Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287558196> ?p ?o ?g. }
Showing items 1 to 70 of
70
with 100 items per page.
- W4287558196 abstract "While Graph Neural Networks (GNNs) are powerful models for learning representations on graphs, most state-of-the-art models do not have significant accuracy gain beyond two to three layers. Deep GNNs fundamentally need to address: 1). expressivity challenge due to oversmoothing, and 2). computation challenge due to neighborhood explosion. We propose a simple deep GNN, shallow sampler design principle to improve both the GNN accuracy and efficiency -- to generate representation of a target node, we use a deep GNN to pass messages only within a shallow, localized subgraph. A properly sampled subgraph may exclude irrelevant or even noisy nodes, and still preserve the critical neighbor features and graph structures. The deep GNN then smooths the informative local signals to enhance feature learning, rather than oversmoothing the global graph signals into just white noise. We theoretically justify why the combination of deep GNNs with shallow samplers yields the best learning performance. We then propose various sampling algorithms and neural architecture extensions to achieve good empirical results. On the largest public graph dataset, ogbn-papers100M, we achieve state-of-the-art accuracy with an order of magnitude reduction in hardware cost." @default.
- W4287558196 created "2022-07-25" @default.
- W4287558196 creator A5002594174 @default.
- W4287558196 creator A5008479914 @default.
- W4287558196 creator A5033086344 @default.
- W4287558196 creator A5033166029 @default.
- W4287558196 creator A5042560222 @default.
- W4287558196 creator A5050729402 @default.
- W4287558196 creator A5050965079 @default.
- W4287558196 creator A5052933431 @default.
- W4287558196 creator A5071515223 @default.
- W4287558196 date "2020-12-02" @default.
- W4287558196 modified "2023-10-04" @default.
- W4287558196 title "Deep Graph Neural Networks with Shallow Subgraph Samplers" @default.
- W4287558196 hasPublicationYear "2020" @default.
- W4287558196 type Work @default.
- W4287558196 citedByCount "0" @default.
- W4287558196 crossrefType "posted-content" @default.
- W4287558196 hasAuthorship W4287558196A5002594174 @default.
- W4287558196 hasAuthorship W4287558196A5008479914 @default.
- W4287558196 hasAuthorship W4287558196A5033086344 @default.
- W4287558196 hasAuthorship W4287558196A5033166029 @default.
- W4287558196 hasAuthorship W4287558196A5042560222 @default.
- W4287558196 hasAuthorship W4287558196A5050729402 @default.
- W4287558196 hasAuthorship W4287558196A5050965079 @default.
- W4287558196 hasAuthorship W4287558196A5052933431 @default.
- W4287558196 hasAuthorship W4287558196A5071515223 @default.
- W4287558196 hasBestOaLocation W42875581961 @default.
- W4287558196 hasConcept C108583219 @default.
- W4287558196 hasConcept C11413529 @default.
- W4287558196 hasConcept C132525143 @default.
- W4287558196 hasConcept C138885662 @default.
- W4287558196 hasConcept C153180895 @default.
- W4287558196 hasConcept C154945302 @default.
- W4287558196 hasConcept C2776401178 @default.
- W4287558196 hasConcept C2984842247 @default.
- W4287558196 hasConcept C41008148 @default.
- W4287558196 hasConcept C41895202 @default.
- W4287558196 hasConcept C45374587 @default.
- W4287558196 hasConcept C50644808 @default.
- W4287558196 hasConcept C80444323 @default.
- W4287558196 hasConceptScore W4287558196C108583219 @default.
- W4287558196 hasConceptScore W4287558196C11413529 @default.
- W4287558196 hasConceptScore W4287558196C132525143 @default.
- W4287558196 hasConceptScore W4287558196C138885662 @default.
- W4287558196 hasConceptScore W4287558196C153180895 @default.
- W4287558196 hasConceptScore W4287558196C154945302 @default.
- W4287558196 hasConceptScore W4287558196C2776401178 @default.
- W4287558196 hasConceptScore W4287558196C2984842247 @default.
- W4287558196 hasConceptScore W4287558196C41008148 @default.
- W4287558196 hasConceptScore W4287558196C41895202 @default.
- W4287558196 hasConceptScore W4287558196C45374587 @default.
- W4287558196 hasConceptScore W4287558196C50644808 @default.
- W4287558196 hasConceptScore W4287558196C80444323 @default.
- W4287558196 hasLocation W42875581961 @default.
- W4287558196 hasOpenAccess W4287558196 @default.
- W4287558196 hasPrimaryLocation W42875581961 @default.
- W4287558196 hasRelatedWork W10116868 @default.
- W4287558196 hasRelatedWork W12563130 @default.
- W4287558196 hasRelatedWork W4919037 @default.
- W4287558196 hasRelatedWork W5374421 @default.
- W4287558196 hasRelatedWork W5738750 @default.
- W4287558196 hasRelatedWork W7497093 @default.
- W4287558196 hasRelatedWork W9043603 @default.
- W4287558196 hasRelatedWork W9190101 @default.
- W4287558196 hasRelatedWork W9362070 @default.
- W4287558196 hasRelatedWork W9770290 @default.
- W4287558196 isParatext "false" @default.
- W4287558196 isRetracted "false" @default.
- W4287558196 workType "article" @default.