1.College of Information and Management Sciences,Henan Agriculture University,Zhengzhou 450046,China
2.Henan Engineering Research Center for Farmland Environmental Monitoring and Big Data Analysis,Zhengzhou 450002,China
3.College of agronomy,Henan Agriculture University,Zhengzhou 450046,China
Show less
文章历史+
Received
Published
2024-10-30
2025-05-01
Issue Date
2026-04-15
PDF (1826K)
摘要
针对农药登记文本中信息密集、逻辑结构复杂、实体间跨度大以及实体长度异质性等特点,同时为克服传统联合抽取方法中面临的三元组重叠、曝光偏差和冗余计算问题,本研究提出一种多特征融合的单阶段实体关系联合抽取模型 (Multi-feature fusion single-stage entity and relation joint extraction model,MF-SERel)。首先,在编码层,通过融合语义与句法特征,丰富字符向量表示,提升模型对复杂语料的表征能力;其次,在多维标注框架层,提出HT-BES多维标注策略,以解决重叠三元组问题。通过并行评分函数与细粒度分类组件,将实体关系联合抽取转化为了基于关系维度的多标签标注任务,该过程不包含相互依赖步骤,从而实现单阶段并行标注,避免了曝光偏差并降低了计算冗余;最后,在解码层依据细粒度分类预测标签,解码出实体关系三元组。将本研究提出的模型与GraphRel、CasRel和TPLinker等基线模型进行对比,在农药数据集(Pesticide registration dataset,PRD)和公开数据集(Dataset of unstructured information extraction,DuIE)上进行检测。结果表明MF-SERel模型在农药数据集PRD和公开数据集DuIE上具有良好的表现。在农药数据集PRD上,本研究提出的模型MF-SERel在推理速度上提升了20%,F1值提升了2.3%,说明MF-SERel模型在农药登记文本中具有良好的知识挖掘能力;在公开数据集DuIE上,MF-SERel模型在推理速度上提升了54%, F1值提升了1.7%,同样取得了较好结果,证明MF-SERel模型具有较好的泛化能力。综上,本研究提出的MF-SERel模型可为农药领域知识的结构化抽取提供新方法。
Abstract
In response to the characteristics of pesticide registration texts, such as information density, complex logical structure, large span between entities, and heterogeneous entity lengths, as well as to address the issues of triplet overlap, exposure bias, and redundant computation encountered in traditional joint extraction methods, a multi-feature fusion single-stage entity and relation joint extraction model (MF-SERel) is proposed in this study. Firstly, at the encoding layer, the model’s ability to represent complex corpora is enhanced through integrating semantic and syntactic features and enriching the meaning of character symble. Secondly, at the multi-dimensional labeling framework layer, an HT-BES multi-dimensional labeling strategy is proposed to resolve the triplet overlap issue. Through parallel scoring functions and fine-grained classification components, the entity-relation joint extraction is transformed into a multi-label labeling task based on the relation dimension. This process does not involve interdependent steps, thereby enabling single-stage parallel labeling, avoiding exposure bias and reducing computational redundancy. Finally, at the decoding layer, labels are predictied based on fine-grained classification, and the entity-relation triplets are decoded. The proposed model is compared with the baseline models such as GraphRel, CasRel, and TPLinker, and evaluated on the pesticide registration dataset (PRD) and the public dataset of unstructured information extraction (DuIE). The results show that the MF-SERel model performs well on the PRD and DuIE. On the PRD, the proposed MF-SERel model improves inference speed by 20% and F1 score by 2.3%, demonstrating its strong knowledge-mining capability in pesticide registration texts. On the DuIE, the MF-SERel model improves inference speed by 54% and F1 score by 1.7%, also achieves better resultsproves its strong generalization ability. In conclusion, the MF-SERel model proposed in this study provides a new method for the structured extraction of knowledge in the pesticide field.
现有的实体关系抽取研究多集中在通用领域和大规模知识图谱构建,在农药领域仍缺乏知识抽取的相关研究。因此,为应对农药登记文本中信息密集、逻辑结构复杂、实体间跨度大以及实体长度异质性所导致的信息抽取复杂问题,同时为克服联合抽取方法中面临的三元组重叠、曝光偏差和冗余计算问题,本研究拟提出一种多特征融合的单阶段实体关系联合抽取模型MF-SERel,并与现行常用的GraphRel、CasRel以及TPLinker方法进行了对比,在农药登记数据集(Pesticide registration dataset,PRD)和公共数据集(Dataset of unstructured information extraction,DuIE)[32]上进行检测和验证,以期为农药领域知识的结构化整合提供一种更准确、更高效的实体关系联合抽取方法。
式中: v 为得分向量;K为预定义关系的数量; R 为关系向量表示矩阵,可视为可训练的权重;𝜎为激活函数LeakyRelu;drop(·)用于防止过拟合; W 为可训练的权重矩阵;[ eif,ejf ]为字符i和j的向量连接运算;b为偏差项;矩阵 R 与 W 通过自适应学习,可以根据输入的字符特征动态调整评分函数,以便更好地评价字符对于预定义关系的依赖性。
SunX M, LuH Y, BianL P, LiuX J, SunA D. Study on intelligent early warning model of pesticide application risk based on registration information[J]. Chinese Journal of Pesticide Science, 2021, 23(5): 1012-1023 (in Chinese)
[3]
LewisK A, TzilivakisJ, WarnerD J, GreenA. An international database for pesticide risk assessments and management[J]. Human and Ecological Risk Assessment, 2016, 22(4): 1050-1064
WeiH, ZhouA, ZhangY J, ChenF, QuW, LuM Y. Review of deep learning-based biomedical entity relation extraction research[J]. Computer Engineering and Applications, 2021, 57(21): 14-23 (in Chinese)
[6]
ZhangJ C, ZainA M, ZhouK Q, ChenX, ZhangR M. A review of recommender systems based on knowledge graph embedding[J]. Expert Systems with Applications, 2024, 250: 123876
HouC, NiuP Y. Review of research status and prospects of agricultural knowledge graphs[J]. Transactions of the Chinese Society for Agricultural Machinery, 2024, 55(6): 1-17 (in Chinese)
[9]
ZhaoK, XuH, ChengY, LiX T, GaoK. Representation iterative fusion based on heterogeneous graph neural network for joint entity and relation extraction[J]. Knowledge-Based Systems, 2021, 219: 106888
HaoZ G, LiuC, QinL. Entity and relationship joint extraction model of food inspection announcement based on Chinese character adjacency graph[J]. Transactions of the Chinese Society of Agricultural Engineering, 2023, 39(14): 283-292 (in Chinese)
[12]
ChoiH S, SongJ Y, ShinK H, ChangJ H, JangB S. Developing prompts from large language model for extracting clinical information from pathology and ultrasound reports in breast cancer[J]. Radiation Oncology Journal, 2023, 41(3): 209-216
[13]
LiuZ R, LiH Z, WangH, LiaoY L, LiuX G, WuG J. A novel pipelined end-to-end relation extraction framework with entity mentions and contextual semantic representation [J]. Expert Systems with Applications, 2023, 228: 120435
[14]
GaoC, ZhangX, LiuH, YunW, JiangJ H. A joint extraction model of entities and relations based on relation decomposition[J]. International Journal of Machine Learning and Cybernetics, 2022, 13(7): 1833-1845
[15]
AiX G, KavuluruR. End-to-end models for chemical‑protein interaction extraction: Better tokenization and span-based pipeline strategies [C]//2023 IEEE 11th International Conference on Healthcare Informatics (ICHI). ,, Houston:IEEE, 2023: 610-618
[16]
YangC L, DengJ M, ChenX L, AnY. SPBERE: Boosting span-based pipeline biomedical entity and relation extraction via entity information [J]. Journal of Biomedical Informatics, 2023, 145: 104456
[17]
ZhengS C, WangF, BaoH Y, HaoY X, ZhouP, XuB. Joint extraction of entities and relations based on a novel tagging scheme[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Vancouver: USAACL, 2017
[18]
ChangH Y, ZanH Y, GuanT F, ZhangK L, SuiZ F. Application of cascade binary pointer tagging in joint entity and relation extraction of Chinese medical text[J]. Mathematical Biosciences and Engineering, 2022, 19(10): 10656-10672
[19]
YuanY, ZhouX F, PanS R, ZhuQ N, SongZ L, GuoL. A relation-specific attention network for joint entity and relation extraction[C]//Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence. ,, Yokohama: ACM, 2021: 4054-4060
[20]
WeiZ P, SuJ L, WangY, TianY, ChangY. A novel cascade binary tagging framework for relational triple extraction [C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: USAACL, 2020: 1476-1488
GeJ W, LiS L, FangY Q. Joint extraction method of Chinese entity relationship based on the mixture of characters and words[J]. Application Research of Computers, 2021, 38(9): 2619-2623 (in Chinese)
ZhangM Y, ShenH L. Joint extraction of entities and relations based on word-pair distance embedding and axial attention mechanism[J]. Computer Science, 2024, 51(12): 234-241 (in Chinese)
[25]
ZhengH Y, WenR, ChenX, YangY F, ZhangY Y, ZhangZ H, ZhengN Y, QinB, MingX, ZhengY F. PRGC: Potential relation and global correspondence based joint relational triple extraction[C]//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: USAACL, 2021: 6225-6235
[26]
ShangY M, HuangH Y, MaoX L. OneRel: Joint entity and relation extraction with one module in one step[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2022, 36(10): 11285-11293
[27]
ChenZ Q, ZhengY J, GeJ K, YuW C, WangZ N. A parallel model for jointly extracting entities and relations[J]. Neural Processing Letters, 2024, 56(3): 165
[28]
WangY C, YuB W, ZhangY Y, LiuT W, ZhuH S, SunL M. TPLinker: Single-stage joint extraction of entities and relations through token pair linking [C]//Proceedings of the 28th International Conference on Computational Linguistics. Barcelona: International Committee on Computational Linguistics, 2020: 1572-1582
ZhuX B, ZhouG, ChenJ, LuJ C, XiangY X. Single-stage joint entity and relation extraction method based on enhanced sequence annotation strategy[J]. Computer Science, 2023, 50(8): 184-192 (in Chinese)
[31]
FuT J, LiP H, MaW Y. GraphRel: Modeling text as relational graphs for joint entity and relation extraction[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: USAACL, 2019: 1409-1418
LyvD D, ChenJ H, MaoD H, ZhangQ C, ZhaoM, HaoZ H. Entity relationship extraction and correlation analysis of agricultural product standard domain knowledge graph[J]. Transactions of the Chinese Society of Agricultural Engineering, 2022, 38(9): 315-323 (in Chinese)
ZhouJ, ZhengP Y, YuanL C, GeW X, LiangJ. Knowledge graph information extraction for rice fertilization based on improved CASREL[J]. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(11): 314-322 (in Chinese)
[40]
JiaX X, ZhengG, DongC Y, XiS Y, ShiL, XiongS F, MaX M, XiL. Joint extraction of wheat germplasm information entity relationship based on deep character and word fusion[J]. Scientific Reports, 2024, 14(1): 10385
WuS S, ZhouA L, XieN F, LiangX H, WangH J, LiX Y, ChenG P. Construction of visualization domain-specific knowledge graph of crop diseases and pests based on deep learning [J]. Transactions of the Chinese Society of Agricultural Engineering, 2020, 36(24): 177-185 (in Chinese)
JiangL H, ZhaoR X, DongC Y, ChangX Y, MaJ J, XieN F, FangS. Construction and verification of the visual knowledge map of aquatic diseases based on deep learning [J]. Transactions of the Chinese Society of Agricultural Engineering, 2023, 39(15): 259-267 (in Chinese)
[45]
LiS J, HeW, ShiY B, JiangW B, LiangH J, JiangY, ZhangY, LyuY J, ZhuY. Duie: A large-scale Chinese dataset for information extraction [C]//Natural Language Processing and Chinese Computing: 8th CCF International Conference, NLPCC 2019, Dunhuang: Springer International Publishing, 2019: 791-800
[46]
ColicN, RinaldiF. Improving spaCy dependency annotation and PoS tagging web service using independent NER services[J]. Genomics & Informatics, 2019, 17(2): e21
[47]
NickelM, RosascoL, PoggioT. Holographic embeddings of knowledge graphs[C]//Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence. Phoenix: ACM, 2016: 1955-1961