FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385567658> ?p ?o ?g. }

• W4385567658 abstract "While having been used widely for large-scale recommendation and online advertising, the Graph Neural Network (GNN) has demonstrated its representation learning capacity to extract embeddings of nodes and edges through passing, transforming, and aggregating information over the graph. In this work, we propose PGLBox1 - a multi-GPU graph learning framework based on PaddlePaddle [24], incorporating with optimized storage, computation, and communication strategies, to train deep GNNs based on web-scale graphs for the recommendation. Specifically, PGLBox adopts a hierarchical storage system with three layers to facilitate I/O, where graphs and embeddings are stored in the HBMs and SSDs, respectively, with MEMs as the cache. To fully utilize multi-GPUs and I/O bandwidth, PGLBox proposes an asynchronous pipeline with three stages - it first samples the subgraphs from the input graph, then pulls & updates embeddings and trains GNNs on the subgraph with parameters updating queued at the end of the pipeline. Thanks to the capacity of PGLBox in handling web-scale graphs, it becomes feasible to unify the view of GNN-based recommendation tasks for multiple advertising verticals and fuse all these graphs into a unified yet huge one. We evaluate PGLBox using a bucket of realistic GNN training tasks for the recommendation, and compare the performance of PGLBox on top of a multi-GPU server (Tesla A100×8) and the legacy training system based on a 40-node MPI cluster at Baidu. The overall comparisons show that PGLBox could save up to 55% monetary cost for training GNN models, and achieve up to 14× training speedup with the same accuracy as the legacy trainer. The open-source implementation of PGLBox is available at https://github.com/PaddlePaddle/PGL/tree/main/apps/PGLBox." @default.
• W4385567658 created "2023-08-05" @default.
• W4385567658 creator A5003318546 @default.
• W4385567658 creator A5004301827 @default.
• W4385567658 creator A5005049423 @default.
• W4385567658 creator A5006865469 @default.
• W4385567658 creator A5008582815 @default.
• W4385567658 creator A5020619158 @default.
• W4385567658 creator A5021438219 @default.
• W4385567658 creator A5028870975 @default.
• W4385567658 creator A5029434627 @default.
• W4385567658 creator A5031886056 @default.
• W4385567658 creator A5032113405 @default.
• W4385567658 creator A5034880019 @default.
• W4385567658 creator A5042074564 @default.
• W4385567658 creator A5081254155 @default.
• W4385567658 creator A5081440599 @default.
• W4385567658 creator A5083871467 @default.
• W4385567658 creator A5084155236 @default.
• W4385567658 creator A5086016146 @default.
• W4385567658 creator A5086810479 @default.
• W4385567658 date "2023-08-04" @default.
• W4385567658 modified "2023-09-27" @default.
• W4385567658 title "PGLBox: Multi-GPU Graph Learning Framework for Web-Scale Recommendation" @default.
• W4385567658 cites W1969282919 @default.
• W4385567658 cites W2016559894 @default.
• W4385567658 cites W2048793727 @default.
• W4385567658 cites W2053923031 @default.
• W4385567658 cites W2119885577 @default.
• W4385567658 cites W2523437372 @default.
• W4385567658 cites W2743104969 @default.
• W4385567658 cites W2914721378 @default.
• W4385567658 cites W2950960796 @default.
• W4385567658 cites W2964182926 @default.
• W4385567658 cites W2984020950 @default.
• W4385567658 cites W3002924435 @default.
• W4385567658 cites W3084512379 @default.
• W4385567658 cites W3100848837 @default.
• W4385567658 cites W3104097132 @default.
• W4385567658 cites W3159894882 @default.
• W4385567658 cites W4205347394 @default.
• W4385567658 doi "https://doi.org/10.1145/3580305.3599885" @default.
• W4385567658 hasPublicationYear "2023" @default.
• W4385567658 type Work @default.
• W4385567658 citedByCount "0" @default.
• W4385567658 crossrefType "proceedings-article" @default.
• W4385567658 hasAuthorship W4385567658A5003318546 @default.
• W4385567658 hasAuthorship W4385567658A5004301827 @default.
• W4385567658 hasAuthorship W4385567658A5005049423 @default.
• W4385567658 hasAuthorship W4385567658A5006865469 @default.
• W4385567658 hasAuthorship W4385567658A5008582815 @default.
• W4385567658 hasAuthorship W4385567658A5020619158 @default.
• W4385567658 hasAuthorship W4385567658A5021438219 @default.
• W4385567658 hasAuthorship W4385567658A5028870975 @default.
• W4385567658 hasAuthorship W4385567658A5029434627 @default.
• W4385567658 hasAuthorship W4385567658A5031886056 @default.
• W4385567658 hasAuthorship W4385567658A5032113405 @default.
• W4385567658 hasAuthorship W4385567658A5034880019 @default.
• W4385567658 hasAuthorship W4385567658A5042074564 @default.
• W4385567658 hasAuthorship W4385567658A5081254155 @default.
• W4385567658 hasAuthorship W4385567658A5081440599 @default.
• W4385567658 hasAuthorship W4385567658A5083871467 @default.
• W4385567658 hasAuthorship W4385567658A5084155236 @default.
• W4385567658 hasAuthorship W4385567658A5086016146 @default.
• W4385567658 hasAuthorship W4385567658A5086810479 @default.
• W4385567658 hasBestOaLocation W43855676581 @default.
• W4385567658 hasConcept C111919701 @default.
• W4385567658 hasConcept C115537543 @default.
• W4385567658 hasConcept C119857082 @default.
• W4385567658 hasConcept C132525143 @default.
• W4385567658 hasConcept C151319957 @default.
• W4385567658 hasConcept C173608175 @default.
• W4385567658 hasConcept C199360897 @default.
• W4385567658 hasConcept C2778692605 @default.
• W4385567658 hasConcept C31258907 @default.
• W4385567658 hasConcept C41008148 @default.
• W4385567658 hasConcept C43521106 @default.
• W4385567658 hasConcept C557471498 @default.
• W4385567658 hasConcept C68339613 @default.
• W4385567658 hasConcept C80444323 @default.
• W4385567658 hasConceptScore W4385567658C111919701 @default.
• W4385567658 hasConceptScore W4385567658C115537543 @default.
• W4385567658 hasConceptScore W4385567658C119857082 @default.
• W4385567658 hasConceptScore W4385567658C132525143 @default.
• W4385567658 hasConceptScore W4385567658C151319957 @default.
• W4385567658 hasConceptScore W4385567658C173608175 @default.
• W4385567658 hasConceptScore W4385567658C199360897 @default.
• W4385567658 hasConceptScore W4385567658C2778692605 @default.
• W4385567658 hasConceptScore W4385567658C31258907 @default.
• W4385567658 hasConceptScore W4385567658C41008148 @default.
• W4385567658 hasConceptScore W4385567658C43521106 @default.
• W4385567658 hasConceptScore W4385567658C557471498 @default.
• W4385567658 hasConceptScore W4385567658C68339613 @default.
• W4385567658 hasConceptScore W4385567658C80444323 @default.
• W4385567658 hasLocation W43855676581 @default.
• W4385567658 hasOpenAccess W4385567658 @default.
• W4385567658 hasPrimaryLocation W43855676581 @default.
• W4385567658 hasRelatedWork W1509211761 @default.
• W4385567658 hasRelatedWork W1531488649 @default.
• W4385567658 hasRelatedWork W2133693067 @default.

• next