The embedding representation of temporal knowledge graphs is one of the hotspots in the field of knowledge engineering. Existing temporal embedding models mostly integrate time information into some static embedding models in different ways to learn the temporal evolution process of entities and relationships. However, it is difficult to mine and learn some fine⁃grained temporal correlation information. Therefore, based on previous research, we propose a temporal graph embedding representation model of contextual temporal correlation in the complex space, which subdivides fine⁃grained temporal information into the relevance of knowledge start time and the consistency of knowledge time intervals. A context⁃aware temporal correlation information mining method is designed to select semantically similar contextual quadruples, mine the temporal correlation information contained in the training quadruples and contextual quadruples, and enhance the embedding model's learning of fine⁃grained temporal information. The proposed method is experimented with two public temporal knowledge graph datasets, YAGO11k and Wikidata12k, and the results show that compared with existing methods, our method has certain improvements in MRR (mean reciprocal rank) and Hits@k (k=1,3,10) indicators.
目前一些时序知识图谱如YAGO3[8]、Wikidata[9]等存储了数十亿的带有时间信息的知识四元组。如在YAGO11k[10]中四元组<Brian Schmidt,graduated From,University of Arizona,[1989,1993]>表示在1989年至1993年,布莱恩·施密特从亚利桑那大学毕业,这个现实世界中的知识仅在[1989—1993]时间间隔内有效。相比于静态知识图谱,时序知识图谱表现出更复杂的时间动态。对于同一个头实体和关系,在不同的时间点可能有多个尾实体使事实成立,其关系与时间相关联。因此如何将知识的表示学习中带有时间信息也是亟须解决的问题之一。许多研究者在TransE模型的基础上引入时间信息,以完成时序知识图谱的嵌入表示,如TTransE[11]、HyTE[10]、ATiSE[12]、TeRo[13]等模型。
在时序知识图谱中,对于语义信息相似的事实,其包含的时间信息往往存在着某种细粒度的关联,并且语义越接近,其带有的时序信息关联性也越高。如在YAGO11k时序知识图谱中,四元组<Brian Schmidt,graduated From, University of Arizona, [1989,1993]>和四元组<Studs Terkel, gra⁃duated From, University of Chicago,[1932,1936]>静态知识的语义信息十分相似,都是描述某个人毕业于某个大学,并且知识的持续时间都为4年,关联性较高。而现有的嵌入模型大多都是基于负采样的方式进行训练,最终使正样本和负样本的得分函数值的差距尽可能地大,以此区分正负样本。因此将语义相似事实的时序关联信息与负采样损失函数相结合,不仅能助力于时序嵌入模型学习和挖掘细粒度的时序信息,同时还能够助力于区分正负样本。
SUNY W, CHENGG, LIX, et al. Graph matching network for interpretable complex question answering over knowledge graphs[J]. Journal of Computer Research and Development, 2021, 58(12): 2673-2683. DOI: 10.7544/issn1000-1239.2021.20211004 (Ch ).
[3]
TAIC Y, HUANGL Y, HUANGC K, et al. User-centric path reasoning towards explainable recommendation[C]//Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2021: 879-889. DOI: 10.1145/3404835.3462847 .
[4]
JIH Z, KEP, HUANGS H, et al. Language generation with multi-hop reasoning on commonsense knowledge graph[C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). Stroudsburg: Association for Computational Linguistics, 2020: 725-736. DOI: 10.18653/v1/2020.emnlp-main.54 .
[5]
BOLLACKERK, EVANSC, PARITOSHP, et al. Freebase: A collaboratively created graph database for structuring human knowledge[C]//Proceedings of the 2008 ACM SIGMOD International Conference on Management of Data. New York: ACM, 2008: 1247-1250. DOI: 10.1145/1376616.1376746 .
[6]
AUERS, BIZERC, KOBILAROVG, et al. DBpedia: A nucleus for a web of open data[M]//The Semantic Web. Berlin: Springer, 2007: 722-735. DOI: 10.1007/978-3-540-76298-0_52 .
[7]
BORDESA, USUNIERN, GARCIA-DURANA, et al. Translating embeddings for modeling multi-relational data[C] //Proc of Advances in Neural Information Processing Systems. Cambridge: MIT Press, 2013: 2787-2795.
[8]
SUNZ Q, DENGZ H, NIEJ Y, et al. Rotate: Knowledge Graph Embedding by Relational Rotation in Complex Space[C/OL].[2020-09-10]. DOI: 10.1109/iccc52777.2021.9580273 .
[9]
MAHDISOLTANIF, BIEGAJ, SUCHANEKF M. YAGO3: A knowledge base from multilingual wikipedias[C/OL].[2022-11-11].
[10]
ERXLEBENF, GÜNTHERM, KRÖTZSCHM, et al. Introducing Wikidata to the linked data web[M]//The Semantic Web — ISWC 2014. Cham: Springer International Publishing, 2014: 50-65. DOI: 10.1007/978-3-319-11964-9_4 .
[11]
DASGUPTAS S, RAYS N, TALUKDARP. HyTE: hyperplane-based temporally aware knowledge graph embedding[C]//Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: Association for Computational Linguistics, 2018: 2001-2011. DOI: 10.18653/v1/d18-1225 .
[12]
LEBLAYJ, CHEKOLM W. Deriving validity time in knowledge graph[C]//Proceedings of the The Web Conference 2018. New York: ACM, 2018: 1771-1776. DOI: 10.1145/3184558.3191639 .
[13]
XUC J, NAYYERIM, ALKHOURYF, et al. Temporal knowledge graph embedding model based on additive time series decomposition[EB/OL]. 2019: arXiv: 1911.07893. DOI: 10.18653/v1/2020.coling-main.139 .
[14]
XUC J, NAYYERIM, ALKHOURYF, et al. TeRo: A time-aware knowledge graph embedding via temporal rotation[C]//Proceedings of the 28th International Conference on Computational Linguistics. Stroudsburg: International Committee on Computational Linguistics, 2020: 1583-1593. DOI: 10.18653/v1/2020.coling-main.139 .
[15]
WANGZ, ZHANGJ W, FENGJ L, et al. Knowledge graph embedding by translating on hyperplanes[C]//Proceedings of the 28th AAAI Conference on Artificial Intelligence. New York: ACM, 2014: 1112-1119. DOI: 10.5555/2893873.2894046 .
[16]
LINY K, LIUZ Y, SUNM S, et al. Learning entity and relation embeddings for knowledge graph completion[C]//Proceedings of the 29th AAAI Conference on Artificial Intelligence. New York: ACM, 2015: 2181-2187. DOI: 10.5555/2886521.2886624 .
[17]
JIG L, HES Z, XUL H, et al. Knowledge graph embedding via dynamic mapping matrix[C]//Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing. Stroudsburg: Association for Computational Linguistics, 2015: 687-696. DOI: 10.3115/v1/p15-1067 .
YAOS Y, ZHAOT Z, WANGR J, et al. Rule-guided joint embedding learning of knowledge graphs[J]. Journal of Computer Research and Development, 2020, 57(12): 2514-2522. DOI: 10.7544/issn1000-1239.2020.20200741(Ch ).
[20]
YANGB S, YIHW T, HEX D, et al. Embedding entities and relations for learning and inference in knowledge bases[C/OL]. [2020-09-10].
[21]
TROUILLONT, WELBLJ, RIEDELS, et al. Complex embeddings for simple link prediction[C]//Proceedings of the 33rd International Conference on International Conference on Machine Learning. New York: ACM, 2016: 2071-2080. DOI: 10.5555/3045390.3045609 .
[22]
ZHANGS, TAY Y, YAOL N, et al. Quaternion knowledge graph embeddings[C]// Proc of the 33rd International Conf on Neural Information Processing Systems. Cambridge: MIT Press, 2019: 2735-2745.
[23]
XUC J, NAYYERIM, CHENY Y, et al. Knowledge graph embeddings in geometric algebras[C]//Proceedings of the 28th International Conference on Computational Linguistics. Stroudsburg: International Committee on Computational Linguistics, 2020: 530-544. DOI: 10.18653/v1/2020.coling-main.46 .
CUIY N, LIJ, SHENL, et al. Duration-HyTE: A time-aware knowledge representation learning method based on duration modeling[J]. Journal of Computer Research and Development, 2020, 57(6): 1239-1251. DOI: 10.7544/issn1000-1239.2020.20190253 (Ch ).
[26]
MAY P, TRESPV, DAXBERGERE A. Embedding models for episodic knowledge graphs[J]. Journal of Web Semantics, 2019, 59: 100490. DOI: 10.1016/j.websem.2018.12.008 .
[27]
GARCÍA-DURÁNA, DUMANČIĆS, NIEPERTM. Learning sequence encoders for temporal knowledge graph completion[C]//Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: Association for Computational Linguistics, 2018: 4816-4821. DOI: 10.18653/v1/d18-1516 .
[28]
GOELR, KAZEMIS M, BRUBAKERM, et al. Diachronic embedding for temporal knowledge graph completion[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(4): 3988-3995. DOI: 10.1609/aaai.v34i04.5815 .
[29]
LACROIXT, OBOZINSKIG, USUNIERN. Tensor decompositions for temporal knowledge base completion[EB/OL]. 2020: arXiv: 2004.04926.