CAREA: Cotraining Attribute and Relation Embeddings for Cross-Lingual Entity Alignment in Knowledge Graphs
Knowledge graphs (KGs) are one of the most widely used techniques of knowledge organizations and have been extensively used in many application fields related to artificial intelligence, for example, web search and recommendations. Entity alignment provides a useful tool for how to integrate multili...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2020-01-01
|
| Series: | Discrete Dynamics in Nature and Society |
| Online Access: | http://dx.doi.org/10.1155/2020/6831603 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832550822021431296 |
|---|---|
| author | Baiyang Chen Xiaoliang Chen Peng Lu Yajun Du |
| author_facet | Baiyang Chen Xiaoliang Chen Peng Lu Yajun Du |
| author_sort | Baiyang Chen |
| collection | DOAJ |
| description | Knowledge graphs (KGs) are one of the most widely used techniques of knowledge organizations and have been extensively used in many application fields related to artificial intelligence, for example, web search and recommendations. Entity alignment provides a useful tool for how to integrate multilingual KGs automatically. However, most of the existing studies evaluated ignore the abundant information of entity attributes except for entity relationships. This paper sets out to investigate cross-lingual entity alignment and proposes an iterative cotraining approach (CAREA) to train a pair of independent models. The two models can extract the attribute and the relation features of multilingual KGs, respectively. In each iteration, the two models alternate to predict a new set of potentially aligned entity pairs. Besides, this method further filters through the dynamic threshold value to enhance the two models’ supervision. Experimental results on three real-world datasets demonstrate the effectiveness and superiority of the proposed method. The CAREA model improves the performance with at least an absolute increase of 3.9% across all experiment datasets. The code is available at https://github.com/ChenBaiyang/CAREA. |
| format | Article |
| id | doaj-art-2d24828d67b6491a9f65416d9d97d870 |
| institution | Kabale University |
| issn | 1026-0226 1607-887X |
| language | English |
| publishDate | 2020-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Discrete Dynamics in Nature and Society |
| spelling | doaj-art-2d24828d67b6491a9f65416d9d97d8702025-02-03T06:05:39ZengWileyDiscrete Dynamics in Nature and Society1026-02261607-887X2020-01-01202010.1155/2020/68316036831603CAREA: Cotraining Attribute and Relation Embeddings for Cross-Lingual Entity Alignment in Knowledge GraphsBaiyang Chen0Xiaoliang Chen1Peng Lu2Yajun Du3School of Computer and Software Engineering, Xihua University, Chengdu 610039, ChinaSchool of Computer and Software Engineering, Xihua University, Chengdu 610039, ChinaDepartment of Computer Science and Operations Research, University of Montreal, Montreal, QC H3C3J7, CanadaSchool of Computer and Software Engineering, Xihua University, Chengdu 610039, ChinaKnowledge graphs (KGs) are one of the most widely used techniques of knowledge organizations and have been extensively used in many application fields related to artificial intelligence, for example, web search and recommendations. Entity alignment provides a useful tool for how to integrate multilingual KGs automatically. However, most of the existing studies evaluated ignore the abundant information of entity attributes except for entity relationships. This paper sets out to investigate cross-lingual entity alignment and proposes an iterative cotraining approach (CAREA) to train a pair of independent models. The two models can extract the attribute and the relation features of multilingual KGs, respectively. In each iteration, the two models alternate to predict a new set of potentially aligned entity pairs. Besides, this method further filters through the dynamic threshold value to enhance the two models’ supervision. Experimental results on three real-world datasets demonstrate the effectiveness and superiority of the proposed method. The CAREA model improves the performance with at least an absolute increase of 3.9% across all experiment datasets. The code is available at https://github.com/ChenBaiyang/CAREA.http://dx.doi.org/10.1155/2020/6831603 |
| spellingShingle | Baiyang Chen Xiaoliang Chen Peng Lu Yajun Du CAREA: Cotraining Attribute and Relation Embeddings for Cross-Lingual Entity Alignment in Knowledge Graphs Discrete Dynamics in Nature and Society |
| title | CAREA: Cotraining Attribute and Relation Embeddings for Cross-Lingual Entity Alignment in Knowledge Graphs |
| title_full | CAREA: Cotraining Attribute and Relation Embeddings for Cross-Lingual Entity Alignment in Knowledge Graphs |
| title_fullStr | CAREA: Cotraining Attribute and Relation Embeddings for Cross-Lingual Entity Alignment in Knowledge Graphs |
| title_full_unstemmed | CAREA: Cotraining Attribute and Relation Embeddings for Cross-Lingual Entity Alignment in Knowledge Graphs |
| title_short | CAREA: Cotraining Attribute and Relation Embeddings for Cross-Lingual Entity Alignment in Knowledge Graphs |
| title_sort | carea cotraining attribute and relation embeddings for cross lingual entity alignment in knowledge graphs |
| url | http://dx.doi.org/10.1155/2020/6831603 |
| work_keys_str_mv | AT baiyangchen careacotrainingattributeandrelationembeddingsforcrosslingualentityalignmentinknowledgegraphs AT xiaoliangchen careacotrainingattributeandrelationembeddingsforcrosslingualentityalignmentinknowledgegraphs AT penglu careacotrainingattributeandrelationembeddingsforcrosslingualentityalignmentinknowledgegraphs AT yajundu careacotrainingattributeandrelationembeddingsforcrosslingualentityalignmentinknowledgegraphs |