时序知识图谱表示学习方法综述

doi:10.11871/jfdc.issn.2096-742X.2026.05.013

数据与计算发展前沿 ›› 2026, Vol. 8 ›› Issue (5): 165-176.

CSTR: 32002.14.jfdc.CN10-1649/TP.2026.05.013

doi: 10.11871/jfdc.issn.2096-742X.2026.05.013

时序知识图谱表示学习方法综述

申舒亦(),王一帆,卜凡亮^*()

中国人民公安大学，信息网络安全学院，北京 100038

收稿日期:2026-03-16 出版日期:2026-10-20 发布日期:2026-10-20
通讯作者: ^*卜凡亮（E-mail: bufanliang@sina.com）
作者简介:申舒亦，中国人民公安大学，硕士研究生，CCF学生会员，主要研究方向为自然语言处理、知识图谱。
本文主要承担工作为论文的撰写和修改。
SHEN Shuyi is a master’s student at People’s Public Security University of China, and a CCF student member. Her main research interests include natural language processing and knowledge graphs.
In this paper, she is primarily responsible for the writing and revision of the manuscript. E-mail: 2024211547@stu.ppsuc.edu.cn|卜凡亮，中国人民公安大学，博士，教授，主要研究方向为安全防范工程、大数据技术与工程和人工智能等。
本文主要承担工作为论文的指导和修改。
BU Fanliang, Ph.D., is a professor at People’s Public Security University of China. His main research interests include security protection engineering, big data technology and engineering, and artificial intelligence.
In this paper, he is primarily responsible for the revision and final review of the manuscript. E-mail: bufanliang@sina.com

A Survey on Temporal Knowledge Graph Representation Learning Methods

SHEN Shuyi(),WANG Yifan,BU Fanliang^*()

School of Information Network Security, People’s Public Security University of China, Beijing 100038, China

Received:2026-03-16 Online:2026-10-20 Published:2026-10-20

摘要/Abstract

摘要：

【目的】时序知识图谱通过引入时间维度刻画实体关系的动态演化过程，是事件预测与动态关系建模的重要基础。本文对时序知识图谱表示学习方法进行系统梳理，分析各类方法的建模思路和研究趋势，为相关领域研究提供参考。【方法】从建模范式角度将现有方法归纳为时间增强嵌入、张量分解、图神经网络和基于历史序列的动态演化四类，并基于时间建模、结构依赖建模和历史信息利用三个维度进行统一比较分析，同时整理了常用数据集与评估指标。【结论】分析表明，不同方法在时间表达粒度、结构建模能力及历史依赖利用方面存在显著差异，现有方法在复杂时间特征刻画和长期依赖建模等方面存在不足，未来研究需关注多源信息融合与统一建模框架的构建。

关键词: 时序知识图谱, 表示学习, 时间建模, 动态演化

Abstract:

[Purpose] Temporal knowledge graphs incorporate temporal information to model the dynamic evolution of entity relations, serving as a foundation for event prediction and dynamic relational modeling. This paper presents a systematic review of representation learning methods, analyzing their modeling paradigms and research trends, and providing a reference for research in related fields. [Methods] Existing approaches are categorized into four paradigms: time-enhanced embedding, tensor decomposition, graph neural networks, and history-based dynamic evolution. A unified comparison is conducted from three aspects: temporal modeling, structural dependency modeling, and historical information utilization. Commonly used datasets and evaluation metrics are also summarized. [Conclusions] Significant differences exist among methods in temporal granularity, structural modeling capability, and the exploitation of historical dependencies. Current approaches remain limited in modeling complex temporal patterns and long-term dependencies. Future work should emphasize multi-source information fusion and unified modeling frameworks.

Key words: temporal knowledge graph, representation learning, temporal modeling, dynamic evolution

申舒亦, 王一帆, 卜凡亮. 时序知识图谱表示学习方法综述[J]. 数据与计算发展前沿, 2026, 8(5): 165-176.

SHEN Shuyi, WANG Yifan, BU Fanliang. A Survey on Temporal Knowledge Graph Representation Learning Methods[J]. Frontiers of Data and Computing, 2026, 8(5): 165-176, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2026.05.013.

图/表 8

图1

图2

图3

图4

表1

表2

表3

表4

参考文献 53

[1]	何巍. 社交网络舆情多模态知识图谱构建框架研究[J]. 情报杂志, 2024, 43(1): 160-166.
[2]	段卓廷, 纪成坤, 张迪, 等. 人工智能驱动的情报实践变革: 知识结构、循证过程与组织形态[J/OL]. 情报杂志, 1-9[2025-12-27]. https://link.cnki.net/urlid/61.1167.G3.20251117.0942.008.
[3]	徐增林, 盛泳潘, 贺丽荣, 等. 知识图谱技术综述[J]. 电子科技大学学报, 2016, 45(4): 589-606.
[4]	刘峤, 李杨, 段宏, 等. 知识图谱构建技术综述[J]. 计算机研究与发展, 2016, 53(3): 582-600.
[5]	LEBLAY J, CHEKOL M W. Deriving validity time in knowledge graph[C]// Companion proceedings of the the web conference 2018. 2018: 1771-1776.
[6]	SIDIROPOULOS N D, DE LATHAUWER L, FU X, et al. Tensor decomposition for signal processing and machine learning[J]. IEEE Transactions on signal processing, 2017, 65(13): 3551-3582.
[7]	HITCHCOCK F L. The expression of a tensor or a polyadic as a sum of products[J]. Journal of Mathematics and Physics, 1927, 6(1-4): 164-189.
[8]	HAN Z, DING Z, MA Y, 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. 2021: 8352-8364.
[9]	DASGUPTA S S, RAY S N, TALUKDAR P. Hyte: Hyperplane-based temporally aware knowledge graph embedding[C]// Proceedings of the 2018 conference on empirical methods in natural language processing. 2018: 2001-2011.
[10]	GARCÍA-DURÁN A, DUMANČIĆ S, NIEPERT M. Learning sequence encoders for temporal knowledge graph completion[C]// Proceedings of the 2018 conference on empirical methods in natural language processing. 2018: 4816-4821.
[11]	GONG X, QIN J, CHAI H, et al. Temporal-relational matching network for few-shot temporal knowledge graph completion[C]//International conference on database systems for advanced applications. Cham: Springer Nature Switzerland, 2023: 768-783.
[12]	王俞涵, 陈子阳, 赵翔, 等. 时序知识图谱表示与推理的研究进展与趋势[J]. 软件学报, 2024, 35(8): 3923-3951.
[13]	LIANG K, MENG L, LIU M, et al. A survey of knowledge graph reasoning on graph types: Static, dynamic, and multi-modal[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024, 46(12): 9456-9478.
[14]	沈英汉, 江旭晖, 王元卓, 等. 时态知识图谱的推理研究综述[J]. 计算机学报, 2023, 46(6): 1272-1301.
[15]	WANG J, WANG B, QIU M, et al. A survey on temporal knowledge graph completion: Taxonomy, progress, and prospects[J]. arXiv preprint arXiv:2308. 02457, 2023.
[16]	XU C, NAYYERI M, ALKHOURY F, et al. Temporal knowledge graph embedding model based on additive time series decomposition[J]. arXiv preprint arXiv:1911.07893, 2019.
[17]	XU C, NAYYERI M, ALKHOURY F, et al. TeRo: A time-aware knowledge graph embedding via temporal rotation[C]// Proceedings of the 28th International Conference on Computational Linguistics. 2020: 1583- 1593.
[18]	SADEGHIAN A, ARMANDPOUR M, COLAS A, et al. Chronor: Rotation based temporal knowledge graph embedding[C]// Proceedings of the AAAI conference on artificial intelligence. 2021, 35(7): 6471-6479.
[19]	NICKEL M, TRESP V, KRIEGEL H P. A three-way model for collective learning on multi-relational data[C]// Icml. 2011, 11(10.5555): 3104482-3104584.
[20]	YANG B, YIH W, HE X, et al. Embedding entities and relations for learning and inference in knowledge bases[J]. arXiv preprint arXiv:1412.6575, 2014.
[21]	TROUILLON T, WELBL J, RIEDEL S, et al. Complex embeddings for simple link prediction[C]// International conference on machine learning. PMLR, 2016: 2071-2080.
[22]	KAZEMI S M, POOLE D. Simple embedding for link prediction in knowledge graphs[J]. Advances in Neural Information Processing Systems, 2018, 31.
[23]	LIN L, SHE K. Tensor decomposition-based temporal knowledge graph embedding[C]// 2020 IEEE 32nd International conference on tools with artificial intelligence (ICTAI). IEEE, 2020: 969-975.
[24]	LACROIX T, OBOZINSKI G, USUNIER N. Tensor decompositions for temporal knowledge base completion[J]. arXiv preprint arXiv:2004.04926, 2020.
[25]	SHAO P, ZHANG D, YANG G, et al. Tucker decomposition-based temporal knowledge graph completion[J]. Knowledge-Based Systems, 2022, 238: 107841.
[26]	BALAŽEVIĆ I, ALLEN C, HOSPEDALES T. Tucker: Tensor factorization for knowledge graph completion[C]// Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language processing (EMNLP-IJCNLP). 2019: 5185-5194.
[27]	CAI B, XIANG Y, GAO L, et al. Temporal knowledge graph completion: A survey[J]. arXiv preprint arXiv:2201.08236, 2022.
[28]	LI Z, JIN X, LI W, et al. Temporal knowledge graph reasoning based on evolutional representation learning[C]// Proceedings of the 44th international ACM SIGIR conference on research and development in information retrieval. 2021: 408-417.
[29]	XU Y, OU J, XU H, et al. Temporal knowledge graph reasoning with historical contrastive learning[C]// Proceedings of the AAAI conference on artificial intelligence. 2023, 37(4): 4765-4773.
[30]	LI Y, SUN S, ZHAO J. Tirgn: Time-guided recurrent graph network with local-global historical patterns for temporal knowledge graph reasoning[C]// IJCAI. 2022: 2152-2158.
[31]	MINGCONG S, ZHU C, ZHANG D, et al. Multi-granularity history and entity similarity learning for temporal knowledge graph reasoning[C]// Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024: 5232-5243.
[32]	DENG S, RANGWALA H, NING Y. Dynamic knowledge graph based multi-event forecasting[C]// Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining. 2020: 1585-1595.
[33]	XU C, SU F, LEHMANN J. Time-aware graph neural network for entity alignment between temporal knowledge graphs[C]// Proceedings of the 2021 conference on empirical methods in natural language processing. 2021: 8999-9010.
[34]	FENG S, YE Z, LIU Q, et al. RPHF-GNN: Recurrent perception of history-future graph neural networks for temporal knowledge graph reasoning[J]. IEEE Access, 2025.
[35]	CHO K, VAN MERRIËNBOER B, GULÇEHRE Ç, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[C]// Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP). 2014: 1724-1734.
[36]	TRIVEDI R, DAI H, WANG Y, et al. Know-evolve: Deep temporal reasoning for dynamic knowledge graphs[C]// international conference on machine learning. PMLR, 2017: 3462-3471.
[37]	JIN W, QU M, JIN X, et al. Recurrent event network: Autoregressive structure inferenceover temporal knowledge graphs[C]// Proceedings of the 2020 conference on empirical methods in natural language processing (EMNLP). 2020: 6669-6683.
[38]	ZHANG M, XIA Y, LIU Q, et al. Learning long-and short-term representations for temporal knowledge graph reasoning[C]// Proceedings of the ACM web conference 2023. 2023: 2412-2422.
[39]	ZHU C, CHEN M, FAN C, et al. Learning from history: Modeling temporal knowledge graphs with sequential copy-generation networks[C]// Proceedings of the AAAI conference on artificial intelligence. 2021, 35(5): 4732-4740.
[40]	LI Z, GUAN S, JIN X, et al. Complex evolutional pattern learning for temporal knowledge graph reasoning[C]//Proceedings of the 60th annual meeting of the association for computational linguistics (volume 2: short papers). 2022: 290-296.
[41]	LI Z, JIN X, GUAN S, et al. Search from history and reason for future: Two-stage reasoning on temporal knowledge graphs[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). 2021: 4732-4743.
[42]	SUN H, ZHONG J, MA Y, et al. Timetraveler: Reinforcement learning for temporal knowledge graph forecasting[C]// Proceedings of the 2021 conference on empirical methods in natural language processing. 2021: 8306-8319.
[43]	HAN Z, CHEN P, MA Y, et al. Explainable subgraph reasoning for forecasting on temporal knowledge graphs[C]// International conference on learning representations. 2020.
[44]	CHEN W, WU Y, WU S, et al. Cogntke: A cognitive temporal knowledge extrapolation framework[C]// Proceedings of the AAAI Conference on Artificial Intelligence. 2025, 39(14): 14815-14823.
[45]	JÄGER K. The limits of studying networks via event data: Evidence from the ICEWS dataset[J]. Journal of Global Security Studies, 2018, 3(4): 498-511.
[46]	LEETARU K, SCHRODT P A. Gdelt: Global data on events, location, and tone, 1979-2012[C]// ISA annual convention. 2013, 2(4): 1-49.
[47]	SUCHANEK F M, KASNECI G, WEIKUM G. Yago: a core of semantic knowledge[C]// Proceedings of the 16th international conference on World Wide Web. 2007: 697-706.
[48]	VRANDEČIĆ D, KRÖTZSCH M. Wikidata: a free collaborative knowledgebase[J]. Communications of the ACM, 2014, 57(10): 78-85.
[49]	JI S, PAN S, CAMBRIA E, et al. A survey on knowledge graphs: Representation, acquisition, and applications[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 33(2): 494-514.
[50]	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. 2020, 34(04): 3988-3995.
[51]	MAVROMATIS C, SUBRAMANYAM P L, IOANNIDIS V N, et al. Tempoqr: temporal question reasoning over knowledge graphs[C]// Proceedings of the AAAI conference on artificial intelligence. 2022, 36(5): 5825-5833.
[52]	FENG S, CHEN B, LIU Q, et al. A contrastive learning framework with dual gates and noise awareness for temporal knowledge graph reasoning[J]. Scientific Reports, 2025, 15(1): 18474.
[53]	ZHANG Y, KONG X, YE K, et al. Tackling sparse facts for temporal knowledge graph completion[C]// Proceedings of the ACM on Web Conference 2025. 2025: 3561-3570.

基于历史事件序列的动态演化建模框架
输入：	历史事件序列
输出：	未来事件预测或事实补全结果
1：	初始化实体、关系和时间嵌入
2：	根据时间戳构建按时间排序的事件序列
3：	对每个时间步编码事件并更新隐藏状态
4：	计算下一事件概率
5：	输出预测结果

方法	代表模型	时间建模能力	结构依赖建模	历史信息利用	计算复杂度	适用场景	局限
时间增强嵌入方法	TTransE^[5]、 HyTE^[9]、ATiSE^[16]	离散时间建模，显式时间向量	基于三元组，较弱	无显式历史建模	低	大规模知识图谱补全	表达能力有限，难建模复杂关系和长期依赖
张量分解方法	T-SimplE^[23]、 T-ComplEx^[24]、TuckERTNT^[25]	时间与关系联合建模	隐式多关系建模，适中	以静态建模为主	较高	复杂关系模式建模	计算复杂度高，对稀疏数据较敏感
图神经网络方法	RE-GCN^[28]、 TiRGN^[30]、MGESL^[31]、RPHF-GNN^[34]	离散或连续时间建模（结合RNN、时间编码）	显式图结构建模，较强	部分历史融合	高	能利用图结构信息，表达能力强	计算开销大，大规模动态图训练困难
基于历史序列的动态演化方法	Know-Evolve^[36]、 RE-NET^[37]、 CyGNet^[39]、CEN^[40]、CognTKE^[44]	连续时间，事件驱动建模	部分结合图结构，适中	显式历史依赖建模	较高	事件预测与时间推理	历史序列长时计算开销大，易信息衰减

数据集	实体数	关系数	训练集	验证集	测试集	时间戳	时间间隔
ICEWS14	7,128	230	72,826	8,941	8,963	365	24hours
ICEWS05-15	10,488	251	386,962	46,275	46,092	4,017	24hours
ICEWS18	23,033	256	373,018	45,995	49,545	365	24hours
GDELT	7,691	240	1,734,399	238,765	305,241	2,975	15mins
YAGO	10,623	10	1,615,40	19,523	20,026	189	1year
WIKI	12,554	24	539,286	67,538	63,110	232	1year

Model	ICEWS14				ICEWS05-15				ICEWS18
Model	MRR	Hits@1	Hits@3	Hits@10	MRR	Hits@1	Hits@3	Hits@10	MRR	Hits@1	Hits@3	Hits@10
TTransE^[5]	25.5	7.4	-	60.1	27.1	8.4	-	61.6	-	-	-	-
TA-TransE^[10]	27.5	9.5	-	62.5	29.9	9.6	-	66.8	-	-	-	-
HyTE^[9]	29.7	10.8	41.6	65.5	31.6	11.6	44.5	68.1	-	-	-	-
ATiSE^[16]	55.0	43.6	62.9	75.0	51.9	37.8	60.6	79.4	-	-	-	-
DistMult^[20]	43.9	32.3	-	67.2	45.6	33.7	-	69.1	-	-	-	-
ComplEx^[21]	46.7	34.7	52.7	71.6	48.1	36.2	53.5	72.9	-	-	-	-
T-ComplEx^[24]	61.9	54.2	66.1	76.7	66.5	58.3	71.6	81.1	-	-	-	-
RE-GCN^[28]	40.4	30.7	45.0	59.2	48.0	37.3	53.9	68.3	30.6	21.0	34.3	48.8
CENET^[29]	39.0	29.6	43.2	57.5	42.0	32.2	47.0	60.4	27.9	18.2	31.6	47.0
TiRGN^[30]	44.0	33.8	49.0	63.8	50.0	39.3	56.1	70.7	33.7	23.2	37.1	54.2
RPHF-GNN^[34]	42.7	31.8	45.0	63.8	46.9	35.3	53.4	68.6	31.3	20.6	35.6	52.2
RE-NET^[37]	38.3	28.7	41.4	54.5	43.3	33.4	47.8	63.1	28.8	19.1	32.4	47.5
CyGNet^[39]	35.0	25.7	39.0	53.6	36.8	26.6	41.6	56.2	24.9	15.9	28.3	42.6
CognTKE^[44]	46.1	36.5	51.1	64.5	53.1	42.6	59.4	72.7	35.2	25.2	39.9	54.7

时序知识图谱表示学习方法综述

A Survey on Temporal Knowledge Graph Representation Learning Methods

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 53

相关文章 1

编辑推荐

Metrics

本文评价