Evolution-Aware Temporal Knowledge Graph Reasoning Method Based on Message Aggregation

ZHU Ke; QIN Tao; ZHU Kuiyu; WANG Chenxu

doi:10.7652/xjtuxb202511018

您当前的位置：

首页 >

文章列表页 >

Evolution-Aware Temporal Knowledge Graph Reasoning Method Based on Message Aggregation

更新时间：2025-11-11

- Evolution-Aware Temporal Knowledge Graph Reasoning Method Based on Message Aggregation
- Journal of Xi'an Jiaotong University Vol. 59, Issue 11, Pages: 187-197(2025)
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：10.7652/xjtuxb202511018
  CLC： TP391
- Received：26 March 2025，
  
  Published：10 November 2025
- 稿件说明：
移动端阅览
ZHU Ke, QIN Tao, ZHU Kuiyu, et al. Evolution-Aware Temporal Knowledge Graph Reasoning Method Based on Message Aggregation[J]. Journal of Xi'an Jiaotong University, 2025, 59(11): 187-197.
DOI：

ZHU Ke, QIN Tao, ZHU Kuiyu, et al. Evolution-Aware Temporal Knowledge Graph Reasoning Method Based on Message Aggregation[J]. Journal of Xi'an Jiaotong University, 2025, 59(11): 187-197. DOI： 10.7652/xjtuxb202511018.

摘要

针对现有时序知识图谱推理补全方法侧重图结构特征建模，忽略其他重要演化依赖信息的问题，提出一

种基于消息聚合的演化感知时序知识图谱补全方法，从局部和全局两个角度建模演化历史。基于邻近时间步内事实的演化具有依赖性的假设，使用演化聚合单元提取邻近时间步内图谱的语义、结构和时序特征。基于全局历史具有重现性的假设，提取全局历史中查询事实相关的交互实体或关系，构建候选实体或关系编码矩阵。解码时采用时间增强的ConvTransE解码器聚合局部依赖性和全局重复性演化信息，实现多粒度融合。训练时引入一种由易到难的课程学习策略，充分捕获历史演化序列的长度多样性。实验结果表明，该方法在3个标准数据集上均优于最优基线方法：在ICEWS14数据集上，平均倒数排名（

）、命中率Hits@3和Hits@10（正确答案出现在预测排名前3、10位的比例）较时序归纳路径神经网络（TiPNN）分别提升了1.13%、1.19%和2.45%；在ICEWS18数据集上，

较基于图卷积网络的循环演化网络（RE-GCN）提升0.98%，Hits@3和Hits@10较TiPNN分别提升0.73%和1.59%；在WIKI数据集上，

较RE-GCN提升0.13%，Hits@3和Hits@10较TiPNN分别提升2.97%和2.89%。在多个数据集上的性能提升验证了所提方法在全面建模演化依赖方面的有效性。

Abstract

To address the issue that existing temporal knowledge graph reasoning and completion methods primarily focus on modeling graph structural features while neglecting other crucial evolutionary dependencies

this study proposes an evolution-aware temporal knowledge graph completion method based on message aggregation

which models evolutionary history from both local and global perspectives. Based on the assumption that the evolution of facts within adjacent timestamps exhibits dependencies

an evolution aggregation unit is employed to extract semantic

structural

and temporal features of the graph within neighboring timestamps. Furthermore

under the hypothesis that global history exhibits recurrence

interactive entities or relations relevant to the query fact are extracted from the global history to construct a candidate entity or relation encoding matrix. During decoding

a time-enhanced ConvTransE decoder is adopted to aggregate local dependency and global recurrence-based evolutionary information

achieving multi-granularity fusion. A curriculum learning strategy

progressing from easy to difficult

is introduced during training to fully capture the length diversity of historical evolutionary sequences. Experimental results demonstrate that the proposed method outperforms state-of-theart baseline

s on three standard datasets：On the ICEWS14 dataset

the mean reciprocal rank (

)

Hits@3

and Hits@10 (the proportion of correct answers appearing in the top 3 and 10 predicted rankings) improve by 1.13%

1.19%

and 2.45%

respectively

compared to TiPNN. On the ICEWS18 dataset

increases by 0.98% over RE-GCN

while Hits@3 and Hits@10 improve by 0.73% and 1.59%

respectively

compared to TiPNN. On the WIKI dataset

rises by 0.13% over RE-GCN

and Hits@3 and Hits@10 increase by 2.97% and 2.89%

respectively

compared to TiPNN. The performance improvement on multiple datasets verifies the effectiveness of the proposed method in comprehensively modeling evolutionary dependencies.

关键词

Keywords

references

SHI Baoxu , WENINGER T . Open-world knowledge graph completion [C ] // Proceedings of the 32nd AAAI Conference on Artificial Intelligence and 30th Innovative Applications of Artificial Intelligence Conference and 8th AAAI Symposium on Educational Advances in Artificial Intelligence . Palo Alto,CA,USA : AAAI Press , 2018 : 1957 - 1964 .

DONG Xin , GABRILOVICH E , HEITZ G , et al . Knowledge vault:a web-scale approach to probabilis tic knowledge fusion [C ] // Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining . New York,USA : ACM , 2014 : 601 - 610 .

BORDES A , USUNIER N , GARCIA-DURÁN A , et al . Translating embeddings for modeling multi-relational data [C ] // Proceedings of the 27th International Conference on Neural Information Processing Systems . Vancouver,Canada : Neural Information Processing Systems Foundation , 2013 : 2787 - 2795 .

NICKEL M , TRESP V , KRIEGEL H . Athree-way model for collective learning on multi-relational data [C ] // Proceedings of the 28th International Conference on Machine Learning . Madison,WI,USA : Omnipress , 2011 : 809 - 816 .

YANG B , YIH S , HE X , et al . Embedding entities and relations for learning and inference in knowledge bases [C ] // Proceedings of the 3rd International Conference on Learning Representations(ICLR 2015) . San Diego,CA,USA : ICLR , 2015 : 20193407343615 .

TROUILLON T , WELBL J , RIEDEL S , et al . Complex embeddings for simple link prediction [C ] // Proceedings of the 3rd International Conference on Machine Learning . New York,USA : JMLR , 2016 : 2071 - 2080 .

BOSCHEE E , LAUTENSCHLAGER J , O′BRIEN S , et al . ICEWS coded event data [EB/OL ] .( 2025-01-01 )[ 2025-01-10 ] . http://doi.org/10.7910/DVN/28075 .

LEETARU K , SCHRODT P . GDELT:global data on events,location,and tone,1979—2012 [C ] // ISA Annual Convention . Los Angeles,CA,USA : ISA , 2013 : 1 - 49 .

LEBLAY J , CHEKOL M W . Deriving validity time in knowledge graph [C ] // Companion Proceedings of the Web Conference 2018.Republic and Canton of Geneva,CHE:International World Wide Web Conferences Steering Committee , 2018 : 1771 - 1776 .

GARCÍA-DURÁN A , DUMANCIC′ S , NIEPERT M . Learning sequence encoders for temporal knowledge graph completion [C ] // Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing . Stroudsburg,PA,USA : ACL , 2018 : 4816 - 4821 .

LACROIX T , OBOZINSKI G , USUNIER N . Tensor decompositions for temporal knowledge base completion [C ] // Proceedings of the 8th International Conference on Learning Representations . San Diego,CA,USA : ICLR , 2020 : 20231313796168 .

JIN W , QU Meng , JIN Xisen , et al . Recurrent event network:autoregressive structure inference over temporal knowledge graphs [C ] // Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP ) . Stroudsburg, PA, USA : ACL , 2020 : 6669 - 6683 .

LI Zixuan , JIN Xiaolong , LI Wei , et al . Temporal knowledge graph reasoning based on evolutional representation learning [C ] // Proceedings of the 44th internationalACM SIGIRconference on research and development in information retrieval . New York,USA : ACM , 2021 : 408 - 417 .

LIU Xinyue , ZHANG Jianan , MA Chi , et al . Temporal knowledge graph reasoning with dynamic hypergraph embedding [C ] // Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024) . Stroudsburg,PA,USA : ACL , 2024 : 15742 - 15751 .

DONG Hao , WANG Pengyang , XIAO Meng , et al . Temporal inductive path neural network for temporal knowledge graph reasoning [J ] . Artificial Intelligence , 2024 , 329 : 104085 .

ZHU Cunchao , CHEN Muhao , FAN Changjun , et al . Learning from history:modeling temporal knowledge graphs with sequential copy-generation networks [C ] // Proceedings of the AAAI Conference on Artificial Intelligence . Palo Alto,CA,USA : AAAI Press , 2021 : 4732 - 4740 .

XU Yi , OU Junjie , XU Hui , et al . Temporal knowledge graph reasoning with historical contrastive learning [C ] // Proceedings of the AAAI Conference on Artificial Intelligence . Palo Alto, CA, USA : AAAI Press , 2023 : 4765 - 4773 .

LI Yujia , SUN Shiliang , ZHAO Jing . TiRGN:timeguided recurrent graph network with local-global historical patterns for temporal knowledge graph reasoning [C ] // Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence . San Mateo, CA,USA : IJCAI , 2022 : 2152 - 2158 .

HE Yongquan , ZHANG Peng , LIU Luchen , et al . HIP network:historical information passing network for extrapolation reasoning on temporal knowledge graph [C ] // Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence . San Mateo, CA,USA : IJCAI , 2021 : 1915 - 1921 .

LV Ao , HUANG Yongzhong , OUYANG Guige , et al . RLGNet: repeating-local-global history network for temporal knowledge graph reasoning [EB/OL ] . ( 2024-03-31 )[ 2025-1-10 ] . https://arxiv.org/abs/2404.00586v1 .

XU Wenjie , LIU Ben , PENG Miao , et al . Pre-trained language model with prompts for temporal knowledge graph completion [C ] // Findings of the Association for Computational Linguistics:ACL 2023 . Stroudsburg, PA,USA : ACL , 2023 : 7790 - 7803 .

LEE D H , AHRABIAN K , JIN W , et al . Temporal knowledge graph forecasting without knowledge using in-context learning [C ] // Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing . Stroudsburg,PA,USA : ACL , 2023 : 544 - 557 .

PENG Miao , LIU Ben , XU Wenjie , et al . Deja vu:contrastive historical modeling with prefix-tuning for temporal knowledge graph reasoning [C ] // Findings of the Association for Computational Linguistics:NAACL 2024 . Stroudsburg,PA,USA : ACL , 2024 : 1178 - 1191 .

XU Wenjie , LIU Ben , PENG Miao , et al . Historical facts learning from long-short terms with language model for temporal knowledge graph reasoning [J ] . Information Processing & Management , 2025 , 62 ( 3 ): 104047 .

SHANG Chao , TANG Yun , HUANG Jing , et al . End-to-end structure-aware convolutional networks for knowledge base completion [C ] // Proceedings of the Thirty-Third AAAI Conference on Artificial Intelligence and Thirty-First Innovative Applications of Artificial Intelligence Conference and Ninth AAAI Symposium on Educational Advances in Artificial Intelligence . Palo Alto, CA,USA : AAAI Press , 2019 : 3060 - 3067 .

GOEL R , KAZEMI S M , BRUBAKER M , et al . Diachronic embedding for temporal knowledge graph completion [C ] // Proceedings of the AAAI Conference on Artificial Intelligence . Palo Alto,CA,USA : AAAI Press , 2020 : 3988 - 3995 .

HAN Zhen , CHEN Peng , MA Yunpu , et al . Explainable subgraph reasoning for forecasting on temporal knowledge graphs [C ] // Proceedings of the 9th International Conference on Learning Representations . New York,USA : ICLR , 2021 : 1 - 24 .

HAN Zhen , DING Zifeng , MA Yunpu , et al . Learning neural ordinary equations for forecasting future links on temporal knowledge graphs [C ] // Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing . Stroudsburg,PA,USA : ACL , 2021 : 8352 - 8364 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Review of the Application of EBSILON Software in Combined Heat and Power System Optimization

Sensitivity Analysis of Foil Geometrical Parameters for the Third-Generation CO2 Gas Foil Bearings Based on Design of Experiments Method

I/O Performance Modeling and Prediction Method for Large-Scale Parallel Applications Using Machine Learning

Electrothermal Coupling Research on the Evolutionary Law of Anti-Corona Performance for High-Altitude Air-Cooled Generators

Mechanism and Optimal Control Strategy for Hysteresis in Rapid Dehumidification Response of Indirect Thermal Systems in Electric Vehicles

Related Author

No data

Related Institution

No data

⁰