Joint Language Semantic and Structure Embedding for Knowledge Graph Completion

Jianhao Shen, Chenguang Wang, Linyuan Gong, Dawn Song


Abstract
The task of completing knowledge triplets has broad downstream applications. Both structural and semantic information plays an important role in knowledge graph completion. Unlike previous approaches that rely on either the structures or semantics of the knowledge graphs, we propose to jointly embed the semantics in the natural language description of the knowledge triplets with their structure information. Our method embeds knowledge graphs for the completion task via fine-tuning pre-trained language models with respect to a probabilistic structured loss, where the forward pass of the language models captures semantics and the loss reconstructs structures. Our extensive experiments on a variety of knowledge graph benchmarks have demonstrated the state-of-the-art performance of our method. We also show that our method can significantly improve the performance in a low-resource regime, thanks to the better use of semantics. The code and datasets are available at https://github.com/pkusjh/LASS.
Anthology ID:
2022.coling-1.171
Volume:
Proceedings of the 29th International Conference on Computational Linguistics
Month:
October
Year:
2022
Address:
Gyeongju, Republic of Korea
Venue:
COLING
SIG:
Publisher:
International Committee on Computational Linguistics
Note:
Pages:
1965–1978
Language:
URL:
https://aclanthology.org/2022.coling-1.171
DOI:
Bibkey:
Cite (ACL):
Jianhao Shen, Chenguang Wang, Linyuan Gong, and Dawn Song. 2022. Joint Language Semantic and Structure Embedding for Knowledge Graph Completion. In Proceedings of the 29th International Conference on Computational Linguistics, pages 1965–1978, Gyeongju, Republic of Korea. International Committee on Computational Linguistics.
Cite (Informal):
Joint Language Semantic and Structure Embedding for Knowledge Graph Completion (Shen et al., COLING 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.coling-1.171.pdf
Code
 pkusjh/lass
Data
FB15kFB15k-237UMLSWN18WN18RR