Rule-based data augmentation for knowledge graph embedding
Knowledge graph (KG) embedding models suffer from the incompleteness issue of observed facts. Different from existing solutions that incorporate additional information or employ expressive and complex embedding techniques, we propose to augment KGs by iteratively mining logical rules from the observ...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
KeAi Communications Co. Ltd.
2021-01-01
|
| Series: | AI Open |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666651021000267 |
| _version_ | 1831651703106043904 |
|---|---|
| author | Guangyao Li Zequn Sun Lei Qian Qiang Guo Wei Hu |
| author_facet | Guangyao Li Zequn Sun Lei Qian Qiang Guo Wei Hu |
| author_sort | Guangyao Li |
| collection | DOAJ |
| description | Knowledge graph (KG) embedding models suffer from the incompleteness issue of observed facts. Different from existing solutions that incorporate additional information or employ expressive and complex embedding techniques, we propose to augment KGs by iteratively mining logical rules from the observed facts and then using the rules to generate new relational triples. We incrementally train KG embeddings with the coming of new augmented triples, and leverage the embeddings to validate these new triples. To guarantee the quality of the augmented data, we filter out the noisy triples based on a propagation mechanism during the validation. The mined rules and rule groundings are human-understandable, and can make the augmentation procedure reliable. Our KG augmentation framework is applicable to any KG embedding models with no need to modify their embedding techniques. Our experiments on two popular embedding-based tasks (i.e., entity alignment and link prediction) show that the proposed framework can bring significant improvement to existing KG embedding models on most benchmark datasets. |
| first_indexed | 2024-12-19T15:33:49Z |
| format | Article |
| id | doaj.art-35b8d65e446149a3b02c28e8a09c76c2 |
| institution | Directory Open Access Journal |
| issn | 2666-6510 |
| language | English |
| last_indexed | 2024-12-19T15:33:49Z |
| publishDate | 2021-01-01 |
| publisher | KeAi Communications Co. Ltd. |
| record_format | Article |
| series | AI Open |
| spelling | doaj.art-35b8d65e446149a3b02c28e8a09c76c22022-12-21T20:15:39ZengKeAi Communications Co. Ltd.AI Open2666-65102021-01-012186196Rule-based data augmentation for knowledge graph embeddingGuangyao Li0Zequn Sun1Lei Qian2Qiang Guo3Wei Hu4State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, ChinaState Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, ChinaState Key Laboratory of Mathematical Engineering and Advanced Computing, Wuxi, ChinaState Key Laboratory of Mathematical Engineering and Advanced Computing, Wuxi, ChinaState Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China; State Key Laboratory of Mathematical Engineering and Advanced Computing, Wuxi, China; Corresponding author. State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China.Knowledge graph (KG) embedding models suffer from the incompleteness issue of observed facts. Different from existing solutions that incorporate additional information or employ expressive and complex embedding techniques, we propose to augment KGs by iteratively mining logical rules from the observed facts and then using the rules to generate new relational triples. We incrementally train KG embeddings with the coming of new augmented triples, and leverage the embeddings to validate these new triples. To guarantee the quality of the augmented data, we filter out the noisy triples based on a propagation mechanism during the validation. The mined rules and rule groundings are human-understandable, and can make the augmentation procedure reliable. Our KG augmentation framework is applicable to any KG embedding models with no need to modify their embedding techniques. Our experiments on two popular embedding-based tasks (i.e., entity alignment and link prediction) show that the proposed framework can bring significant improvement to existing KG embedding models on most benchmark datasets.http://www.sciencedirect.com/science/article/pii/S2666651021000267Knowledge graph embeddingData augmentationLogical rules |
| spellingShingle | Guangyao Li Zequn Sun Lei Qian Qiang Guo Wei Hu Rule-based data augmentation for knowledge graph embedding AI Open Knowledge graph embedding Data augmentation Logical rules |
| title | Rule-based data augmentation for knowledge graph embedding |
| title_full | Rule-based data augmentation for knowledge graph embedding |
| title_fullStr | Rule-based data augmentation for knowledge graph embedding |
| title_full_unstemmed | Rule-based data augmentation for knowledge graph embedding |
| title_short | Rule-based data augmentation for knowledge graph embedding |
| title_sort | rule based data augmentation for knowledge graph embedding |
| topic | Knowledge graph embedding Data augmentation Logical rules |
| url | http://www.sciencedirect.com/science/article/pii/S2666651021000267 |
| work_keys_str_mv | AT guangyaoli rulebaseddataaugmentationforknowledgegraphembedding AT zequnsun rulebaseddataaugmentationforknowledgegraphembedding AT leiqian rulebaseddataaugmentationforknowledgegraphembedding AT qiangguo rulebaseddataaugmentationforknowledgegraphembedding AT weihu rulebaseddataaugmentationforknowledgegraphembedding |