English

应用科学学报 >

2026 , Vol. 44 >Issue 1: 50 - 66

DOI: https://doi.org/10.3969/j.issn.0255-8297.2026.01.004

计算机应用专辑

基于高斯度量学习的不确定性知识图谱推理模型

  • 张钰婷 ,
  • 滕飞 ,
  • 叶晓庆
展开
  • 西南交通大学 计算机与人工智能学院, 四川 成都 611756

收稿日期: 2025-08-11

  网络出版日期: 2026-02-03

基金资助

国家自然科学基金（No.62272398）；四川省科技计划（No.2024NSFJQ0019）

收起

Uncertain Knowledge Graph Reasoning Model Based on Gaussian Metric Learning

  • ZHANG Yuting ,
  • TENG Fei ,
  • YE Xiaoqing
Expand
  • School of Computing and Artificial Intelligence, Southwest Jiaotong University, Chengdu 611756, Sichuan, China

Received date: 2025-08-11

  Online published: 2026-02-03

Fold

摘要

现实知识图谱中普遍存在一些仅包含少量事实的长尾关系，少样本知识图谱补全旨在解决这一数据稀疏问题。然而，现有方法往往忽视了实体与三元组的内在不确定性，限制了模型在噪声干扰或样本匮乏场景下的推理性能。为此，本文提出了一种面向不确定性补全的协方差优化高斯度量学习模型（covariance-optimizedGaussian metric learning for uncertain completion,CoGMUC），用于解决少样本知识图谱的不确定性推理问题。该模型将知识图谱中的实体和关系建模为高斯分布，利用协方差矩阵有效捕捉其内在不确定性，并通过设计协方差感知的多重匹配网络计算语义相似度，实现对缺失事实的补全及置信度预测。此外，引入了困难负样本挖掘策略，进一步增强模型的辨别能力与泛化性能。在公开数据集NL27K和CN15K上的实验结果表明，相较于现有基于高斯度量学习的少样本不确定性知识图谱补全模型，CoGMUC模型在链接预测任务中，平均倒数排名分别提升了21.8%和2.3%，Hits@10分别提升了9.6%和21.5%，在置信度预测任务中均方误差分别降低了14.3%和7.7%。研究结果表明，CoGMUC模型能有效建模并利用不确定性信息，显著提升了少样本知识图谱补全的性能。

关键词： 知识图谱; 不确定性; 知识图谱推理; 高斯度量学习; 协方差建模

本文引用格式

张钰婷 , 滕飞 , 叶晓庆 . 基于高斯度量学习的不确定性知识图谱推理模型[J]. 应用科学学报, 2026 , 44(1) : 50 -66 . DOI: 10.3969/j.issn.0255-8297.2026.01.004

Abstract

Long-tail relations containing only a small number of facts are prevalent in real-world knowledge graphs, and few-shot knowledge graph completion aims to address this data sparsity issue. However, existing approaches often neglect the inherent uncertainty of entities and triples, which limits model reasoning performance under noisy or data-scarce conditions. This paper proposed a covariance-optimized Gaussian metric learning model for uncertain completion (CoGMUC) model to tackle uncertain reasoning in few-shot knowledge graphs. This model represented entities and relations within knowledge graphs as Gaussian distributions, effectively capturing their inherent uncertainty through covariance matrices. It computed semantic similarity with a covariance-aware multi-matching network to complete missing facts and predict confidence levels. Furthermore, a difficult negative sample mining strategy was introduced to enhance the discriminative capability and generalization performance of the model. Experimental results on the public datasets NL27K and CN15K demonstrate that compared with the existing few-shot uncertain knowledge graph completion model based on Gaussian metric learning, CoGMUC improves mean reciprocal rank (MRR) by 21.8% and 2.3% and increases Hits@10 by 9.6% and 21.5%, respectively in the link prediction task. Meanwhile, in the confidence prediction task, the mean squared error (MSE) is reduced by 14.3% and 7.7%, respectively. The findings demonstrate that the CoGMUC model effectively models and leverages uncertainty information, significantly enhancing the performance of few-shot knowledge graph completion.

Key words： knowledge graph; uncertainty; knowledge graph reasoning; Gaussian metric learning; covariance modeling

参考文献

[1] Nickel M, Tresp V, Kriegel H P. A three-way model for collective learning on multirelational data [C]//International Conference on Machine Learning, 2011: 809-816.
[2] Yang B, Yih W, He X, et al. Embedding entities and relations for learning and inference in knowledge bases [DB/OL]. (2014-12-20) [2025-08-11]. https://arxiv.org/abs/1412.6575.
[3] Bordes A, Usunier N, Garcia-Duran A, et al. Translating embeddings for modeling multirelational data [C]//Neural Information Processing Systems, 2013: 26.
[4] Lin Y K, Liu Z Y, Sun M S, et al. Learning entity and relation embeddings for knowledge graph completion [C]//AAAI Conference on Artificial Intelligence, 2015, 29(1): 2181-2187.
[5] Trouillon T, Welbl J, Riedel S, et al. Complex embeddings for simple link prediction [C]//International Conference on Machine Learning, 2016: 2071-2080.
[6] 吴国栋, 刘涵伟, 何章伟, 等. 知识图谱补全技术研究综述[J]. 小型微型计算机系统, 2023, 44(3): 471-482. Wu G D, Liu H W, He Z W, et al. A summary of the research on knowledge map completion technology [J]. Journal of Chinese Computer Systems, 2023, 44(3): 471-482. (in Chinese)
[7] Xiong W, Yu M, Chang S, et al. One-shot relational learning for knowledge graphs [DB/OL]. (2018-08-27) [2025-08-11]. https://arxiv.org/abs/1808.09040.
[8] Zhang C X, Yao H X, Huang C, et al. Few-shot knowledge graph completion [J]. AAAI Conference on Artificial Intelligence, 2020, 34(3): 3041-3048.
[9] Chen X, Chen M, Shi W, et al. Embedding uncertain knowledge graphs [J]. AAAI Conference on Artificial Intelligence, 2019, 33(1): 3363-3370.
[10] Mitchell T, Cohen W, Hruschka E, et al. Never-ending learning [J]. Communications of the ACM, 2018, 61(5): 103-115.
[11] Speer R, Chin J, Havasi C. ConceptNet 5.5: an open multilingual graph of general knowledge [J]. AAAI Conference on Artificial Intelligence, 2017, 31(1): 4444-4451.
[12] Zhang J T, Wu T X, Qi G L. Gaussian metric learning for few-shot uncertain knowledge graph completion [C]//Database Systems for Advanced Applications, 2021: 256-271.
[13] Lao N, Cohen W W. Relational retrieval using a combination of path-constrained random walks [J]. Machine Learning, 2010, 81(1): 53-67.
[14] Tan X Y, Wang X Y, Liu Q, et al. Paths-over-graph: knowledge graph empowered large language model reasoning [C]// ACM on Web Conference 2025, 2025: 3505-3522.
[15] 张天成, 田雪, 孙相会, 等. 知识图谱嵌入技术研究综述[J]. 软件学报, 2023, 34(1): 277-311. Zhang T C, Tian X, Sun X H, et al. Overview on knowledge graph embedding technology research [J]. Journal of Software, 2023, 34(1): 277-311. (in Chinese)
[16] 马恒志, 钱育蓉, 冷洪勇, 等. 知识图谱嵌入研究进展综述[J]. 计算机工程, 2025, 51(2): 18-34. Ma H Z, Qian Y R, Leng H Y, et al. Review of research progress on knowledge graph embedding [J]. Computer Engineering, 2025, 51(2): 18-34. (in Chinese)
[17] Wang Z, Zhang J, Feng J, et al. Knowledge graph embedding by translating on hyperplanes [J]. AAAI Conference on Artificial Intelligence, 2014, 28(1): 1112-1119.
[18] Ji G L, He S Z, Xu L H, et al. Knowledge graph embedding via dynamic mapping matrix [C]//53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, 2015: 687-696.
[19] Socher R, Chen D, Manning C D, et al. Reasoning with neural tensor networks for knowledge base completion [J]. Advances in Neural Information Processing Systems, 2013, 26: 926-934.
[20] Dettmers T, Minervini P, Stenetorp P, et al. Convolutional 2D knowledge graph embeddings [J]. AAAI Conference on Artificial Intelligence, 2018, 32(1): 1811-1818.
[21] Schlichtkrull M, Kipf T N, Bloem P, et al. Modeling relational data with graph convolutional networks [C]//The Semantic Web. Cham: Springer, 2018: 593-607.
[22] Sheng J, Guo S, Chen Z, et al. Adaptive attentional network for few-shot knowledge graph completion [DB/OL]. (2020-10-19) [2025-08-11]. https://arxiv.org/abs/2010.09638.
[23] Chen M, Zhang W, Zhang W, et al. Meta relational learning for few-shot link prediction in knowledge graphs [DB/OL]. (2019-09-04) [2025-08-11]. https://arxiv.org/abs/1909.01515.
[24] Li L Y, Zhang X, Ma Y B, et al. A knowledge graph completion model based on contrastive learning and relation enhancement method [J]. Knowledge-Based Systems, 2022, 256: 109889.
[25] Wu H, Yin J, Rajaratnam B, et al. Hierarchical relational learning for few-shot knowledge graph completion [DB/OL]. (2022-09-02) [2025-08-11]. https://arxiv.org/abs/2209.01205.
[26] Kertkeidkachorn N, Liu X, Ichise R. GTransE: generalizing translation-based model on uncertain knowledge graph embedding [C]//Advances in Artificial Intelligence, 2020: 170-178.
[27] Pai S, Costabello L. Learning embeddings from knowledge graphs with numeric edge attributes [DB/OL]. (2021-05-18) [2025-08-11]. https://arxiv.org/abs/2105.08683.
[28] Chen X, Boratko M, Chen M, et al. Probabilistic box embeddings for uncertain knowledge graph reasoning [DB/OL]. (2021-04-09) [2025-08-11]. https://arxiv.org/abs/2104.04597.
[29] 李健京, 李贯峰, 秦飞舟, 等. 基于不确定知识图谱嵌入的多关系近似推理模型[J]. 计算机应用, 2024, 44(6): 1751-1759. Li J J, Li G F, Qin F Z, et al. Multi-relation approximate reasoning model based on uncertain knowledge graph embedding [J]. Journal of Computer Applications, 2024, 44(6): 1751-1759. (in Chinese)
[30] 赵一晴, 李喜华, 邓彬. 基于语义感知的不确定知识推理模型[J]. 系统工程理论与实践, 2025: 1-12. Zhao Y Q, Li X H, Deng B. Uncertain knowledge reasoning model based on semantic perception [J]. Systems Engineering Theory and Practice, 2025: 1-12. (in Chinese)
[31] Ma R X, Wu H, Wang X R, et al. Multi-view semantic enhancement model for few-shot knowledge graph completion [J]. Expert Systems with Applications, 2024, 238: 122086.
[32] Wang J T, Wu T X, Zhang J T. Incorporating uncertainty of entities and relations into few-shot uncertain knowledge graph embedding [C]//China Conference on Knowledge Graph and Semantic Computing, 2022: 16-28.
[33] Chen Z M, Yeh M Y, Kuo T W. PASSLEAF: a pool-based semi-supervised learning framework for uncertain knowledge graph embedding [J]. AAAI Conference on Artificial Intelligence, 2021, 35(5): 4019-4026.
[34] Tseng Y C, Chen Z M, Yeh M Y, et al. UPGAT: uncertainty-aware pseudo-neighbor augmented knowledge graph attention network [C]//Advances in Knowledge Discovery and Data Mining, 2023: 53-65.
Options
文章导航

/

〈 〉