Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385965402> ?p ?o ?g. }
Showing items 1 to 49 of
49
with 100 items per page.
- W4385965402 abstract "The task of inductive knowledge graph completion requires models to learn inference patterns from a training graph, which can then be used to make predictions on a disjoint test graph. Rule-based methods seem like a natural fit for this task, but in practice they significantly underperform state-of-the-art methods based on Graph Neural Networks (GNNs), such as NBFNet. We hypothesise that the underperformance of rule-based methods is due to two factors: (i) implausible entities are not ranked at all and (ii) only the most informative path is taken into account when determining the confidence in a given link prediction answer. To analyse the impact of these factors, we study a number of variants of a rule-based approach, which are specifically aimed at addressing the aforementioned issues. We find that the resulting models can achieve a performance which is close to that of NBFNet. Crucially, the considered variants only use a small fraction of the evidence that NBFNet relies on, which means that they largely keep the interpretability advantage of rule-based methods. Moreover, we show that a further variant, which does look at the full KG, consistently outperforms NBFNet." @default.
- W4385965402 created "2023-08-18" @default.
- W4385965402 creator A5004654520 @default.
- W4385965402 creator A5009887598 @default.
- W4385965402 creator A5027783013 @default.
- W4385965402 creator A5060245641 @default.
- W4385965402 date "2023-08-14" @default.
- W4385965402 modified "2023-10-17" @default.
- W4385965402 title "Inductive Knowledge Graph Completion with GNNs and Rules: An Analysis" @default.
- W4385965402 doi "https://doi.org/10.48550/arxiv.2308.07942" @default.
- W4385965402 hasPublicationYear "2023" @default.
- W4385965402 type Work @default.
- W4385965402 citedByCount "0" @default.
- W4385965402 crossrefType "posted-content" @default.
- W4385965402 hasAuthorship W4385965402A5004654520 @default.
- W4385965402 hasAuthorship W4385965402A5009887598 @default.
- W4385965402 hasAuthorship W4385965402A5027783013 @default.
- W4385965402 hasAuthorship W4385965402A5060245641 @default.
- W4385965402 hasBestOaLocation W43859654021 @default.
- W4385965402 hasConcept C119857082 @default.
- W4385965402 hasConcept C132525143 @default.
- W4385965402 hasConcept C154945302 @default.
- W4385965402 hasConcept C2776214188 @default.
- W4385965402 hasConcept C2781067378 @default.
- W4385965402 hasConcept C41008148 @default.
- W4385965402 hasConcept C80444323 @default.
- W4385965402 hasConceptScore W4385965402C119857082 @default.
- W4385965402 hasConceptScore W4385965402C132525143 @default.
- W4385965402 hasConceptScore W4385965402C154945302 @default.
- W4385965402 hasConceptScore W4385965402C2776214188 @default.
- W4385965402 hasConceptScore W4385965402C2781067378 @default.
- W4385965402 hasConceptScore W4385965402C41008148 @default.
- W4385965402 hasConceptScore W4385965402C80444323 @default.
- W4385965402 hasLocation W43859654021 @default.
- W4385965402 hasOpenAccess W4385965402 @default.
- W4385965402 hasPrimaryLocation W43859654021 @default.
- W4385965402 hasRelatedWork W1986582023 @default.
- W4385965402 hasRelatedWork W2355108509 @default.
- W4385965402 hasRelatedWork W3006943036 @default.
- W4385965402 hasRelatedWork W4200511449 @default.
- W4385965402 hasRelatedWork W4206534706 @default.
- W4385965402 hasRelatedWork W4229079080 @default.
- W4385965402 hasRelatedWork W4299487748 @default.
- W4385965402 hasRelatedWork W4385767940 @default.
- W4385965402 hasRelatedWork W4385957992 @default.
- W4385965402 hasRelatedWork W4385965371 @default.
- W4385965402 isParatext "false" @default.
- W4385965402 isRetracted "false" @default.
- W4385965402 workType "article" @default.