Matches in SemOpenAlex for { <https://semopenalex.org/work/W4226142803> ?p ?o ?g. }
Showing items 1 to 54 of
54
with 100 items per page.
- W4226142803 abstract "Knowledge graph completion (KGC) aims to reason over known facts and infer the missing links. Text-based methods such as KGBERT (Yao et al., 2019) learn entity representations from natural language descriptions, and have the potential for inductive KGC. However, the performance of text-based methods still largely lag behind graph embedding-based methods like TransE (Bordes et al., 2013) and RotatE (Sun et al., 2019b). In this paper, we identify that the key issue is efficient contrastive learning. To improve the learning efficiency, we introduce three types of negatives: in-batch negatives, pre-batch negatives, and self-negatives which act as a simple form of hard negatives. Combined with InfoNCE loss, our proposed model SimKGC can substantially outperform embedding-based methods on several benchmark datasets. In terms of mean reciprocal rank (MRR), we advance the state-of-the-art by +19% on WN18RR, +6.8% on the Wikidata5M transductive setting, and +22% on the Wikidata5M inductive setting. Thorough analyses are conducted to gain insights into each component. Our code is available at https://github.com/intfloat/SimKGC ." @default.
- W4226142803 created "2022-05-05" @default.
- W4226142803 creator A5014548132 @default.
- W4226142803 creator A5050004923 @default.
- W4226142803 creator A5072097629 @default.
- W4226142803 creator A5086664647 @default.
- W4226142803 date "2022-01-01" @default.
- W4226142803 modified "2023-10-14" @default.
- W4226142803 title "SimKGC: Simple Contrastive Knowledge Graph Completion with Pre-trained Language Models" @default.
- W4226142803 doi "https://doi.org/10.18653/v1/2022.acl-long.295" @default.
- W4226142803 hasPublicationYear "2022" @default.
- W4226142803 type Work @default.
- W4226142803 citedByCount "21" @default.
- W4226142803 countsByYear W42261428032022 @default.
- W4226142803 countsByYear W42261428032023 @default.
- W4226142803 crossrefType "proceedings-article" @default.
- W4226142803 hasAuthorship W4226142803A5014548132 @default.
- W4226142803 hasAuthorship W4226142803A5050004923 @default.
- W4226142803 hasAuthorship W4226142803A5072097629 @default.
- W4226142803 hasAuthorship W4226142803A5086664647 @default.
- W4226142803 hasBestOaLocation W42261428031 @default.
- W4226142803 hasConcept C119857082 @default.
- W4226142803 hasConcept C132525143 @default.
- W4226142803 hasConcept C154945302 @default.
- W4226142803 hasConcept C204321447 @default.
- W4226142803 hasConcept C41008148 @default.
- W4226142803 hasConcept C41608201 @default.
- W4226142803 hasConcept C44083865 @default.
- W4226142803 hasConcept C80444323 @default.
- W4226142803 hasConceptScore W4226142803C119857082 @default.
- W4226142803 hasConceptScore W4226142803C132525143 @default.
- W4226142803 hasConceptScore W4226142803C154945302 @default.
- W4226142803 hasConceptScore W4226142803C204321447 @default.
- W4226142803 hasConceptScore W4226142803C41008148 @default.
- W4226142803 hasConceptScore W4226142803C41608201 @default.
- W4226142803 hasConceptScore W4226142803C44083865 @default.
- W4226142803 hasConceptScore W4226142803C80444323 @default.
- W4226142803 hasLocation W42261428031 @default.
- W4226142803 hasLocation W42261428032 @default.
- W4226142803 hasOpenAccess W4226142803 @default.
- W4226142803 hasPrimaryLocation W42261428031 @default.
- W4226142803 hasRelatedWork W2893186803 @default.
- W4226142803 hasRelatedWork W2923818335 @default.
- W4226142803 hasRelatedWork W3035116611 @default.
- W4226142803 hasRelatedWork W3044354590 @default.
- W4226142803 hasRelatedWork W3094605108 @default.
- W4226142803 hasRelatedWork W4212923699 @default.
- W4226142803 hasRelatedWork W4226361842 @default.
- W4226142803 hasRelatedWork W4287763734 @default.
- W4226142803 hasRelatedWork W4310879833 @default.
- W4226142803 hasRelatedWork W4385279070 @default.
- W4226142803 isParatext "false" @default.
- W4226142803 isRetracted "false" @default.
- W4226142803 workType "article" @default.