AI帮你理解科学

AI 生成解读视频

AI抽取解析论文重点内容自动生成视频


pub
生成解读视频

AI 溯源

AI解析本论文相关学术脉络


Master Reading Tree
生成 溯源树

AI 精读

AI抽取本论文的概要总结


微博一下
This paper presents a new entity-pair embedding approach for Knowledge Graph alignment

Knowledge Graph Alignment with Entity Pair Embedding

EMNLP 2020, pp.1672-1680, (2020)

被引用0|浏览317
下载 PDF 全文
引用
微博一下

摘要

Knowledge Graph (KG) alignment is to match entities in different KGs, which is important to knowledge fusion and integration. Recently, a number of embedding-based approaches for KG alignment have been proposed and achieved promising results. These approaches first embed entities in low-dimensional vector spaces, and then obtain entity al...更多

代码

数据

0
简介
  • Knowledge graphs (KGs) have been built and applied in several domains, including question answering (Zhang et al, 2018), recommendation (Sun et al, 2018b), and information extraction (Yang and Mitchell, 2017).
  • A number of embedding-based approaches have been proposed, including MTransE (Chen et al, 2017), JAPE (Sun et al, 2017), IPTransE (Zhu et al, 2017), GCN-Align (Wang et al, 2018), RDGCN (Wu et al, 2019), and MultiKE (Zhang et al, 2019), etc
  • These approaches first embed entities in low-dimensional vector spaces, and obtain the entity alignments by computations on their vector representations.
  • Equivalent relations of entities can be accurately predicted based on the learned embeddings of entity-pairs
重点内容
  • Knowledge graphs (KGs) have been built and applied in several domains, including question answering (Zhang et al, 2018), recommendation (Sun et al, 2018b), and information extraction (Yang and Mitchell, 2017)
  • To get more accurate alignment results, we propose an entity-pair embedding approach for KG alignment (EPEA)
  • We introduce the definition of pairwise connectivity graph (PCG) of KGs, whose nodes are entity-pairs and edges correspond to relation-pairs
  • We propose a similarity feature extraction method based on convolutional neural network (CNN), which automatically generates feature vectors of entity-pairs encoding their attribute similarities
  • We propose an attribute feature extraction method based on Convolutional Neural Network (CNN)
  • Experiments on five real-world datasets show that our approach can achieve the stateof-the-art KG alignment results
  • This paper presents a new entity-pair embedding approach for KG alignment
方法
  • Five datasets are used to evaluate the approach, each dataset contains two knowledge graphs to be aligned.
  • DBP15KZH−EN, DBP15KJA−EN and DBP15KFR−EN were built by (Sun et al, 2017).
  • They are generated from DBpedia and each dataset contains 15 thousand aligned entity.
  • The authors use Hits@k and MRR(Mean reciprocal ranking) as the evaluation metrics, which are popular and widely used in other KG alignment work.
  • Best configurations for two models in the approach are selected based on the MRR
结果
  • Because all the approaches use the same sets of seeding and testing alignments in each dataset, the results of the compared approaches are obtained from their original papers.
  • It shows that the approach EPEA achieves promising improvements compared with the previous approaches.
  • In terms of Hits@1, CEA performs better than RDGCN and MultiKE on cross-lingual and monolingual datasets, respectively.
  • CEA performs better than EPEA on DBPFR−EN and DBP-WD, but the results of two approaches are close, with small differences of 1.7% and 2.3%
结论
  • This paper presents a new entity-pair embedding approach for KG alignment.
  • The authors' approach first extracts useful attribute features of entity-pairs by using a convolutional neural network, and propagates the features among the neighbors of entitypairs, by using a graph neural network with edgeaware attentions.
  • The embeddings are learned with the object of separating equivalent and nonequivalent entity-pairs.
  • Experiments on five real-world datasets show that the approach achieves the stateof-the-art results
总结
  • Introduction:

    Knowledge graphs (KGs) have been built and applied in several domains, including question answering (Zhang et al, 2018), recommendation (Sun et al, 2018b), and information extraction (Yang and Mitchell, 2017).
  • A number of embedding-based approaches have been proposed, including MTransE (Chen et al, 2017), JAPE (Sun et al, 2017), IPTransE (Zhu et al, 2017), GCN-Align (Wang et al, 2018), RDGCN (Wu et al, 2019), and MultiKE (Zhang et al, 2019), etc
  • These approaches first embed entities in low-dimensional vector spaces, and obtain the entity alignments by computations on their vector representations.
  • Equivalent relations of entities can be accurately predicted based on the learned embeddings of entity-pairs
  • Methods:

    Five datasets are used to evaluate the approach, each dataset contains two knowledge graphs to be aligned.
  • DBP15KZH−EN, DBP15KJA−EN and DBP15KFR−EN were built by (Sun et al, 2017).
  • They are generated from DBpedia and each dataset contains 15 thousand aligned entity.
  • The authors use Hits@k and MRR(Mean reciprocal ranking) as the evaluation metrics, which are popular and widely used in other KG alignment work.
  • Best configurations for two models in the approach are selected based on the MRR
  • Results:

    Because all the approaches use the same sets of seeding and testing alignments in each dataset, the results of the compared approaches are obtained from their original papers.
  • It shows that the approach EPEA achieves promising improvements compared with the previous approaches.
  • In terms of Hits@1, CEA performs better than RDGCN and MultiKE on cross-lingual and monolingual datasets, respectively.
  • CEA performs better than EPEA on DBPFR−EN and DBP-WD, but the results of two approaches are close, with small differences of 1.7% and 2.3%
  • Conclusion:

    This paper presents a new entity-pair embedding approach for KG alignment.
  • The authors' approach first extracts useful attribute features of entity-pairs by using a convolutional neural network, and propagates the features among the neighbors of entitypairs, by using a graph neural network with edgeaware attentions.
  • The embeddings are learned with the object of separating equivalent and nonequivalent entity-pairs.
  • Experiments on five real-world datasets show that the approach achieves the stateof-the-art results
表格
  • Table1: Details of the datasets
  • Table2: Results of KG alignment
Download tables as Excel
相关工作
  • A number of embedding-based entity alignment approaches have been proposed recently. Some approaches mainly rely on the structure information in KGs to find alignments, including MTransE (Chen et al, 2017), IPTransE (Zhu et al, 2017), BootEA (Sun et al, 2018a),
基金
  • The work is supported by the National Key Research and Development Program of China (No 2017YFB1402105). Zhichun Wang, Qingsong Lv, Xiaohan Lan, and Yu Zhang. 2018
研究对象与分析
real-world datasets: 5
To get desirable embeddings, a convolutional neural network is used to generate similarity features of entity-pairs from their attributes; and a graph neural network is employed to propagate the similarity features and get the final embeddings of entitypairs. Experiments on five real-world datasets show that our approach can achieve the stateof-the-art KG alignment results. 4.1 Datasets

Five datasets are used to evaluate our approach, each dataset contains two knowledge graphs to be aligned

datasets: 5
Experiments on five real-world datasets show that our approach can achieve the stateof-the-art KG alignment results. 4.1 Datasets

Five datasets are used to evaluate our approach, each dataset contains two knowledge graphs to be aligned
. Table 1 outlines the detail information of these datasets

real-world datasets: 5
To get desirable embeddings, a convolutional neural network is used to generate similarity features of entity-pairs from their attributes; and a graph neural network is employed to propagate the similarity features and get the final embeddings of entitypairs. Experiments on five real-world datasets show that our approach can achieve the stateof-the-art KG alignment results. Knowledge graphs (KGs) have been built and applied in several domains, including question answering (Zhang et al, 2018), recommendation (Sun et al, 2018b), and information extraction (Yang and Mitchell, 2017)

aligned entity pairs: 100000
DBPWD and DBP-YG were first used in (Sun et al, 2018a), which are generated from DBpedia, Wikidata and YAGO3. Each dataset contains 100 thousand aligned entity pairs. For all the datasets, we use the same training/testing split of aligned entity pairs with previous work (Sun et al, 2017, 2018a), 30% for training and 70% for testing

datasets: 5
4.1 Datasets. Five datasets are used to evaluate our approach, each dataset contains two knowledge graphs to be aligned. Table 1 outlines the detail information of these datasets

datasets: 5
It shows that our approach EPEA achieves promising improvements compared with the previous approaches. Our approach outperforms all the compared approaches other than CEA on five datasets, in terms of Hits@1, Hits@10 and MRR. Taking no account of CEA, RDGCN achieved the stateof-the-art results on three cross-lingual datasets

cross-lingual datasets: 3
Our approach outperforms all the compared approaches other than CEA on five datasets, in terms of Hits@1, Hits@10 and MRR. Taking no account of CEA, RDGCN achieved the stateof-the-art results on three cross-lingual datasets. Compared with RDGCN, our approach gets improvements of 17.7%, 15.7%, and 6.9% of Hits@1 on these datasets

real-world datasets: 5
The embeddings are learned with the object of separating equivalent and nonequivalent entity-pairs. Experiments on five real-world datasets show that our approach achieves the stateof-the-art results. The work is supported by the National Key Research and Development Program of China (No 2017YFB1402105)

引用论文
  • Yixin Cao, Zhiyuan Liu, Chengjiang Li, Zhiyuan Liu, Juanzi Li, and Tat-Seng Chua. 2019. Multi-channel graph neural network for entity alignment. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (ACL2019), pages 1452–1461.
    Google ScholarLocate open access versionFindings
  • Muhao Chen, Yingtao Tian, Mohan Yang, and Carlo Zaniolo. 2017. Multilingual knowledge graph embeddings for cross-lingual knowledge alignment. In Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence (AAAI2017), pages 1511–1517.
    Google ScholarLocate open access versionFindings
  • Ernesto Jimenez-Ruiz and Bernardo Cuenca Grau. 2011. Logmap: Logic-based and scalable ontology matching. In Proceedings of the Tenth International Semantic Web Conference(ISWC2011), pages 273– 288.
    Google ScholarLocate open access versionFindings
  • S. Melnik, H. Garcia-Molina, and E. Rahm. 2002. Similarity flooding: a versatile graph matching algorithm and its application to schema matching. In Proceedings 18th International Conference on Data Engineering (ICDE2002), pages 117–128.
    Google ScholarLocate open access versionFindings
  • Zequn Sun, Wei Hu, and Chengkai Li. 2017. Cross-lingual entity alignment via joint attributepreserving embedding. In Proceedings of the Sixteenth International Semantic Web Conference (ISWC2017), pages 628–644.
    Google ScholarLocate open access versionFindings
  • Zequn Sun, Wei Hu, Qingheng Zhang, and Yuzhong Qu. 2018a. Bootstrapping entity alignment with knowledge graph embedding. In Proceedings of the Twenty-Seventh international joint conference on Artificial Intelligence (IJCAI2018), pages 4396–4402.
    Google ScholarLocate open access versionFindings
  • Zhu Sun, Jie Yang, Jie Zhang, Alessandro Bozzon, Long-Kai Huang, and Chi Xu. 2018b. Recurrent knowledge graph embedding for effective recommendation. In Proceedings of the 12th ACM Conference on Recommender Systems(RecSys2018), pages 297–305.
    Google ScholarLocate open access versionFindings
  • Petar Velickovic, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Lio, and Yoshua Bengio. 2017. Graph attention networks. CoRR, abs/1710.10903.
    Findings
  • Yuting Wu, Xiao Liu, Yansong Feng, Zheng Wang, Rui Yan, and Dongyan Zhao. 201Relationaware entity alignment for heterogeneous knowledge graphs. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI2019), pages 5278–5284.
    Google ScholarLocate open access versionFindings
  • Bishan Yang and Tom Mitchell. 2017. Leveraging knowledge bases in LSTMs for improving machine reading. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (ACL2017), pages 1436–1446.
    Google ScholarLocate open access versionFindings
  • W. Zeng, X. Zhao, J. Tang, and X. Lin. 2020. Collective entity alignment via adaptive features. In 2020 IEEE 36th International Conference on Data Engineering (ICDE2020), pages 1870–1873.
    Google ScholarLocate open access versionFindings
  • Wei Hu Muhao Chen Jian Dai Wei Zhang Yuzhong Qu Zequn Sun, Chengming Wang. 2020. Knowledge graph alignment network with gated multi-hop neighborhood aggregation. In Proceedings of the Twenty-Nineth International Joint Conference on Artificial Intelligence (AAAI2020).
    Google ScholarLocate open access versionFindings
  • Qingheng Zhang, Zequn Sun, Wei Hu, Muhao Chen, Lingbing Guo, and Yuzhong Qu. 2019. Multiview knowledge graph embedding for entity alignment. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI2019), pages 5429–5435.
    Google ScholarLocate open access versionFindings
  • Yuyu Zhang, Hanjun Dai, Zornitsa Kozareva, Alexander J Smola, and Le Song. 2018. Variational reasoning for question answering with knowledge graph. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18), pages 6069–6076.
    Google ScholarLocate open access versionFindings
  • Hao Zhu, Ruobing Xie, Zhiyuan Liu, and Maosong Sun. 2017. Iterative entity alignment via joint knowledge embeddings. In Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence (IJCAI2017), pages 4258–4264.
    Google ScholarLocate open access versionFindings
  • Qiannan Zhu, Xiaofei Zhou, Jia Wu, Jianlong Tan, and Li Guo. 2019. Neighborhood-aware attentional representation for multilingual knowledge graphs. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI2019), pages 1943–1949.
    Google ScholarLocate open access versionFindings
  • Zhichun Wang, Juanzi Li, Zhigang Wang, and Jie Tang. 2012. Cross-lingual knowledge linking across wiki knowledge bases. In Proceedings of the 21st International Conference on World Wide Web (WWW2012), pages 459–468.
    Google ScholarLocate open access versionFindings
作者
Jinjian Yang
Jinjian Yang
Xiaoju Ye
Xiaoju Ye
您的评分 :
0

 

标签
评论
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn
小科