Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287121301> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4287121301 abstract "We present GNNAutoScale (GAS), a framework for scaling arbitrary message-passing GNNs to large graphs. GAS prunes entire sub-trees of the computation graph by utilizing historical embeddings from prior training iterations, leading to constant GPU memory consumption in respect to input node size without dropping any data. While existing solutions weaken the expressive power of message passing due to sub-sampling of edges or non-trainable propagations, our approach is provably able to maintain the expressive power of the original GNN. We achieve this by providing approximation error bounds of historical embeddings and show how to tighten them in practice. Empirically, we show that the practical realization of our framework, PyGAS, an easy-to-use extension for PyTorch Geometric, is both fast and memory-efficient, learns expressive node representations, closely resembles the performance of their non-scaling counterparts, and reaches state-of-the-art performance on large-scale graphs." @default.
- W4287121301 created "2022-07-25" @default.
- W4287121301 creator A5002742256 @default.
- W4287121301 creator A5004108638 @default.
- W4287121301 creator A5073462022 @default.
- W4287121301 creator A5091272738 @default.
- W4287121301 date "2021-06-10" @default.
- W4287121301 modified "2023-10-05" @default.
- W4287121301 title "GNNAutoScale: Scalable and Expressive Graph Neural Networks via Historical Embeddings" @default.
- W4287121301 doi "https://doi.org/10.48550/arxiv.2106.05609" @default.
- W4287121301 hasPublicationYear "2021" @default.
- W4287121301 type Work @default.
- W4287121301 citedByCount "0" @default.
- W4287121301 crossrefType "posted-content" @default.
- W4287121301 hasAuthorship W4287121301A5002742256 @default.
- W4287121301 hasAuthorship W4287121301A5004108638 @default.
- W4287121301 hasAuthorship W4287121301A5073462022 @default.
- W4287121301 hasAuthorship W4287121301A5091272738 @default.
- W4287121301 hasBestOaLocation W42871213011 @default.
- W4287121301 hasConcept C105795698 @default.
- W4287121301 hasConcept C11413529 @default.
- W4287121301 hasConcept C127413603 @default.
- W4287121301 hasConcept C132525143 @default.
- W4287121301 hasConcept C195818886 @default.
- W4287121301 hasConcept C2524010 @default.
- W4287121301 hasConcept C2781089630 @default.
- W4287121301 hasConcept C33923547 @default.
- W4287121301 hasConcept C41008148 @default.
- W4287121301 hasConcept C45374587 @default.
- W4287121301 hasConcept C48044578 @default.
- W4287121301 hasConcept C62611344 @default.
- W4287121301 hasConcept C66938386 @default.
- W4287121301 hasConcept C77088390 @default.
- W4287121301 hasConcept C80444323 @default.
- W4287121301 hasConcept C99844830 @default.
- W4287121301 hasConceptScore W4287121301C105795698 @default.
- W4287121301 hasConceptScore W4287121301C11413529 @default.
- W4287121301 hasConceptScore W4287121301C127413603 @default.
- W4287121301 hasConceptScore W4287121301C132525143 @default.
- W4287121301 hasConceptScore W4287121301C195818886 @default.
- W4287121301 hasConceptScore W4287121301C2524010 @default.
- W4287121301 hasConceptScore W4287121301C2781089630 @default.
- W4287121301 hasConceptScore W4287121301C33923547 @default.
- W4287121301 hasConceptScore W4287121301C41008148 @default.
- W4287121301 hasConceptScore W4287121301C45374587 @default.
- W4287121301 hasConceptScore W4287121301C48044578 @default.
- W4287121301 hasConceptScore W4287121301C62611344 @default.
- W4287121301 hasConceptScore W4287121301C66938386 @default.
- W4287121301 hasConceptScore W4287121301C77088390 @default.
- W4287121301 hasConceptScore W4287121301C80444323 @default.
- W4287121301 hasConceptScore W4287121301C99844830 @default.
- W4287121301 hasLocation W42871213011 @default.
- W4287121301 hasOpenAccess W4287121301 @default.
- W4287121301 hasPrimaryLocation W42871213011 @default.
- W4287121301 hasRelatedWork W2022544890 @default.
- W4287121301 hasRelatedWork W2043523297 @default.
- W4287121301 hasRelatedWork W2113405914 @default.
- W4287121301 hasRelatedWork W2260963831 @default.
- W4287121301 hasRelatedWork W2331305369 @default.
- W4287121301 hasRelatedWork W2394097730 @default.
- W4287121301 hasRelatedWork W2734546921 @default.
- W4287121301 hasRelatedWork W2769996735 @default.
- W4287121301 hasRelatedWork W2950042906 @default.
- W4287121301 hasRelatedWork W4235249401 @default.
- W4287121301 isParatext "false" @default.
- W4287121301 isRetracted "false" @default.
- W4287121301 workType "article" @default.