基于跨度和语义特征的实体关系抽取模型

doi:10.13229/j.cnki.jdxbgxb.20240474

吉林大学学报(工学版) ›› 2025, Vol. 55 ›› Issue (12): 4045-4051.doi: 10.13229/j.cnki.jdxbgxb.20240474

• 计算机科学与技术 • 上一篇

基于跨度和语义特征的实体关系抽取模型

冯萍^1,²(),杨茈茜¹,王韧杰¹,冯师语¹,吴航¹,孙宇³

^1.长春大学计算机科学技术学院，长春 130022
^2.长春大学残障人士智能康复与无障碍教育部重点实验室，长春 130022
^3.长春大学特殊教育学院，长春 130022

收稿日期:2024-04-30 出版日期:2025-12-01 发布日期:2026-02-03
作者简介:冯萍（1977-），女，副教授，博士研究生.研究方向：人工智能、知识图谱、康复工程.E-mail：fengping@ccu.edu.cn
基金资助:
吉林省科技厅科技发展计划项目(2023JB405L07);国家自然科学基金面上项目(62377006);吉林省发改委计划项目(2019C048-6)

Entity relationship extraction method based on span and semantic features

Ping FENG^1,²(),Zi-qian YANG¹,Ren-jie WANG¹,Shi-yu FENG¹,Hang WU¹,Yu SUN³

^1.College of Computer Science and Technology，Changchun University，Changchun 130022，China
^2.Ministry of Education Key Laboratory of Intelligent Rehabilitation and Barrier-free Access for the Disabled，Changchun University，Changchun 130022，China
^3.College of Special Education，Changchun University，Changchun 130022，China

Received:2024-04-30 Online:2025-12-01 Published:2026-02-03

摘要/Abstract

摘要：

针对传统实体关系抽取方法依赖距离度量或简单跨度识别，难以捕捉实体间潜在关系，且模型计算复杂度高、易引发误差传递的问题，提出了一种基于跨度和语义特征的端到端实体关系抽取模型。首先，将文本向量随机分割成跨度序列，以便模型能够学习更广泛的语义特征信息。其次，对语义关系进行判定，以筛选出候选关系子集，从而减少信息冗余。最后，将候选关系转换为包含重要关系语义的关系-跨度组合，并利用Transformer解码器实现实体关系联合抽取。实验结果表明，相较于其他基线模型，该模型在NYT数据集和WebNLG数据集上的F₁值均有显著提升，证明了其有效性。

关键词: 实体关系抽取, 跨度, 语义特征, 上下文信息

Abstract:

A end-to-end entity relationship extraction model based on span and semantic features was proposed to address the problems of traditional entity relationship extraction methods relying on distance measurement or simple span recognition， which make it difficult to capture potential relationships between entities， and the high computational complexity and error propagation of the model. Firstly， the text vectors were randomly segmented into span sequences so that the model can learn a wider range of semantic feature information. Secondly， semantic relationships were judged to screen out subsets of candidate relationships， thus reducing information redundancy. Finally， the candidate relationships were transformed into relationship-span combinations containing important relationship semantics， and the Transformer decoder was used to achieve the joint extraction of entity relationships. The experimental results show that the F₁ value of this model is significantly improved in the NYT and WebNLG datasets compared to other baseline models， proving its effectiveness.

Key words: entity relationship extraction, span, semantic feature, context information

中图分类号:

TP391

冯萍,杨茈茜,王韧杰,冯师语,吴航,孙宇. 基于跨度和语义特征的实体关系抽取模型[J]. 吉林大学学报(工学版), 2025, 55(12): 4045-4051.

Ping FENG,Zi-qian YANG,Ren-jie WANG,Shi-yu FENG,Hang WU,Yu SUN. Entity relationship extraction method based on span and semantic features[J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(12): 4045-4051.

图/表 8

图1

图2

图3

表1

表2

图4

表3

图5

参考文献 19

[1]	李冬梅, 张扬, 李东远, 等. 实体关系抽取方法研究综述[J]. 计算机研究与发展, 2020, 57(7): 1424-1448.
	Li Dong-mei, Zhang Yang, Li Dong-yuan, et al. Review of studies on entity relationship extraction methods[J]. Computer Research and Development, 2020,57(7): 1424-1448.
[2]	Yang Y, Wu Z L, Yang Y X, et al.A Survey of information extraction based on deep learning[J].Applied Sciences, 2022, 12(19): No.9691.
[3]	Li J, Sun A X, Han J L, et al. A survey on deep learning for named entity recognition[J]. IEEE Transactionson Knowledge and Data Engineering, 2020: 50-70.
[4]	王进, 蒋诗琪. 基于多路局部特征整合的嵌套命名实体识别方法[J]. 江苏大学学报:自然科学版, 2025, 46(4): 431-437.
	Wang Jin, Jiang Shi-qi. Nested named entity recognition method based on multiplexed local feature integration[J]. Journal of Jiangsu University (Natural Science Edition), 2025, 46(4): 431-437.
[5]	王进, 王猛旗, 张昕跃, 等. 基于多头注意力机制字词联合的中文命名实体识别[J]. 江苏大学学报:自然科学版, 2024, 45(1): 77-84.
	Wang Jin, Wang Meng-qi, Zhang Xin-yue, et al. Chinese named entity recognition based on multi-head attention character word integration[J]. Journal of Jiangsu University (Natural Science Edition), 2024, 45(1): 77-84.
[6]	Tuo M, Yang W Z, Wei F Y, et al.A novel chinese overlapping entity relation extraction model using word-label based on cascade binary tagging[J]. Electronics, 2023, 12(4): No.1013.
[7]	Wang P. A survey of research on deep learning entity relationship extraction[J]. Natural Language Processing and Speech Recognition, 2019, 1(1): 1-5.
[8]	Li J L, Xu Y J, Lin H Z, et al.Semantic-consistent learning for one-shot joint entity and relation extraction[J]. Applied Intelligence, 2022,53(5): 5963-5976.
[9]	Lin Y, Shen S, Liu Z, et al. Neural relation extraction with selective attention over instances[C]∥Proceedings of the 54the Annual Meeting of the Association for Computational Linguistics. Berlin: Association for Computational Linguistics, 2016: 2124-2133.
[10]	Kalpit D, Yaser A.Span-Level Model for Relation Extraction[C]∥Proceedings of 57th Annual Meeting of the Association for Computational Linguistics (ACL 2019), Florence, Italy, 2019: 5308-5314.
[11]	Eberts M, Ulges A.Span-based joint entity and relation extraction with transformer pre-Training[C]∥ Proceedings of the 24th European Conference on Artificial Intelligence. Amsterdam: IOS Press, 2020: 2006-2013.
[12]	Wan Q, Wei L, Zhao S, et al.A span-based multi-modal attention network for joint entity-relation extraction[J].Knowledge-Based Systems, 2023, 262:No.110228.
[13]	Devlin J, Chang M W, Lee K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[EB/OL]. [2024-03-15].
[14]	Ma L, Ren H, Zhang X.Effective cascade dual-decoder model for joint entity and relation extraction[EB/OL]. [2024-03-17].
[15]	Surdeanu M, Tibshirani J, Nallapati R, Manning C D. Multi-instance multi-label learning for relation extraction[C]∥Proceedings of the Joint Conference on Empirical Methods in Natural Language Processing and Computationl Natural Language Learning. Je ju Island: Association for Computational Linguistics, 2012: 455-465.
[16]	Riedel S, Yao L, McCallum A.Modeling relations and their mentions without labeled text[C]∥Proceedings of the European Conference on Machine Learning and Knowledge Discovery in Databases (ECML PKDD). Berlin: Springer, 2010: 148-163.
[17]	Zeng X R, Zeng D J, He S Z,et al.Extracting relational facts by an end-to-end neural model with copy mechanism[C]∥Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics,Melbourne, Australia, 2018: 506-514.
[18]	Ren F L, Zhang L H, Yin S J,et al.A novel global feature-oriented relational triple extraction model based on table filling[C]∥Proceedings of the Conference on Empirical Methods in Natural Language Processing, Online and Punta Cana, Dominican Republic, 2021: 2646-2656.
[19]	Xie X Y, Xie M, Moshayedi A J, et al. A hybrid improved neural networks algorithm based on L2 and dropout regularization[J].Mathematical Problems in Engineering, 2022: 1-19.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

参数	值
GPU/（24 G）	4 090
内存/G	32
batch size	16
学习率	0.000 1
max span	6
epoch	100
BERT_emb	768

模型	NYT					WebNLG
模型	P	R		F₁		P	R	F₁
CopyRE	61.0		56.6		58.7	37.7	36.4	37.1
GraphRel	63.9		60.0		61.9	44.7	41.1	42.9
PRGC	89.6		82.3		85.8	90.6	88.5	89.5
GRTE	92.9		93.1		93.0	93.7	94.2	93.9
Ours	93.1		93.8		93.4	94.0	94.1	94.0

数据集	模式	F₁/%
NYT	语义筛选关系子集	93.4
NYT	常规	90.3
WebNLG	语义筛选关系子集	94.0
WebNLG	常规	88.7

基于跨度和语义特征的实体关系抽取模型

Entity relationship extraction method based on span and semantic features

RICH HTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 19

相关文章 8

Metrics

本文评价

推荐阅读 0

[1]	才华,王玉瑶,付强,马智勇,王伟刚,张晨洁. 基于注意力机制和特征融合的语义分割网络[J]. 吉林大学学报(工学版), 2025, 55(4): 1384-1395.
[2]	郭晓然,王铁君,闫悦. 基于局部注意力和本地远程监督的实体关系抽取方法[J]. 吉林大学学报(工学版), 2025, 55(1): 307-315.
[3]	张磊,焦晶,李勃昕,周延杰. 融合机器学习和深度学习的大容量半结构化数据抽取算法[J]. 吉林大学学报(工学版), 2024, 54(9): 2631-2637.
[4]	王春华,李恩泽,肖敏. 多特征融合和孪生注意力网络的高分辨率遥感图像目标检测[J]. 吉林大学学报(工学版), 2024, 54(1): 240-250.
[5]	冯宇,郝键铭,王峰,张久鹏,黄晓明. 非平稳极端风作用下大跨桥梁瞬态风致效应分析[J]. 吉林大学学报(工学版), 2023, 53(6): 1638-1649.
[6]	王雪,李占山,吕颖达. 基于多尺度感知和语义适配的医学图像分割算法[J]. 吉林大学学报(工学版), 2022, 52(3): 640-647.
[7]	郭殊伦,钟铁毅,闫志刚. 大跨度斜拉桥拉索的抖振响应计算方法[J]. 吉林大学学报(工学版), 2021, 51(5): 1756-1762.
[8]	蔡健，陈国栋，左志亮，吴轶 . 受压弦杆填充混凝土的带悬挑预应力矩形钢管桁架[J]. 吉林大学学报(工学版), 2009, 39(02): 393-0397.