligently identify case entities in legal texts and improve case processing efficiency, the dynamic word vector model is studied, and a recognition method based on the bidirectional encoder representations from transformers model as the input layer is proposed. Based on the fusion of bi-directional long short-term memory and conditional random fields models, a high-precision method of intelligent identification of legal case entities is constructed. The performance of the model is verifiedthrough experiments.

Key words: natural language processing, intelligent legal affairs, bidirectional encoder representations from transformers model

中图分类号:

TP391.1

郭知鑫, 邓小龙. 基于BERT-BiLSTM-CRF的法律案件实体智能识别方法[J]. 北京邮电大学学报, 2021, 44(4): 129-134.

GUO Zhi-xin, DENG Xiao-long. Intelligent Identification Method of Legal Case Entity Based on BERT-BiLSTM-CRF[J]. Journal of Beijing University of Posts and Telecommunications, 2021, 44(4): 129-134.

参考文献

[1] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems. MountainView:[s.n.], 2017:5998-6008.
[2] Grishman R, Sundheim B. Message understanding conference-6:a brief history[C]//Proceedings of the 16th Conference on Computational Linguistics. Copenhagen:Association for Computational Linguistics, 1996:466-471.
[3] Bikel D M, Schwartz R, Weischedel R M. An algorithm that learns what's in a name[J]. Machine Learning, 1999, 34(1/2/3):211-231.
[4] Sekine S, Grishman R, Shinnou H. A decision tree method for finding and classifying names in Japanese texts[J]. Proceeding of the 6th Workshop on Very Large Corpora, 1998(5):171-178.
[5] Borthwick A. A maximum entropy approach to named entity recognition[J]. Thesis New York University, 1999, 36(1):4701-4708.
[6] McCallum A, Li Wei. Early results for named entity recognition with conditional random fields, feature induction and web-enhanced lexicons[C]//Proceedings of the Seventh Conference on Natural Language Learning at HLT-NAACL 2003. Edmonton:Association for Computational Linguistics, 2003:188-191.
[7] Zhang Yue, Yang Jie. Chinese NER using lattice LSTM[C]//Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1:Long Papers). Melbourne:Association for Computational Linguistics, 2018:1554-1564.
[8] Jawahar G, Sagot B, Seddah D. What does BERT learn about the structure of language?[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Florence:Association for Computational Linguistics, 2019:3651-3657.
[9] Peinelt N, Nguyen D, Liakata M. tBERT:topic models and BERT joining forces for semantic similarity detection[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg:Association for Computational Linguistics, 2020:7047-7055.
[10] Chai Duo, Wu Wei, Han Qinghong, et al. Description based text classification with reinforcement learning[C]//PMLR. Austria Vienna:[s.n.], 2020:1371-1382.
[11] Qu Chen, Yang Liu, Qiu Minghui, et al. BERT with history answer embedding for conversational question answering[C]//Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. Paris:ACM, 2019:1133-1136.
[12] Dai Zhuyun, Callan J. Deeper text understanding for IR with contextual neural language modeling[C]//Procee-dings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. Paris:ACM, 2019:985-988.
[13] 赵平, 孙连英, 万莹, 等. 基于BERT+BiLSTM+CRF的中文景点命名实体识别[J]. 计算机系统应用, 2020, 29(6):169-174. Zhao Ping, Sun Lianying, Wan Ying, et al. Chinese scenic spot named entity recognition based on BERT+BiLSTM+CRF[J]. Computer Systems & Applications, 2020, 29(6):169-174.
[14] Hochreiter S, Schmidhuber J. Long short-term memory[J]. Neural Computation, 1997, 9(8):1735-1780.
[15] 张玉帅, 赵欢, 李博. 基于BERT和BiLSTM的语义槽填充[J]. 计算机科学, 2021, 48(1):247-252. Zhang Yushuai, Zhao Huan, Li Bo. Semantic slot filling based on BERT and BiLSTM[J]. Computer Science, 2021, 48(1):247-252.
[16] 李丽双, 郭元凯. 基于CNN-BLSTM-CRF模型的生物医学命名实体识别[J]. 中文信息学报, 2018, 32(1):116-122. Li Lishuang, Guo Yuankai. Biomedical named entity recognition with CNN-BLSTM-CRF[J]. Journal of Chinese Information Processing, 2018, 32(1):116-122.

基于BERT-BiLSTM-CRF的法律案件实体智能识别方法

Intelligent Identification Method of Legal Case Entity Based on BERT-BiLSTM-CRF

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 7

编辑推荐

Metrics

本文评价

[1]	付洁琼, 孙亚伟, 刘建毅, 李金斌. 一种基于对抗训练的医疗问答匹配方法[J]. 北京邮电大学学报, 2022, 45(4): 37-43.
[2]	黄诗琳郑小林陈德人. 针对产品命名实体识别的半监督学习方法[J]. 北京邮电大学学报, 2013, 36(2): 20-23.
[3]	李荣军, 王小捷, 周延泉. PageRank模型在中文情感词极性判别中的应用[J]. 北京邮电大学学报, 2010, 33(5): 141-144.
[4]	王大亮1, 张德政1, 涂序彦1, 郑雪峰1, 佟子健2. 基于相对条件熵的搭配抽取方法[J]. 北京邮电大学学报, 2007, 30(6): 40-45.
[5]	刘耀　穗志方胡永伟冀铁亮. 领域Ontology自动构建研究[J]. 北京邮电大学学报, 2006, 29(s2): 65-69.
[6]	李蕾, 许晶, 龚媛, 周延泉, 王枞, 钟义信. 基于全信息自然语言理解的语音识别后文本处理[J]. 北京邮电大学学报, 2006, 29(s2): 192-196.
[7]	李蕾,孙春葵,杨晓兰,钟义信. 一种特定领域中文自动摘要系统[J]. 北京邮电大学学报, 2000, 23(1): 6-10.