Matches in SemOpenAlex for { <https://semopenalex.org/work/W3214086726> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W3214086726 abstract "State-of-the-art Graph Neural Networks (GNNs) have limited scalability with respect to the graph and model sizes. On large graphs, increasing the model depth often means exponential expansion of the scope (i.e., receptive field). Beyond just a few layers, two fundamental challenges emerge: 1. degraded expressivity due to oversmoothing, and 2. expensive computation due to neighborhood explosion. We propose a design principle to decouple the depth and scope of GNNs -- to generate representation of a target entity (i.e., a node or an edge), we first extract a localized subgraph as the bounded-size scope, and then apply a GNN of arbitrary depth on top of the subgraph. A properly extracted subgraph consists of a small number of critical neighbors, while excluding irrelevant ones. The GNN, no matter how deep it is, smooths the local neighborhood into informative representation rather than oversmoothing the global graph into white noise. Theoretically, decoupling improves the GNN expressive power from the perspectives of graph signal processing (GCN), function approximation (GraphSAGE) and topological learning (GIN). Empirically, on seven graphs (with up to 110M nodes) and six backbone GNN architectures, our design achieves significant accuracy improvement with orders of magnitude reduction in computation and hardware cost." @default.
- W3214086726 created "2021-11-22" @default.
- W3214086726 creator A5002594174 @default.
- W3214086726 creator A5008479914 @default.
- W3214086726 creator A5014103790 @default.
- W3214086726 creator A5033086344 @default.
- W3214086726 creator A5042560222 @default.
- W3214086726 creator A5050729402 @default.
- W3214086726 creator A5052933431 @default.
- W3214086726 creator A5071515223 @default.
- W3214086726 creator A5085643676 @default.
- W3214086726 date "2022-01-19" @default.
- W3214086726 modified "2023-10-09" @default.
- W3214086726 title "Decoupling the Depth and Scope of Graph Neural Networks" @default.
- W3214086726 doi "https://doi.org/10.48550/arxiv.2201.07858" @default.
- W3214086726 hasPublicationYear "2022" @default.
- W3214086726 type Work @default.
- W3214086726 sameAs 3214086726 @default.
- W3214086726 citedByCount "1" @default.
- W3214086726 countsByYear W32140867262021 @default.
- W3214086726 crossrefType "posted-content" @default.
- W3214086726 hasAuthorship W3214086726A5002594174 @default.
- W3214086726 hasAuthorship W3214086726A5008479914 @default.
- W3214086726 hasAuthorship W3214086726A5014103790 @default.
- W3214086726 hasAuthorship W3214086726A5033086344 @default.
- W3214086726 hasAuthorship W3214086726A5042560222 @default.
- W3214086726 hasAuthorship W3214086726A5050729402 @default.
- W3214086726 hasAuthorship W3214086726A5052933431 @default.
- W3214086726 hasAuthorship W3214086726A5071515223 @default.
- W3214086726 hasAuthorship W3214086726A5085643676 @default.
- W3214086726 hasBestOaLocation W32140867261 @default.
- W3214086726 hasConcept C11413529 @default.
- W3214086726 hasConcept C114614502 @default.
- W3214086726 hasConcept C127413603 @default.
- W3214086726 hasConcept C132525143 @default.
- W3214086726 hasConcept C133731056 @default.
- W3214086726 hasConcept C134306372 @default.
- W3214086726 hasConcept C184720557 @default.
- W3214086726 hasConcept C199360897 @default.
- W3214086726 hasConcept C205606062 @default.
- W3214086726 hasConcept C2778012447 @default.
- W3214086726 hasConcept C33923547 @default.
- W3214086726 hasConcept C34388435 @default.
- W3214086726 hasConcept C41008148 @default.
- W3214086726 hasConcept C45374587 @default.
- W3214086726 hasConcept C48044578 @default.
- W3214086726 hasConcept C77088390 @default.
- W3214086726 hasConcept C80444323 @default.
- W3214086726 hasConceptScore W3214086726C11413529 @default.
- W3214086726 hasConceptScore W3214086726C114614502 @default.
- W3214086726 hasConceptScore W3214086726C127413603 @default.
- W3214086726 hasConceptScore W3214086726C132525143 @default.
- W3214086726 hasConceptScore W3214086726C133731056 @default.
- W3214086726 hasConceptScore W3214086726C134306372 @default.
- W3214086726 hasConceptScore W3214086726C184720557 @default.
- W3214086726 hasConceptScore W3214086726C199360897 @default.
- W3214086726 hasConceptScore W3214086726C205606062 @default.
- W3214086726 hasConceptScore W3214086726C2778012447 @default.
- W3214086726 hasConceptScore W3214086726C33923547 @default.
- W3214086726 hasConceptScore W3214086726C34388435 @default.
- W3214086726 hasConceptScore W3214086726C41008148 @default.
- W3214086726 hasConceptScore W3214086726C45374587 @default.
- W3214086726 hasConceptScore W3214086726C48044578 @default.
- W3214086726 hasConceptScore W3214086726C77088390 @default.
- W3214086726 hasConceptScore W3214086726C80444323 @default.
- W3214086726 hasLocation W32140867261 @default.
- W3214086726 hasOpenAccess W3214086726 @default.
- W3214086726 hasPrimaryLocation W32140867261 @default.
- W3214086726 hasRelatedWork W1568173680 @default.
- W3214086726 hasRelatedWork W2023563208 @default.
- W3214086726 hasRelatedWork W2155649197 @default.
- W3214086726 hasRelatedWork W2167499506 @default.
- W3214086726 hasRelatedWork W2375463041 @default.
- W3214086726 hasRelatedWork W2396430780 @default.
- W3214086726 hasRelatedWork W2915336237 @default.
- W3214086726 hasRelatedWork W4236163602 @default.
- W3214086726 hasRelatedWork W4288563611 @default.
- W3214086726 hasRelatedWork W2944165644 @default.
- W3214086726 isParatext "false" @default.
- W3214086726 isRetracted "false" @default.
- W3214086726 magId "3214086726" @default.
- W3214086726 workType "article" @default.