面向工艺文本中的命名实体,该文提出一种融入领域知识的神经网络命名实体识别方法,旨在对零件、工程图纸、参考标准、属性等12类命名实体进行识别。该方法针对工艺实体的特点,利用领域词典及规则预识别出部分实体,形成预识别实体特征,将预识别实体特征加入CNN-BiLSTM-CRF神经网络模型,指导训练与预测。实验结果表明,该方法在工艺文本中能较好地完成命名实体识别任务,在提高词典及规则覆盖的实体识别效果的同时,还能够提高其他类实体的识别效果,通过加入预识别实体特征,使得F1值从90.99%提升到93.03%,验证了该文方法的有效性。
Abstract
This paper proposes, a method of identifying named entities based on neural network with domain knowledge to identify 12 types of process entities including parts, engineering drawings, reference standards and attributes. According to the characteristics of process entities, this method uses domain dictionaries and rules to pre-identify candidate entities to form pre-recognition features, which are then fed to the CNN-BiLSTM-CRF neural network model. The experimental results show that, by adding pre-recognition entity features, the F1 value is increased from 90.99% to 93.03%.
关键词
工艺文本 /
命名实体识别 /
领域词典及规则 /
CNN-BiLSTM-CRF
{{custom_keyword}} /
Key words
process text /
named entity recognition /
domain dictionaries and rules /
CNN-BiLSTM-CRF
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] Chinchor N, Robinson P. MUC-7 named entity task definition[C]//Proceedings of the 7th Conference on Message Understanding, 1997, 29: 1-21.
[2] 任民山, 蔡红霞, 李钧. 基于本体的民机装配工艺知识库系统构建[J]. 工业控制计算机, 2018, 31(4): 122-123.
[3] 田桂中, 朱钰萍, 刘金锋, 等. 面向机加工艺智能生成的工艺知识建模与管理方法[J]. 计算机集成制造系统, 2019, 25(07): 1695-1705.
[4] 赵宇, 单忠臣, 顾金凤, 等. 基于特征的回转体零件工艺文件自动生成方法研究[J]. 应用科技, 2004, 31(8): 5-7.
[5] 刘持平, 董成. 非回转体CAPP系统结构与工艺过程自动生成研究[J]. 新技术新工艺, 2011(09): 81-83.
[6] 石春丹, 秦岭. 基于BGRU-CRF的中文命名实体识别方法[J]. 计算机科学, 2019, 46(09): 237-242.
[7] 张聪品, 方滔, 刘昱良. 基于LSTM-CRF命名实体识别技术的研究与应用[J]. 计算机技术与发展, 2019, 29(02): 112-114.
[8] Sahu S K, Anand A . Recurrent neural network models for disease name recognition using domain invariant features[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. 2016: 2216-2225.
[9] Hammerton J. Named entity recognition with long short-term memory[C]//Proceedings of the 7th Conference on Natural Language Learning at HLT-NAACL. Association for Computational Linguistics, 2003: 172-175.
[10] Huang Z, Xu W, Yu K. Bidirectional LSTM-CRF models for sequence tagging[J]. arXiv preprint arXiv: 1508.01991, 2015: 1-10.
[11] Chiu J P C, Nichols E. Named entity recognition with bidirectional LSTM-CNNs[J]. Transactions of the Association for Computational Linguistics, 2016: 357-370.
[12] Lample G, Ballesteros M, Subramanian S, et al. Neural architectures for named entity recognition[C]//Proceedings of the NAACL-HLT, 2016: 260-270.
[13] 殷章志, 李欣子, 黄德根, 等. 融合字词模型的中文命名实体识别研究[J]. 中文信息学报, 2019, 33(11): 95-102.
[14] Zhang, J Yang, Chinese NER using lattice LSTM[C]//Proceedings of ACL, 2018: 1554-1564.
[15] Seyler D, Dembelova T, Corro L D, et al. A study of the importance of external knowledge in the named entity recognition task[C]//Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, 2018: 241-246.
[16] Song C H, Lawrie D, Finin T, et al. Improving neural named entity recognition with gazetteers[C]//Proceedings of the 33th International Florida Artificial Intelligence Research Society Conference, 2020.
[17] 马建红, 王立芹, 姚爽. 面向化学资源文本的命名实体识别[J]. 郑州大学学报 (理学版), 2018, 50(4): 14-20.
[18] 冯蕴天, 张宏军, 郝文宁. 面向军事文本的命名实体识别[J]. 计算机科学, 2015, 42(7): 15-18.
[19] 朱顺乐. 基于深度学习的维吾尔语命名实体识别模型[J]. 计算机工程与设计, 2019 (10): 2874-2878.
[20] 代培, 杨爱民. 一种基于逆向最大匹配的中文分词算法: CN. CN102999534 A[P],2013.
[21] Mikolov, Tomas, Sutskever, et al. Distributed Representations of words and phrases and their compositionality[J].Advances in neural information processing systems, 2013, 26: 3111-3119.
[22] Mikolov T, Corrado G, Chen K, et al. Efficient estimation of word representations in vector space[C]//Proceedings of the International Conference on Learning Representations, 2013.
[23] Lai S . Word and document embeddings based on neural network approaches[J]. arXiv preprint arXiv: 1611.05962,2016.
[24] Santos C N D, Zadrozny B . Learning character-level representations for part-of-speech tagging[C]//Proceedings of the International Conference on Machine Learning, 2014.
[25] Srivastava N, Hinton G, Krizhevsky A, et al. Dropout: a simple way to prevent neural networks from overfitting[J].The Journal of Machine Learning Research, 2014, 15(1): 1929-1958.
[26] Chris Dyer, Miguel Ballesteros, Wang Ling,et al. Transition-based dependency parsing with stack long short-term memory[C]//Proceedings of the ACL, 2015: 334-343.
[27] John Lafferty, Andrew McCallum, Fernando C N Pereira. Conditional random fields: probabilistic models for segmenting and labeling sequence data[C]//Proceedings of the 18th International Conference on Machine Learning, 2001.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
基金
辽宁省重点研发计划(2019JH2/10100020);辽宁省自然科学基金(20170540705);沈阳市重大科技创新研发计划(Y19-1-011)
{{custom_fund}}
PDF(2128 KB)
