Dual-Channel Multi-Granularity Feature Fusion for Named Entity Recognition in Ancient Poetry A Case Study of the Tang and Song Dynasties

Abstract

Abstract:

In response to the scarcity of training data in the field of ancient Chinese poetry we constructthe POEM-NER dataset for ancient poetry focusing on collecting descriptive poetry data related to naturalscenery and utilizing the BIOES method for entity annotation. To overcome the limitations of traditionalnamed entity recognition NER methods in handling the complex sentence structures of ancient poetryand the extraction of relatively singular features we propose the DMFF-APNER method based on dualchannelmulti-granularity feature fusion for ancient poetry NER. First we pre-train the SikuBERTmodel on ancient poetry corpora obtain multi-granularity feature vectors using Word2Vec and use FLATfor embedding fusion. Next the information vectors are input into the BiLSTM and IDCNN dualchannelsfor parallel feature extraction. The extracted contextual and local features are then dynamicallyfused through an attention mechanism. Finally predicted sequence labels are obtained through CRFdecoding. Comparative and ablation experiments on the open-source C-CLUE dataset and the POEM-NERdataset demonstrate that DMFF-APNER model can effectively utilize multi-granularity features to enhancethe model’s semantic representation capability. Furthermore the use of dual-channel technology in the feature extraction layer complements the two types of features thereby significantly improving entityextraction performance. The F1 values on the C-CLUE dataset and the POEM-NER ancient poetry datasetreach 82. 66% and 86. 02% respectively.

Key words:

Named entity recognition, Multi-granularity feature fusion, Dual-channel, Ancient Poetry, Digital humanities

Zhao Jingsheng1 2 Yang Xinyi1 Qu Weilong1 Zheng Jiashang1Zhu Qiaoming2（1 School of Information and Control Engineering Qingdao University of Technology 2 School of Computer Science and Technology Soochow University）. Dual-Channel Multi-Granularity Feature Fusion for Named Entity Recognition in Ancient Poetry A Case Study of the Tang and Song Dynasties[J]. Libraly Journal, 2025, 44(415): 75-87.

References

[ 1 ] Grishman R. Message understanding conference-6a brief history C / / Proceedings of the 16thconference on computational linguistics 1996.

[ 2 ] Nadeau D Sekine S. A survey of named entityrecognition and classification J . LingvisticaeInvestigationes 2007 30 1 3 26.

[ 3 ] Wang W Chang L Bin C et al. ESN-NERentity storage network using attention mechanism forChinese NER C / / The International Conference2019.
[ 4 ] Zhu P Cheng D Yang F et al. ImprovingChinese named entity recognition by large-scalesyntactic dependency graph J . IEEE / ACMTransactions on Audio Speech and LanguageProcessing 2022 30 979 991.

[ 5 ] 韩立帆,季紫荆,陈子睿,等. 数字人文视域下面向历史古籍的信息抽取方法研究[J]. 大数据, 2022, 8(6):26 39.

[ 6 ] 张玥杰,徐智婷,薛向阳. 融合多特征的最大熵汉语命名实体识别模型[J]. 计算机研究与发展, 2008(6):1004 1010.

[ 7 ] 何炎祥,罗楚威,胡彬尧. 基于CRF 和规则相结合的地理命名实体识别方法[J]. 计算机应用与软件,2015, 32(1):179 185; 202.

[ 8 ] 乐娟,赵玺. 基于HMM 的京剧机构命名实体识别算法[J]. 计算机工程,2013, 39 (6):266 271;286. [ 9 ] 米健霞,谢红薇. 面向招标物料的命名实体识别研究及应用[J]. 计算机工程与应用,2023, 59(2): 314 320.

[10] 包振山,宋秉彦,张文博,等. 基于半监督学习和规则相结合的中医古籍命名实体识别研究[J].中文信息学报,2022, 36(6):90 100.

[11] 张晓艳,王挺,陈火旺. 基于混合统计模型的汉语命名实体识别方法[J]. 计算机工程与科学, 2006(6):135 139.

[12] Lample G Ballesteros M Subramanian S et al.Neural architectures for named entity recognitionC / / Proceedings of the 2016 Conference of theNorth American Chapter of the Association forComputational Linguistics Human LanguageTechnologies. 2016 260 270.

[13] Strubell E Verga P Belanger D et al. Fast andaccurate entity recognition with iterated dilatedconvolutions C / / Proceedings of the 2017Conference on Empirical Methods in NaturalLanguage Processing. 2017 2670 2680.

[14] 林立涛,王东波,刘江峰,等. 数字人文视域下典籍动物命名实体识别研究———以SikuBERT 预训练模型为例[J]. 图书馆论坛,2022, 42 (10 ):4250.

[15] 陈雪松,詹子依,王浩畅. 融合SikuBERT 模型与MHA 的古汉语命名实体识别[J]. 吉林大学学报(信息科学版),2023, 41(5):866 875.

[16] Zhang Y Yang J. Chinese NER using latticeLSTM C / / Proceedings of the 56th AnnualMeeting of the Association for ComputationalLinguistics Volume 1 Long Papers . 2018 15541564.
[17] Li X Yan H Qiu X et al. FLAT Chinese NERusing flat-lattice transformer C / / Proceedings ofthe 58th Annual Meeting of the Association forComputational Linguistics. Association forComputational Linguistics 2020 611 617.

[18] 崔丹丹,刘秀磊,陈若愚,等. 基于Lattice LSTM的古汉语命名实体识别[J]. 计算机科学,2020,47(增刊2):18 22.

[19] 张朦,刘忠宝. 数字人文环境下融入多特征的词命名实体识别[J]. 计算机系统应用,2023, 32(3):
[20] 武帅,杨秀璋,何琳,等. 基于句法特征和Bert-BiLSTM-MHA-CRF 的细粒度古籍实体识别研究[J].数据分析与知识发现,2024,8(12):136 148.
[21] 崔竞烽,郑德俊,王东波,等. 基于深度学习模型的菊花古典诗词命名实体识别[J]. 情报理论与实践,2020, 43(11):150 155.

[22] 余馨玲,常娥. 基于DA-BERT-CRF 模型的古诗词地名自动识别研究———以金陵古诗词为例[J].图书馆杂志,2023, 42(10):87 94; 73.

[23] 徐润华,陈小荷. 一种利用注疏的《左传》分词新方法[J]. 中文信息学报,2012, 26(2):13 17;45.
[24] 黄建年. 农业古籍的计算机断句标点与分词标引研究[D]. 南京:南京农业大学,2009.

[25] Fu X Yuan T Li X et al. Research on themethod and system of word segmentation and POStagging for ancient Chinese medicine literatureC / / 2019 IEEE International Conference onBioinformatics and Biomedicine BIBM . IEEE2020. [26] 李娜,包平. 面向数字人文的馆藏方志古籍地名自动识别模型构建[ J]. 图书馆,2018 ( 5 ): 67 73.
[27] 王菁薇,肖莉,骆嘉伟,等. 基于《伤寒论》的命名实体识别研究[J]. 计算机与数字工程, 2021,49(8):1584 1587.

[28] 俞士汶,段慧明,朱学锋,等. 北京大学现代汉语语料库基本加工规范[J]. 中文信息学报, 2002(5):49 64.
[29] 黄水清,王东波. 新时代人民日报分词语料库构建、性能及应用(一)———语料库构建及测评[J].图书情报工作,2019, 63(22):5 12.

[30] 蔡竹娟,赵丹群. 引文分析视角下的中国古诗词用典问题初探[J]. 图书情报工作,2022, 66(20): 82 92.

[31 ] 王东波, 刘畅, 朱子赫, 等. SikuBERT 与SikuRoBERTa:面向数字人文的《四库全书》预训练模型构建及应用研究[J]. 图书馆论坛,2022,42(6):31 43.

[1]	Pan Jun, Xiao Mingcheng, Tao Xiangxing（School of Science Zhejiang University of Science and Technology）. Machine Reading Comprehension and Lexicon-Enhanced BERT Based Named Entity Recognition Ancient Chinese Science and Technology Texts [J]. Libraly Journal, 2025, 44(415): 64-74.
[2]	Liu Maolin, Zhao Meng , Wang Hao( School of Information Management, NanjingUniversity, Jiangsu Key Laboratory of Data Engineering and Knowledge Service). Research on the Construction Method and Distribution Law of Object-Image Database for Ancient Poetry [J]. Libraly Journal, 2024, 43(393): 96-108.
[3]	Yu Xinling, Chang E ( School of Economics and Management, Southeast University, Southeast University Library). Automatic Recognition of Place Names in Ancient PoetryBased on DA-BERT-CRF Models: Taking the AncientPoetries of Nanjing as an Example [J]. Libraly Journal, 2023, 42(390): 87-94.