[1]刘 纳,季 喆,吴克东,等.基于三重提示和对比学习的中文医疗命名实体识别[J].郑州大学学报(工学版),2026,47(5):77-84.[doi:10.13705/j.issn.1671-6833.2026.05.002]
　LIU Na,JI Zhe,WU Kedong,et al.Chinese Medical Named Entity Recognition Based on Triple Hint and Contrast Learning[J].Journal of Zhengzhou University (Engineering Science),2026,47(5):77-84.[doi:10.13705/j.issn.1671-6833.2026.05.002]
点击复制

基于三重提示和对比学习的中文医疗命名实体识别()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
47
期数:
2026年5期
页码:
77-84
栏目:
出版日期:
2026-09-09

文章信息/Info

Title:
Chinese Medical Named Entity Recognition Based on Triple Hint and Contrast Learning
文章编号:
1671-6833(2026)05-0077-08
作者:
刘 纳1,2, 季 喆1,2, 吴克东1,2, 刘 磊1,2
1. 北方民族大学 计算机科学与工程学院,宁夏 银川 750021;2. 北方民族大学 图形图像智能处理国家民委重点实验室,宁夏 银川 750021
Author(s):
LIU Na1,2, JI Zhe1,2, WU Kedong1,2, LIU Lei1,2
1. School of Computer Science and Engineering, North Minzu University, Yinchuan 750021, China;2. State Key Laboratory of Intelligent Processing of Graphics and Images, North Minzu University, Yinchuan 750021, China
关键词:
小样本;  命名实体识别;  提示学习;  特征交互;  对比学习
Keywords:
small sample;  named entity recognition;  cue learning;  feature interaction;  contrast learning
分类号:
TP391
DOI:
10.13705/j.issn.1671-6833.2026.05.002
文献标志码:
A
摘要:
针对中文医疗小样本命名实体识别任务中预训练知识表征不足、边界模糊及类别混淆等问题，提出一种基于三重提示引导的多阶段识别方法。首先，构建三重提示学习机制，通过边界定位、类型判别与语境验证的分步推理范式，结合离散模板的可解释性与连续向量的可优化性，增强模型对实体结构的解析能力；其次，设计双向门控网络，动态调控提示信息与文本表征的特征交互，强化复杂术语的语义捕获效果；最后，引入对比学习优化特征空间分布，提升实体类内聚合度与类间区分度。实验结果表明：该模型在 IMCS‑V2‑NER、cMedQANER 和 CCKS2019 3个医疗命名实体识别数据集上 F1 值分别达到91.86%、84.06%和88.93%。证明该方法能够稳定提升实体识别性能，并在低资源场景下表现出较好的泛化能力。
Abstract:
To address the issues of insufficient pretrained knowledge representation, ambiguous boundaries, and category confusion in Chinese medical few‑shot named entity recognition, a multi‑stage recognition approach guided by triple prompts was proposed in this study. Firstly, a triple‑prompt learning mechanism was constructed, in which a stepwise reasoning paradigm, comprising boundary localization, type discrimination, and contextual verification was adopted. By integrating the interpretability of discrete templates with the optimizability of continuous vectors, the capability of the model to parse entity structures was enhanced. Secondly, a bidirectional gating network was designed to dynamically regulate feature interactions between prompt information and textual representations, thereby strengthening semantic capture for complex terminology. Finally, contrastive learning was introduced to optimize the distribution of the feature space, improving intra‑class compactness and inter‑class separability. Experimental results indicated that F1 scores of 91.86%, 84.06%, and 88.93% were achieved on the IMCS‑V2‑NER, cMedQANER, and CCKS2019 datasets, respectively. These findings demonstrated that the proposed method consistently improved entity recognition performance and exhibited strong generalization capability under low‑resource settings.

参考文献/References:

[1] Kang Hui, Xiao Jingwu, Zhang Yunpeng, et al. A research toward Chinese named entity recognition based on transfer learning[J]. International Journal of Computational Intelligence Systems, 2023, 16(1):56.
[2] Liu Xin, Xu Hongzhen, Liu Aihua, et al. Geological named entity recognition based on MacBERT and R‑Drop[J]. Journal of Zhengzhou University (Engineering Science), 2024, 45(3):89‑95.
[3] Lin Lingde, Liu Na, Xu Zhenshun, et al. Chinese medical named entity recognition based on multi‑layer dynamic fusion[J]. Computer Engineering and Applications, 2024, 60(15):161‑169.
[4] Yang Haotian, Wang Li, Yang Yanpeng, et al. Named entity recognition in electronic medical records incorporating pre‑trained and multi‑head attention[J]. IAENG International Journal of Computer Science, 2024, 51(4):401‑408.
[5] Zhang Ziqi, Zheng Xiangwei, Zhang Jinsong. Machine reading comprehension based named entity recognition for medical text[J]. Multimedia Tools and Applications, 2025, 84(28):33431‑33451.
[6] Cui Jinman, Li Dongmei, Tian Xuan, et al. Survey on prompt learning[J]. Computer Engineering and Applications, 2024, 60(23):1‑27.
[7] Brown T B, Mann B, Ryder N, et al. Language models are few‑shot learners[C]//Proceedings of the 34th International Conference on Neural Information Processing Systems. New York: ACM, 2020:1877‑1901.
[8] Zheng Guofeng, Liu Na, Li Chen, et al. Chinese medical named entity recognition based on prompt tuning and contrastive learning[J]. Computer Engineering and Applications, 2026, 62(1):231‑242.
[9] Schick T, Schütze H. Exploiting cloze‑questions for few‑shot text classification and natural language inference[C]//Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume. Stroudsburg: ACL, 2021:255‑269.
[10] Cui Leyang, Wu Yu, Liu Jian, et al. Template‑based named entity recognition using BART[C]//Proceedings of the Findings of the Association for Computational Linguistics: ACL‑IJCNLP 2021. Stroudsburg: ACL, 2021:1835‑1845.
[11] Petroni F, Rocktäschel T, Riedel S, et al. Language models as knowledge bases?[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP‑IJCNLP). Stroudsburg: ACL, 2019:2463‑2473.
[12] Jiang Cong, Xu Qingyang, Song Yong, et al. Discrete sequence rearrangement based self‑supervised Chinese named entity recognition for robot instruction parsing[J]. Intelligence & Robotics, 2023, 3(3):337‑354.
[13] Mai Chengcheng, Chen Yu, Gong Ziyu, et al. Prompt‑CNER: a segmentation‑based method for few‑shot Chinese NER with prompt‑tuning[J]. ACM Transactions on Asian and Low‑Resource Language Information Processing, 2025, 24(9):1‑24.
[14] Liu Jiang, Fei Hao, Li Fei, et al. TKDP: threefold knowledge‑enriched deep prompt tuning for few‑shot named entity recognition[J]. IEEE Transactions on Knowledge and Data Engineering, 2024, 36(11):6397‑6409.
[15] Chen Yuxiang, Dan Zhiping, Gao Zhun, et al. Chinese medical continual named entity recognition based on continual learning and knowledge distillation[C]//Proceedings of the 2024 IEEE International Conference on High Performance Computing and Communications (HPCC). Piscataway: IEEE, 2024:885‑893.
[16] Chen Junqi, Ding Zhaoyun, Jin Guang, et al. The optimal sampling strategy for few‑shot named entity recognition based on prototypical network[C]//Proceedings of the 2024 International Joint Conference on Neural Networks (IJCNN). Piscataway: IEEE, 2024:1‑8.
[17] Wang Chengyu, Zhao Shan, Yan Tianwei, et al. Hierarchical label‑enhanced contrastive learning for Chinese NER[J]. IEEE Transactions on Neural Networks and Learning Systems, 2025, 36(6):10504‑10514.
[18] Qin Xiao, Tan Sijing, Chun Xin, et al. Simplify prompt template and optimum label‑select for few‑shot NER[C]//Applied intelligence. Singapore: Springer Nature Singapore, 2025:370‑381.
[19] Bhushan R C, Donthi R K, Chilukuri Y, et al. Biomedical named entity recognition using improved green ana‑conda‑assisted Bi‑GRU‑based hierarchical ResNet model[J]. BMC Bioinformatics, 2025, 26:34.
[20] Huang Yucheng, He Kai, Wang Yige, et al. COPNER: contrastive learning with prompt guiding for few‑shot named entity recognition[C]//Proceedings of the 29th International Conference on Computational Linguistics. Stroudsburg: ACL, 2025‑2527.
[21] Zhu Wei, Wang Xiaoling, Zheng Huanran, et al. PromptCBLUE: a Chinese prompt tuning benchmark for the medical domain[PP/OL]. (2023‑10‑22)[2025‑12‑24]. https://doi.org/10.48550/arXiv.2310.14151.
[22] Liu Feifei, Liu Mingtong, Li Meiting, et al. Automatic knowledge extraction from Chinese electronic medical records and rheumatoid arthritis knowledge graph construction[J]. Quantitative Imaging in Medicine and Surgery, 2023, 13(6):3873‑3890.
[23] Wang Haitao, He Zhengqiu, Zhu Tong, et al. CCKS2019 shared task on inter‑personal relationship extraction[PP/OL]. (2019‑08‑29)[2025‑12‑24]. https://doi.org/10.48550/arXiv.1908.11337.
[24] Xu Liang, Tong Yu, Dong Qianqian, et al. CLUENER2020: fine‑grained named entity recognition dataset and benchmark for Chinese[PP/OL]. (2020‑01‑20)[2025‑12‑24]. https://doi.org/10.48550/arXiv.2001.04351.
[25] Yang Yi, Katiyar A. Simple and effective few‑shot named entity recognition with structured nearest neighbor learning[C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). Stroudsburg: ACL, 2020:6365‑6375.
[26] Ma Jie, Ballesteros M, Doss S, et al. Label semantics for few shot named entity recognition[C]//Proceedings of the Findings of the Association for Computational Linguistics. Stroudsburg: ACL, 2022:1956‑1971.
[27] Shen Yongliang, Tan Zeqi, Wu Shuhui, et al. Prompt‑NER: prompt locating and typing for named entity recognition[C]//Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Stroudsburg: ACL, 2023:12492‑12507.
[28] Chen Xiang, Li Lei, Deng Shumin, et al. LightNER: a lightweight tuning paradigm for low‑resource NER via pluggable prompting[C]//Proceedings of the 29th International Conference on Computational Linguistics. Stroudsburg: ACL, 2022:2374‑2387.
[29] Chen Yanru, Zheng Yanan, Yang Zhilin. Prompt‑based metric learning for few‑shot NER[C]//Proceedings of the Findings of the Association for Computational Linguistics: ACL 2023. Stroudsburg: ACL, 2023:7199‑7212.

更新日期/Last Update: 2026-09-07