基于属性增强的对偶图实体对齐算法
Attribute Augmentation Based Alignment Algorithm for Pairs of Dyadic Graph Entity Alignment
DOI: 10.12677/CSA.2023.135114, PDF,   
作者: 姚 荣：成都信息工程大学软件工程学院，四川 成都
关键词: 属性增强；对偶关系图；实体对齐；知识图谱；Attribute Augmentation； Dyadic Relationship Graphs； Entity Alignment； Knowledge Graphs
摘要: 本文提出结合属性信息的对偶图实体对齐算法针对基于关系感知的双对偶关系图算法中没有考虑到的属性信息进行优化，对属性结构嵌入向量使用图卷积神经网络算法对邻居节点抽取信息，并使用对偶关系图和注意力机制抽取实体对中的关系信息，最后通过结合实体对的关系信息和属性信息的相似度，判断是否为同一实体。针对原算法中识别效率不高的异构知识图谱实体对提升效果明显。在数据集DBP15K的三个跨语言数据集ZH-EN，JA-EN，FR-EN上实验，实验结果验证了对偶注意力以及属性信息对实体对齐方法的有效性。
Abstract: In this paper, we propose the dyadic graph entity alignment algorithm DAI combining attribute information to optimize the attribute information that is not considered in the dual dyadic relational graph algorithm RDGCN based on relationship awareness, use the graph convolutional neural network algorithm for attribute structure embedding vector to extract information from neighbor nodes, and use the dyadic relational graph and attention mechanism to extract the relational in-formation in entity pairs, and finally by combining the relational. The similarity of relationship information and attribute information of entity pairs is finally judged whether they are the same entity or not. The improvement effect is obvious for the heterogeneous knowledge mapping entity pairs which are not recognized efficiently in the original algorithm. Experiments on three cross-lingual datasets ZH-EN, JA-EN, and FR-EN of dataset DBP15K are conducted, and the experi-mental results verify the effectiveness of pairwise attention and attribute information on entity alignment methods.
文章引用：姚荣. 基于属性增强的对偶图实体对齐算法[J]. 计算机科学与应用, 2023, 13(5): 1166-1177. https://doi.org/10.12677/CSA.2023.135114

参考文献

[1] Elfeky, M.G., Verykios, V.S. and Elmagarmid, A.K. (2002) TAILOR: A Record Linkage Toolbox. Proceedings 18th International Conference on Data Engineering, San Jose, CA, USA, 26 February 2002-1 March 2002, 17-28. [Google Scholar] [CrossRef] 
[2] Sarawagi, S. and Bhamidipaty, A. (2002) Interactive Deduplica-tion Using Active Learning. Proceedings of the Eighth ACM SIGKDD International Conference on Knowledge Discov-ery and Data Mining, Edmonton, Canada, 23-26 July 2002, 269-278. [Google Scholar] [CrossRef] 
[3] Lacoste-Julien, S., Palla, K., Davies, A., et al. (2013) SiGMa: Simple Greedy Matching for Aligning Large Knowledge Bases. Proceedings of the 19th ACM SIGKDD International Confer-ence on Knowledge Discovery and Data Mining, Chicago, IL, 11-14 August 2013, 572-580. [Google Scholar] [CrossRef] 
[4] Scharffe, F., Liu, Y. and Zhou, C. (2009) Rdf-Ai: An Architecture for Rdf Datasets Matching, Fusion and Interlink.
https://www.researchgate.net/publication/228972747_Rdf-ai_an_architecture_for_rdf_datasets_matching_fusion_and_interlink
[5] Volz, J., Bizer, C., Gaedke, M., et al. (2009) Discovering and Maintaining Links on the Web of Data. In: Bernstein, A., Karger, D.R., Heath, T., Feigenbaum, L., Maynard, D., Motta, E. and Thirunarayan, K., eds., The Se-mantic Web—ISWC 2009, Springer, Berlin, 650-665. [Google Scholar] [CrossRef] 
[6] Hao, Y., Zhang, Y., He, S., et al. (2016) A Joint Embedding Method for Entity Alignment of Knowledge Bases. In: Chen, H., Ji, H., Sun, L., Wang, H., Qian, T. and Ruan, T., eds., CCKS 2016: Knowledge Graph and Semantic Computing: Semantic, Knowledge, and Linked Big Data, Springer, Singapore, 3-14. [Google Scholar] [CrossRef] 
[7] Bordes, A., Usunier, N., Garcia-Duran, A., et al. (2013) Trans-lating Embeddings for Modeling Multi-Relational Data.
https://proceedings.neurips.cc/paper_files/paper/2013/file/1cecc7a77928ca8133fa24680a88d2f9-Paper.pdf
[8] Lin, Y., Liu, Z., Sun, M., et al. (2015) Learning Entity and Relation Embeddings for Knowledge Graph Completion. Pro-ceedings of the AAAI Conference on Artificial Intelligence, 29, 254-266. [Google Scholar] [CrossRef] 
[9] Guan, S., Jin, X., Wang, Y., et al. (2019) Self-Learning and Embed-ding Based Entity Alignment. Knowledge and Information Systems, 59, 361-386. [Google Scholar] [CrossRef] 
[10] Sun, Z., Wei, H., Zhang, Q., et al. (2018) Bootstrapping Entity Alignment with Knowledge Graph Embedding. Proceedings of the 27th International Joint Conference on Artificial Intel-ligence, Stockholm, Sweden, 13-19 July 2018, 4396-4402. [Google Scholar] [CrossRef] 
[11] Sun, Z., Hu, W. and Li, C. (2017) Cross-Lingual Entity Alignment via Joint Attribute-Preserving Embedding. In: d’Amato, C., Fernandez, M., Tamma, V., Lecue, F., Cudré-Mauroux, P., Sequeda, J., Lange, C. and Heflin, J., eds., The Semantic Web—ISWC 2017, Springer, Cham, 628-644. [Google Scholar] [CrossRef] 
[12] Trisedya, B.D., Qi, J. and Zhang, R. (2019) Entity Alignment between Knowledge Graphs Using Attribute Embeddings. Proceedings of the AAAI Conference on Artificial Intelligence, 33, 297-304. [Google Scholar] [CrossRef] 
[13] He, F., Li, Z., Qiang, Y., et al. (2019) Unsupervised Entity Alignment Using Attribute Triples and Relation Triples. In: Li, G., Yang, J., Gama, J., Natwichai, J. and Tong, Y., eds., DASFAA 2019: Database Systems for Advanced Applications, Springer, Cham, 367-382. [Google Scholar] [CrossRef]