Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387596598> ?p ?o ?g. }
Showing items 1 to 69 of
69
with 100 items per page.
- W4387596598 abstract "We analyze the performance of graph neural network (GNN) architectures from the perspective of random graph theory. Our approach promises to complement existing lenses on GNN analysis, such as combinatorial expressive power and worst-case adversarial analysis, by connecting the performance of GNNs to typical-case properties of the training data. First, we theoretically characterize the nodewise accuracy of one- and two-layer GCNs relative to the contextual stochastic block model (cSBM) and related models. We additionally prove that GCNs cannot beat linear models under certain circumstances. Second, we numerically map the recoverability thresholds, in terms of accuracy, of four diverse GNN architectures (GCN, GAT, SAGE, and Graph Transformer) under a variety of assumptions about the data. Sample results of this second analysis include: heavy-tailed degree distributions enhance GNN performance, GNNs can work well on strongly heterophilous graphs, and SAGE and Graph Transformer can perform well on arbitrarily noisy edge data, but no architecture handled sufficiently noisy feature data well. Finally, we show how both specific higher-order structures in synthetic data and the mix of empirical structures in real data have dramatic effects (usually negative) on GNN performance." @default.
- W4387596598 created "2023-10-13" @default.
- W4387596598 creator A5006335922 @default.
- W4387596598 creator A5039807393 @default.
- W4387596598 creator A5044401499 @default.
- W4387596598 creator A5048835300 @default.
- W4387596598 creator A5050658708 @default.
- W4387596598 creator A5088606210 @default.
- W4387596598 creator A5090460464 @default.
- W4387596598 date "2023-10-11" @default.
- W4387596598 modified "2023-10-14" @default.
- W4387596598 title "Global Minima, Recoverability Thresholds, and Higher-Order Structure in GNNS" @default.
- W4387596598 doi "https://doi.org/10.48550/arxiv.2310.07667" @default.
- W4387596598 hasPublicationYear "2023" @default.
- W4387596598 type Work @default.
- W4387596598 citedByCount "0" @default.
- W4387596598 crossrefType "posted-content" @default.
- W4387596598 hasAuthorship W4387596598A5006335922 @default.
- W4387596598 hasAuthorship W4387596598A5039807393 @default.
- W4387596598 hasAuthorship W4387596598A5044401499 @default.
- W4387596598 hasAuthorship W4387596598A5048835300 @default.
- W4387596598 hasAuthorship W4387596598A5050658708 @default.
- W4387596598 hasAuthorship W4387596598A5088606210 @default.
- W4387596598 hasAuthorship W4387596598A5090460464 @default.
- W4387596598 hasBestOaLocation W43875965981 @default.
- W4387596598 hasConcept C11413529 @default.
- W4387596598 hasConcept C121332964 @default.
- W4387596598 hasConcept C124101348 @default.
- W4387596598 hasConcept C132525143 @default.
- W4387596598 hasConcept C134306372 @default.
- W4387596598 hasConcept C153180895 @default.
- W4387596598 hasConcept C154945302 @default.
- W4387596598 hasConcept C165801399 @default.
- W4387596598 hasConcept C186633575 @default.
- W4387596598 hasConcept C33923547 @default.
- W4387596598 hasConcept C41008148 @default.
- W4387596598 hasConcept C62520636 @default.
- W4387596598 hasConcept C66322947 @default.
- W4387596598 hasConcept C80444323 @default.
- W4387596598 hasConceptScore W4387596598C11413529 @default.
- W4387596598 hasConceptScore W4387596598C121332964 @default.
- W4387596598 hasConceptScore W4387596598C124101348 @default.
- W4387596598 hasConceptScore W4387596598C132525143 @default.
- W4387596598 hasConceptScore W4387596598C134306372 @default.
- W4387596598 hasConceptScore W4387596598C153180895 @default.
- W4387596598 hasConceptScore W4387596598C154945302 @default.
- W4387596598 hasConceptScore W4387596598C165801399 @default.
- W4387596598 hasConceptScore W4387596598C186633575 @default.
- W4387596598 hasConceptScore W4387596598C33923547 @default.
- W4387596598 hasConceptScore W4387596598C41008148 @default.
- W4387596598 hasConceptScore W4387596598C62520636 @default.
- W4387596598 hasConceptScore W4387596598C66322947 @default.
- W4387596598 hasConceptScore W4387596598C80444323 @default.
- W4387596598 hasLocation W43875965981 @default.
- W4387596598 hasOpenAccess W4387596598 @default.
- W4387596598 hasPrimaryLocation W43875965981 @default.
- W4387596598 hasRelatedWork W136674370 @default.
- W4387596598 hasRelatedWork W2050203848 @default.
- W4387596598 hasRelatedWork W2052387497 @default.
- W4387596598 hasRelatedWork W2105527480 @default.
- W4387596598 hasRelatedWork W2354676191 @default.
- W4387596598 hasRelatedWork W2375684291 @default.
- W4387596598 hasRelatedWork W2909957174 @default.
- W4387596598 hasRelatedWork W3188646203 @default.
- W4387596598 hasRelatedWork W4246278799 @default.
- W4387596598 hasRelatedWork W4320035781 @default.
- W4387596598 isParatext "false" @default.
- W4387596598 isRetracted "false" @default.
- W4387596598 workType "article" @default.