Zero-shot knowledge graph relation prediction method via dual-channel embedding fusion

doi:10.11772/j.issn.1001-9081.2025080945

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (10): 3117-3123.DOI: 10.11772/j.issn.1001-9081.2025080945

• Artificial intelligence • Previous Articles

Zero-shot knowledge graph relation prediction method via dual-channel embedding fusion

Bing GAO¹, Jinzhuo LIN¹, Qijie ZOU¹(), Yongqiang YAO²

^1.College of Information Engineering，Dalian University，Dalian Liaoning 116622，China
^2.Affiliated Zhongshan Hospital of Dalian University，Dalian Liaoning 116001，China

Received:2025-08-19 Revised:2025-10-07 Accepted:2025-10-16 Online:2025-11-18 Published:2026-10-10
Contact: Qijie ZOU
About author:GAO Bing， born in 1976， Ph. D.， associate professor. His research interests include data mining， big data analysis.
LIN Jinzhuo， born in 2001， M. S. candidate. His research interests include knowledge graph.
YAO Yongqiang， born in 1976， Ph. D.， professor. His research interests include clinical breast and thyroid surgery.
Supported by:
Scientific Research Project of Liaoning Province(LJKZ1180);Interdisciplinary General Project of Dalian University(DLUXK-2024-YB-002);Dalian University Discipline Development Special Project(DLUXK-2025-FX-001)

基于双通道嵌入融合的零样本知识图谱关系预测方法

高兵¹, 林锦卓¹, 邹启杰¹(), 姚永强²

^1.大连大学信息工程学院，辽宁大连 116622
^2.大连大学附属中山医院，辽宁大连 116001

通讯作者: 邹启杰
作者简介:高兵（1976—），男，黑龙江哈尔滨人，副教授，博士，CCF会员，主要研究方向：数据挖掘、大数据分析
林锦卓（2001—），男，河北邢台人，硕士研究生，主要研究方向：知识图谱
姚永强（1976—），男，辽宁大连人，教授，博士，主要研究方向：乳腺和甲状腺外科临床。
基金资助:
辽宁省科学研究经费资助项目(LJKZ1180);大连大学学科交叉一般类项目(DLUXK-2024-YB-002);大连大学学科建设专项(DLUXK-2025-FX-001)

Abstract

Abstract:

Zero-shot knowledge graph relation prediction aims to infer triplets with unseen relations using the existing fact triplets. The existing methods often underutilize entity descriptions， only considering text information and graph structure information of relations. Text information for relations only relies on dataset-provided text descriptions， so it is difficult to capture the detailed semantics of relations and the overall semantic information of knowledge graph triplets. To address this issue， a zero-shot knowledge graph relation prediction method via dual-channel embedding fusion was proposed. First， a structural encoder and a semantic encoder were used to generate structural and semantic features of triplet entities. Then， the obtained features were fused via a dual-channel entity multimodal information fusion module， and the text generation ability of Large Language Models （LLMs） was used to generate description information after enhancing relation text semantics. Third， through training， semantically enhanced relation features and multimodal entity features were aligned semantically in vector space. Finally， for zero-shot relations， zero-shot knowledge graph relation prediction was realized according to semantically enhanced text description information. Experimental results on two real datasets show that compared to the optimal baseline method， on the NELL_ZS dataset， the proposed method improves Mean Reciprocal Rank （MRR）， Hit@1， Hit@5， and Hit@10 by 8.7， 7.2， 9.5， and 11.2 percentage points， respectively； on the WN18RR_ZSL dataset， the proposed method improves MRR， Hit@1， Hit@5， and Hit@10 by 6.6， 9.2， 3.9， and 0.8 percentage points， respectively， demonstrating the effectiveness of the proposed method.

Key words: Knowledge Graph Completion (KGC), Zero-Shot Knowledge Graph Completion (ZS-KGC), relation prediction, multi-modal fusion, Generative Adversarial Network (GAN)

摘要：

零样本知识图谱关系预测任务旨在根据知识图谱已有的事实三元组推理预测包含未见关系的事实三元组。已有方法大多仅考虑关系的文本信息和图结构信息，而未充分利用实体的描述信息；针对关系的文本信息仅依赖于数据集本身的文本描述，难以捕捉关系的详细语义与知识图谱三元组整体的语义信息。为解决上述问题，本文提出一种基于双通道嵌入融合的零样本知识图谱关系预测方法。首先，使用结构编码器和语义编码器分别生成三元组实体部分的结构特征与语义特征；其次，通过双通道实体多模态信息融合模块融合得到的特征，同时利用大语言模型（LLM）的文本生成能力生成关系文本语义增强后的描述信息；再次，经过训练，语义增强后的语义特征与实体的多模态特征在向量空间中满足语义一致性；最后，针对零样本关系，根据语义增强后的文本描述信息完成零样本知识图谱关系预测。在两个真实数据集上的实验结果表明，相较于最优基线模型，在NELL_ZS数据集上，所提方法的平均倒数排名（MRR）、Hit@1、Hit@5和Hit@10分别提高了8.7、7.2、9.5和11.2个百分点；在WN18RR_ZSL数据集上，所提方法的MRR、Hit@1、Hit@5和Hit@10分别提高了6.6、9.2、3.9和0.8个百分点，验证了所提方法的有效性。

关键词: 知识图谱补全, 零样本知识图谱补全, 关系预测, 多模态融合, 生成对抗网络

CLC Number:

TP391.1

Bing GAO, Jinzhuo LIN, Qijie ZOU, Yongqiang YAO. Zero-shot knowledge graph relation prediction method via dual-channel embedding fusion[J]. Journal of Computer Applications, 2026, 46(10): 3117-3123.

高兵, 林锦卓, 邹启杰, 姚永强. 基于双通道嵌入融合的零样本知识图谱关系预测方法[J]. 《计算机应用》唯一官方网站, 2026, 46(10): 3117-3123.

Figures/Tables 9

References 20

[1]	Bollacker K， Evans C， Paritosh P， et al. Freebase： a collaboratively created graph database for structuring human knowledge［C］// SIGMOD 2008. New York： ACM， 2008： 1247-1250.
[2]	Suchanek F M， Kasneci G， Weikum G. YAGO： a core of semantic knowledge［C］// WWW 2007. New York： ACM， 2007： 697-706.
[3]	Mitchell T， Cohen W， Hruschka E， et al. Never-ending learning［J］. Communications of the ACM， 2018， 61（5）： 103-115.
[4]	Vrandečić D， Krötzsch M. Wikidata： a free collaborative knowledgebase［J］. Communications of the ACM， 2014， 57（10）： 78-85.
[5]	Fdez-Díaz M， Quevedo J R， Montañés E. Target inductive methods for zero-shot regression［J］. Information Sciences， 2022， 599： 44-63.
[6]	Hao Y， Cao X， Fang Y， et al. Inductive link prediction for nodes having only attribute information［C］// IJCAI 2020. California： IJCAI， 2020： 1209-1215.
[7]	郑修林，周鹏，李培培，等. 知识图谱补全技术及应用［J］. 软件学报， 2025， 36（12）： 5599-5628.
	Zheng Xiulin， Zhou Peng， Li Peipei， et al. Knowledge graph completion： techniques and applications［J］. Journal of Software， 2025， 36（12）： 5599-5628.
[8]	Qin P， Wang X， Chen W， et al. Generative adversarial zero-shot relational learning for knowledge graphs［J］. Proceedings of the AAAI Conference on Artificial Intelligence， 2020， 34（5）： 8673-8680.
[9]	Goodfellow I J， Pouget-Abadie J， Mirza M， et al. Generative adversarial nets［C］// NeurIPS 2014. Cambridge： MIT Press， 2014： 2672-2680.
[10]	Geng Y， Chen J， Chen Z， et al. OntoZSL： ontology-enhanced zero-shot learning［C］// WWW 2021. New York： ACM， 2021： 3325-3336.
[11]	Yang Y， Hospedales T M. A unified perspective on multi-domain and multi-task learning［PP/OL］. V2. arXiv （2015-03-26）［2025-12-23］..
[12]	Ba J L， Swersky K， Fidler S. Predicting deep zero-shot convolutional neural networks using textual descriptions［C］// ICCV 2015. Piscataway： IEEE， 2015： 4247-4255.
[13]	Geng Y， Chen J， Zhang W， et al. Disentangled ontology embedding for zero-shot learning［C］// KDD 2022. New York： ACM， 2022： 443-453.
[14]	Schlichtkrull M， Kipf T N， Bloem P， et al. Modeling relational data with graph convolutional networks［C］// ESWC 2018， LNCS 10843. Cham： Springer， 2018： 593-607.
[15]	Wei Y， Huang Q， Kwok J T， et al. KICGPT： large language model with knowledge in context for knowledge graph completion［C］// Findings of the Association for Computational Linguistics： EMNLP 2023. Stroudsburg： ACL， 2023： 8667-8683.
[16]	Bordes A， Usunier N， Garcia-Durán A， et al. Translating embeddings for modeling multi-relational data［C］// NeurIPS 2013. Cambridge： MIT Press， 2013： 2787-2795.
[17]	Yang B， Yih W T， He X， et al. Embedding entities and relations for learning and inference in knowledge bases［PP/OL］. V4. arXiv （2015-08-29）［2025-12-20］..
[18]	Trouillon T， Welbl J， Riedel S， et al. Complex embeddings for simple link prediction［J］. Proceedings of Machine Learning Research， 2016， 48： 2071-2080.
[19]	Wang B， Shen T， Long G， et al. Structure-augmented text representation learning for efficient knowledge graph completion［C］// WWW 2021. New York： ACM， 2021： 1737-1748.
[20]	Hinton G E， Roweis S. Stochastic neighbor embedding［C］// NeurIPS 2002. Cambridge： MIT Press， 2002： 857-864.

数据集	实体数	三元组数	关系数
NELL_ZS	65 567	188 392	139/10/32
WN18RR_ZSL	40 943	87 667	6/2/3

数据集	实体数	三元组数	关系数
NELL_ZS	65 567	188 392	139/10/32
WN18RR_ZSL	40 943	87 667	6/2/3

方法	NELL_ZS				WN18RR_ZSL
方法	MRR	Hit@10/%	Hit@5/%	Hit@1/%	MRR	Hit@10/%	Hit@5/%	Hit@1/%
ZS_TransE	0.097	20.3	14.7	4.3	0.043	10.2	6.1	0.8
ZS_DistMult	0.235	32.6	28.4	18.5	0.109	17.6	10.1	5.3
ZS_ComplEx	0.216	31.6	26.7	16.0	0.098	14.0	8.6	3.3
StAR	0.192	37.2	26.6	10.9	0.271	47.9	37.4	16.7
ZSGAN	0.253	37.6	31.6	19.4	0.149	28.3	20.6	8.4
Onto-ZSL	0.256	38.5	31.8	18.8	—	—	—	—
本文方法（TransE）	0.343	49.4	41.1	26.6	0.337	48.7	41.3	25.9
本文方法（DistMult）	0.340	49.7	41.3	25.8	0.303	47.4	38.9	21.6

方法	NELL_ZS				WN18RR_ZSL
方法	MRR	Hit@10/%	Hit@5/%	Hit@1/%	MRR	Hit@10/%	Hit@5/%	Hit@1/%
ZS_TransE	0.097	20.3	14.7	4.3	0.043	10.2	6.1	0.8
ZS_DistMult	0.235	32.6	28.4	18.5	0.109	17.6	10.1	5.3
ZS_ComplEx	0.216	31.6	26.7	16.0	0.098	14.0	8.6	3.3
StAR	0.192	37.2	26.6	10.9	0.271	47.9	37.4	16.7
ZSGAN	0.253	37.6	31.6	19.4	0.149	28.3	20.6	8.4
Onto-ZSL	0.256	38.5	31.8	18.8	—	—	—	—
本文方法（TransE）	0.343	49.4	41.1	26.6	0.337	48.7	41.3	25.9
本文方法（DistMult）	0.340	49.7	41.3	25.8	0.303	47.4	38.9	21.6

数据集	Fusion	LLM	MRR	Hit@10/%	Hit@5/%	Hit@1/%
NELL_ZS	√		0.325	47.9	40.2	24.5
		√	0.301	45.2	37.2	22.5
	√	√	0.343	49.4	41.1	26.6
WN18RR_ZSL	√		0.254	37.0	31.6	19.0
		√	0.256	42.1	33.9	16.7
	√	√	0.337	48.7	41.3	25.9

Zero-shot knowledge graph relation prediction method via dual-channel embedding fusion

基于双通道嵌入融合的零样本知识图谱关系预测方法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 9

References 20

Related Articles 15

Recommended Articles

Metrics

数据集	模型	MRR	Hit@10/%	Hit@5/%	Hit@1/%
NELL_ZS	Qwen-MAX	0.343	49.4	41.1	26.6
NELL_ZS	Qwen3-30B	0.320	46.7	39.1	24.4
WN18RR_ZSL	Qwen-MAX	0.337	48.7	41.3	25.9
WN18RR_ZSL	Qwen3-30B	0.293	46.9	37.3	20.9

[1]	Xuan CUI, Bo LIU. Unsupervised face attribute editing method based on dynamic convolutional autoencoder [J]. Journal of Computer Applications, 2026, 46(4): 1300-1308.
[2]	Haoyang ZHANG, Liping ZHANG, Sheng YAN, Na LI, Xuefei ZHANG. Review of large language model methods for knowledge graph completion [J]. Journal of Computer Applications, 2026, 46(3): 683-695.
[3]	Min CHEN, Xiaolin QIN, Shaohan LI, Hao YANG, Taohong LI. Review of deep learning applications in severe convective weather prediction [J]. Journal of Computer Applications, 2026, 46(3): 980-992.
[4]	Fei WANG, Ye TAO, Jiawang LIU, Wei LI, Xiugong QIN, Ning ZHANG. Bimodal fusion method for constructing spatio-temporal knowledge graph in smart home space [J]. Journal of Computer Applications, 2026, 46(1): 52-59.
[5]	Na FAN, Chuang LUO, Zehui ZHANG, Mengyao ZHANG, Ding MU. Semantic privacy protection mechanism of vehicle trajectory based on improved generative adversarial network [J]. Journal of Computer Applications, 2026, 46(1): 169-180.
[6]	Lifang WANG, Wenjing REN, Xiaodong GUO, Rongguo ZHANG, Lihua HU. Trident generative adversarial network for low-dose CT image denoising [J]. Journal of Computer Applications, 2026, 46(1): 270-279.
[7]	Yinlong JIAN, Xuebin CHEN, Zhongrui JING, Qi ZHONG, Zhenbo ZHANG. Data augmentation scheme based on conditional generative adversarial network in federated learning [J]. Journal of Computer Applications, 2026, 46(1): 21-32.
[8]	Yanan LI, Mengyang GUO, Guojun DENG, Yunfeng CHEN, Jianji REN, Yongliang YUAN. Method for life prediction of parallel branching engine based on multi-modal fusion features [J]. Journal of Computer Applications, 2026, 46(1): 305-313.
[9]	Yilin DENG, Fajiang YU. Pseudo random number generator based on LSTM and separable self-attention mechanism [J]. Journal of Computer Applications, 2025, 45(9): 2893-2901.
[10]	Jin ZHOU, Yuzhi LI, Xu ZHANG, Shuo GAO, Li ZHANG, Jiachuan SHENG. Modulation recognition network for complex electromagnetic environments [J]. Journal of Computer Applications, 2025, 45(8): 2672-2682.
[11]	Ying HUANG, Shengmei GAO, Guang CHEN, Su LIU. Low-light image enhancement network combining signal-to-noise ratio guided dual-branch structure and histogram equalization [J]. Journal of Computer Applications, 2025, 45(6): 1971-1979.
[12]	Hui LI, Bingzhi JIA, Chenxi WANG, Ziyu DONG, Jilong LI, Zhaoman ZHONG, Yanyan CHEN. Generative adversarial network underwater image enhancement model based on Swin Transformer [J]. Journal of Computer Applications, 2025, 45(5): 1439-1446.
[13]	Sheping ZHAI, Qing YANG, Yan HUANG, Rui YANG. Knowledge graph completion using hierarchical attention fusing directed relationships and relational paths [J]. Journal of Computer Applications, 2025, 45(4): 1148-1156.
[14]	Lihu PAN, Shouxin PENG, Rui ZHANG, Zhiyang XUE, Xuzhen MAO. Video anomaly detection for moving foreground regions [J]. Journal of Computer Applications, 2025, 45(4): 1300-1309.
[15]	Hong SHANGGUAN, Huiying REN, Xiong ZHANG, Xinglong HAN, Zhiguo GUI, Yanling WANG. Low-dose CT denoising model based on dual encoder-decoder generative adversarial network [J]. Journal of Computer Applications, 2025, 45(2): 624-632.