Matches in SemOpenAlex for { <https://semopenalex.org/work/W3189231572> ?p ?o ?g. }
Showing items 1 to 94 of
94
with 100 items per page.
- W3189231572 abstract "Mining from graph-structured data is an integral component of graph data management. A recent trending technique, graph convolutional network (GCN), has gained momentum in the graph mining field, and plays an essential part in numerous graph-related tasks. Although the emerging GCN optimization techniques bring improvements to specific scenarios, they perform diversely in different applications and introduce many trial-and-error costs for practitioners. Moreover, existing GCN models often suffer from oversmoothing problem. Besides, the entanglement of various graph patterns could lead to non-robustness and harm the final performance of GCNs. In this work, we propose a simple yet efficient graph decomposition approach to improve the performance of general graph neural networks. We first empirically study existing graph decomposition methods and propose an automatic connectivity-ware graph decomposition algorithm, DeGNN. To provide a theoretical explanation, we then characterize GCN from the information-theoretic perspective and show that under certain conditions, the mutual information between the output after l layers and the input of GCN converges to 0 exponentially with respect to l. On the other hand, we show that graph decomposition can potentially weaken the condition of such convergence rate, alleviating the information loss when GCN becomes deeper. Extensive experiments on various academic benchmarks and real-world production datasets demonstrate that graph decomposition generally boosts the performance of GNN models. Moreover, our proposed solution DeGNN achieves state-of-the-art performances on almost all these tasks." @default.
- W3189231572 created "2021-08-16" @default.
- W3189231572 creator A5004750246 @default.
- W3189231572 creator A5005029405 @default.
- W3189231572 creator A5005191608 @default.
- W3189231572 creator A5007322337 @default.
- W3189231572 creator A5013718214 @default.
- W3189231572 creator A5014615052 @default.
- W3189231572 creator A5015552951 @default.
- W3189231572 creator A5016955253 @default.
- W3189231572 creator A5022714894 @default.
- W3189231572 creator A5022969296 @default.
- W3189231572 creator A5034686488 @default.
- W3189231572 creator A5034995105 @default.
- W3189231572 creator A5041619025 @default.
- W3189231572 creator A5061643764 @default.
- W3189231572 creator A5062357883 @default.
- W3189231572 creator A5064255264 @default.
- W3189231572 creator A5072823320 @default.
- W3189231572 creator A5074733691 @default.
- W3189231572 creator A5083848214 @default.
- W3189231572 creator A5086979603 @default.
- W3189231572 date "2021-08-14" @default.
- W3189231572 modified "2023-09-25" @default.
- W3189231572 title "DeGNN" @default.
- W3189231572 cites W2070232376 @default.
- W3189231572 cites W2123469175 @default.
- W3189231572 cites W2594083602 @default.
- W3189231572 cites W2612186437 @default.
- W3189231572 cites W2788284887 @default.
- W3189231572 cites W2945827377 @default.
- W3189231572 cites W2996320484 @default.
- W3189231572 cites W3012816161 @default.
- W3189231572 cites W3023884480 @default.
- W3189231572 cites W3024834561 @default.
- W3189231572 cites W3030204405 @default.
- W3189231572 cites W3090347762 @default.
- W3189231572 cites W3101434632 @default.
- W3189231572 doi "https://doi.org/10.1145/3447548.3467312" @default.
- W3189231572 hasPublicationYear "2021" @default.
- W3189231572 type Work @default.
- W3189231572 sameAs 3189231572 @default.
- W3189231572 citedByCount "11" @default.
- W3189231572 countsByYear W31892315722021 @default.
- W3189231572 countsByYear W31892315722022 @default.
- W3189231572 countsByYear W31892315722023 @default.
- W3189231572 crossrefType "proceedings-article" @default.
- W3189231572 hasAuthorship W3189231572A5004750246 @default.
- W3189231572 hasAuthorship W3189231572A5005029405 @default.
- W3189231572 hasAuthorship W3189231572A5005191608 @default.
- W3189231572 hasAuthorship W3189231572A5007322337 @default.
- W3189231572 hasAuthorship W3189231572A5013718214 @default.
- W3189231572 hasAuthorship W3189231572A5014615052 @default.
- W3189231572 hasAuthorship W3189231572A5015552951 @default.
- W3189231572 hasAuthorship W3189231572A5016955253 @default.
- W3189231572 hasAuthorship W3189231572A5022714894 @default.
- W3189231572 hasAuthorship W3189231572A5022969296 @default.
- W3189231572 hasAuthorship W3189231572A5034686488 @default.
- W3189231572 hasAuthorship W3189231572A5034995105 @default.
- W3189231572 hasAuthorship W3189231572A5041619025 @default.
- W3189231572 hasAuthorship W3189231572A5061643764 @default.
- W3189231572 hasAuthorship W3189231572A5062357883 @default.
- W3189231572 hasAuthorship W3189231572A5064255264 @default.
- W3189231572 hasAuthorship W3189231572A5072823320 @default.
- W3189231572 hasAuthorship W3189231572A5074733691 @default.
- W3189231572 hasAuthorship W3189231572A5083848214 @default.
- W3189231572 hasAuthorship W3189231572A5086979603 @default.
- W3189231572 hasConcept C11413529 @default.
- W3189231572 hasConcept C124101348 @default.
- W3189231572 hasConcept C132525143 @default.
- W3189231572 hasConcept C41008148 @default.
- W3189231572 hasConcept C80444323 @default.
- W3189231572 hasConceptScore W3189231572C11413529 @default.
- W3189231572 hasConceptScore W3189231572C124101348 @default.
- W3189231572 hasConceptScore W3189231572C132525143 @default.
- W3189231572 hasConceptScore W3189231572C41008148 @default.
- W3189231572 hasConceptScore W3189231572C80444323 @default.
- W3189231572 hasLocation W31892315721 @default.
- W3189231572 hasOpenAccess W3189231572 @default.
- W3189231572 hasPrimaryLocation W31892315721 @default.
- W3189231572 hasRelatedWork W2347219288 @default.
- W3189231572 hasRelatedWork W2348097614 @default.
- W3189231572 hasRelatedWork W2351491280 @default.
- W3189231572 hasRelatedWork W2354822586 @default.
- W3189231572 hasRelatedWork W2366221835 @default.
- W3189231572 hasRelatedWork W2371447506 @default.
- W3189231572 hasRelatedWork W2386767533 @default.
- W3189231572 hasRelatedWork W303980170 @default.
- W3189231572 hasRelatedWork W3149424243 @default.
- W3189231572 hasRelatedWork W4317655900 @default.
- W3189231572 isParatext "false" @default.
- W3189231572 isRetracted "false" @default.
- W3189231572 magId "3189231572" @default.
- W3189231572 workType "article" @default.