Matches in SemOpenAlex for { <https://semopenalex.org/work/W4283766095> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4283766095 abstract "Due to the over-smoothing issue, most existing graph neural networks can only capture limited dependencies with their inherently finite aggregation layers. To overcome this limitation, we propose a new kind of graph convolution, called Graph Implicit Nonlinear Diffusion (GIND), which implicitly has access to infinite hops of neighbors while adaptively aggregating features with nonlinear diffusion to prevent over-smoothing. Notably, we show that the learned representation can be formalized as the minimizer of an explicit convex optimization objective. With this property, we can theoretically characterize the equilibrium of our GIND from an optimization perspective. More interestingly, we can induce new structural variants by modifying the corresponding optimization objective. To be specific, we can embed prior properties to the equilibrium, as well as introducing skip connections to promote training stability. Extensive experiments show that GIND is good at capturing long-range dependencies, and performs well on both homophilic and heterophilic graphs with nonlinear diffusion. Moreover, we show that the optimization-induced variants of our models can boost the performance and improve training stability and efficiency as well. As a result, our GIND obtains significant improvements on both node-level and graph-level tasks." @default.
- W4283766095 created "2022-07-02" @default.
- W4283766095 creator A5004999983 @default.
- W4283766095 creator A5014708668 @default.
- W4283766095 creator A5016399094 @default.
- W4283766095 creator A5017320345 @default.
- W4283766095 creator A5029941351 @default.
- W4283766095 date "2022-06-29" @default.
- W4283766095 modified "2023-10-01" @default.
- W4283766095 title "Optimization-Induced Graph Implicit Nonlinear Diffusion" @default.
- W4283766095 doi "https://doi.org/10.48550/arxiv.2206.14418" @default.
- W4283766095 hasPublicationYear "2022" @default.
- W4283766095 type Work @default.
- W4283766095 citedByCount "0" @default.
- W4283766095 crossrefType "posted-content" @default.
- W4283766095 hasAuthorship W4283766095A5004999983 @default.
- W4283766095 hasAuthorship W4283766095A5014708668 @default.
- W4283766095 hasAuthorship W4283766095A5016399094 @default.
- W4283766095 hasAuthorship W4283766095A5017320345 @default.
- W4283766095 hasAuthorship W4283766095A5029941351 @default.
- W4283766095 hasBestOaLocation W42837660951 @default.
- W4283766095 hasConcept C111472728 @default.
- W4283766095 hasConcept C112972136 @default.
- W4283766095 hasConcept C11413529 @default.
- W4283766095 hasConcept C119857082 @default.
- W4283766095 hasConcept C121332964 @default.
- W4283766095 hasConcept C126255220 @default.
- W4283766095 hasConcept C132525143 @default.
- W4283766095 hasConcept C138885662 @default.
- W4283766095 hasConcept C158622935 @default.
- W4283766095 hasConcept C159985019 @default.
- W4283766095 hasConcept C189950617 @default.
- W4283766095 hasConcept C192562407 @default.
- W4283766095 hasConcept C204323151 @default.
- W4283766095 hasConcept C31972630 @default.
- W4283766095 hasConcept C33923547 @default.
- W4283766095 hasConcept C3770464 @default.
- W4283766095 hasConcept C41008148 @default.
- W4283766095 hasConcept C62520636 @default.
- W4283766095 hasConcept C80444323 @default.
- W4283766095 hasConceptScore W4283766095C111472728 @default.
- W4283766095 hasConceptScore W4283766095C112972136 @default.
- W4283766095 hasConceptScore W4283766095C11413529 @default.
- W4283766095 hasConceptScore W4283766095C119857082 @default.
- W4283766095 hasConceptScore W4283766095C121332964 @default.
- W4283766095 hasConceptScore W4283766095C126255220 @default.
- W4283766095 hasConceptScore W4283766095C132525143 @default.
- W4283766095 hasConceptScore W4283766095C138885662 @default.
- W4283766095 hasConceptScore W4283766095C158622935 @default.
- W4283766095 hasConceptScore W4283766095C159985019 @default.
- W4283766095 hasConceptScore W4283766095C189950617 @default.
- W4283766095 hasConceptScore W4283766095C192562407 @default.
- W4283766095 hasConceptScore W4283766095C204323151 @default.
- W4283766095 hasConceptScore W4283766095C31972630 @default.
- W4283766095 hasConceptScore W4283766095C33923547 @default.
- W4283766095 hasConceptScore W4283766095C3770464 @default.
- W4283766095 hasConceptScore W4283766095C41008148 @default.
- W4283766095 hasConceptScore W4283766095C62520636 @default.
- W4283766095 hasConceptScore W4283766095C80444323 @default.
- W4283766095 hasLocation W42837660951 @default.
- W4283766095 hasOpenAccess W4283766095 @default.
- W4283766095 hasPrimaryLocation W42837660951 @default.
- W4283766095 hasRelatedWork W2090191433 @default.
- W4283766095 hasRelatedWork W2351491280 @default.
- W4283766095 hasRelatedWork W2359407220 @default.
- W4283766095 hasRelatedWork W2386767533 @default.
- W4283766095 hasRelatedWork W2389896045 @default.
- W4283766095 hasRelatedWork W3005990459 @default.
- W4283766095 hasRelatedWork W303980170 @default.
- W4283766095 hasRelatedWork W3168493052 @default.
- W4283766095 hasRelatedWork W4225891298 @default.
- W4283766095 hasRelatedWork W4309795810 @default.
- W4283766095 isParatext "false" @default.
- W4283766095 isRetracted "false" @default.
- W4283766095 workType "article" @default.