Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387323047> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W4387323047 abstract "Federated learning (FL) is a popular paradigm for private and collaborative model training on the edge. In centralized FL, the parameters of a global architecture (such as a deep neural network) are maintained and distributed by a central server/controller to clients who transmit model updates (gradients) back to the server based on local optimization. While many efforts have focused on reducing the communication complexity of gradient transmission, the vast majority of compression-based algorithms assume that each participating client is able to download and train the current and full set of parameters, which may not be a practical assumption depending on the resource constraints of smaller clients such as mobile devices. In this work, we propose a simple yet effective novel algorithm, Comfetch, which allows clients to train large networks using reduced representations of the global architecture via the count sketch, which reduces local computational and memory costs along with bi-directional communication complexity. We provide a nonconvex convergence guarantee and experimentally demonstrate that it is possible to learn large models, such as a deep convolutional network, through federated training on their sketched counterparts. The resulting global models exhibit competitive test accuracy over CIFAR10/100 classification when compared against un-compressed model training." @default.
- W4387323047 created "2023-10-04" @default.
- W4387323047 creator A5008905506 @default.
- W4387323047 creator A5016163266 @default.
- W4387323047 creator A5036382130 @default.
- W4387323047 creator A5037162168 @default.
- W4387323047 creator A5060725247 @default.
- W4387323047 creator A5066702474 @default.
- W4387323047 creator A5082528804 @default.
- W4387323047 creator A5086440196 @default.
- W4387323047 date "2021-09-17" @default.
- W4387323047 modified "2023-10-05" @default.
- W4387323047 title "Comfetch: Federated Learning of Large Networks on Constrained Clients via Sketching" @default.
- W4387323047 doi "https://doi.org/10.48550/arxiv.2109.08346" @default.
- W4387323047 hasPublicationYear "2021" @default.
- W4387323047 type Work @default.
- W4387323047 citedByCount "0" @default.
- W4387323047 crossrefType "posted-content" @default.
- W4387323047 hasAuthorship W4387323047A5008905506 @default.
- W4387323047 hasAuthorship W4387323047A5016163266 @default.
- W4387323047 hasAuthorship W4387323047A5036382130 @default.
- W4387323047 hasAuthorship W4387323047A5037162168 @default.
- W4387323047 hasAuthorship W4387323047A5060725247 @default.
- W4387323047 hasAuthorship W4387323047A5066702474 @default.
- W4387323047 hasAuthorship W4387323047A5082528804 @default.
- W4387323047 hasAuthorship W4387323047A5086440196 @default.
- W4387323047 hasBestOaLocation W43873230471 @default.
- W4387323047 hasConcept C111472728 @default.
- W4387323047 hasConcept C111919701 @default.
- W4387323047 hasConcept C11413529 @default.
- W4387323047 hasConcept C120314980 @default.
- W4387323047 hasConcept C123657996 @default.
- W4387323047 hasConcept C138236772 @default.
- W4387323047 hasConcept C138885662 @default.
- W4387323047 hasConcept C142362112 @default.
- W4387323047 hasConcept C153349607 @default.
- W4387323047 hasConcept C154945302 @default.
- W4387323047 hasConcept C162307627 @default.
- W4387323047 hasConcept C162324750 @default.
- W4387323047 hasConcept C177264268 @default.
- W4387323047 hasConcept C199360897 @default.
- W4387323047 hasConcept C203479927 @default.
- W4387323047 hasConcept C2777303404 @default.
- W4387323047 hasConcept C2779231336 @default.
- W4387323047 hasConcept C2780586882 @default.
- W4387323047 hasConcept C41008148 @default.
- W4387323047 hasConcept C50522688 @default.
- W4387323047 hasConcept C6557445 @default.
- W4387323047 hasConcept C79974875 @default.
- W4387323047 hasConcept C81363708 @default.
- W4387323047 hasConcept C86803240 @default.
- W4387323047 hasConceptScore W4387323047C111472728 @default.
- W4387323047 hasConceptScore W4387323047C111919701 @default.
- W4387323047 hasConceptScore W4387323047C11413529 @default.
- W4387323047 hasConceptScore W4387323047C120314980 @default.
- W4387323047 hasConceptScore W4387323047C123657996 @default.
- W4387323047 hasConceptScore W4387323047C138236772 @default.
- W4387323047 hasConceptScore W4387323047C138885662 @default.
- W4387323047 hasConceptScore W4387323047C142362112 @default.
- W4387323047 hasConceptScore W4387323047C153349607 @default.
- W4387323047 hasConceptScore W4387323047C154945302 @default.
- W4387323047 hasConceptScore W4387323047C162307627 @default.
- W4387323047 hasConceptScore W4387323047C162324750 @default.
- W4387323047 hasConceptScore W4387323047C177264268 @default.
- W4387323047 hasConceptScore W4387323047C199360897 @default.
- W4387323047 hasConceptScore W4387323047C203479927 @default.
- W4387323047 hasConceptScore W4387323047C2777303404 @default.
- W4387323047 hasConceptScore W4387323047C2779231336 @default.
- W4387323047 hasConceptScore W4387323047C2780586882 @default.
- W4387323047 hasConceptScore W4387323047C41008148 @default.
- W4387323047 hasConceptScore W4387323047C50522688 @default.
- W4387323047 hasConceptScore W4387323047C6557445 @default.
- W4387323047 hasConceptScore W4387323047C79974875 @default.
- W4387323047 hasConceptScore W4387323047C81363708 @default.
- W4387323047 hasConceptScore W4387323047C86803240 @default.
- W4387323047 hasLocation W43873230471 @default.
- W4387323047 hasOpenAccess W4387323047 @default.
- W4387323047 hasPrimaryLocation W43873230471 @default.
- W4387323047 hasRelatedWork W2885461866 @default.
- W4387323047 hasRelatedWork W2918879532 @default.
- W4387323047 hasRelatedWork W3113627641 @default.
- W4387323047 hasRelatedWork W3176734149 @default.
- W4387323047 hasRelatedWork W3201779876 @default.
- W4387323047 hasRelatedWork W3208911535 @default.
- W4387323047 hasRelatedWork W4238142035 @default.
- W4387323047 hasRelatedWork W4313339048 @default.
- W4387323047 hasRelatedWork W4321606826 @default.
- W4387323047 hasRelatedWork W4386004629 @default.
- W4387323047 isParatext "false" @default.
- W4387323047 isRetracted "false" @default.
- W4387323047 workType "article" @default.