FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287704986> ?p ?o ?g. }

• W4287704986 abstract "Federated learning has emerged recently as a promising solution for distributing machine learning tasks through modern networks of mobile devices. Recent studies have obtained lower bounds on the expected decrease in model loss that is achieved through each round of federated learning. However, convergence generally requires a large number of communication rounds, which induces delay in model training and is costly in terms of network resources. In this paper, we propose a fast-convergent federated learning algorithm, called FOLB, which performs intelligent sampling of devices in each round of model training to optimize the expected convergence speed. We first theoretically characterize a lower bound on improvement that can be obtained in each round if devices are selected according to the expected improvement their local models will provide to the current global model. Then, we show that FOLB obtains this bound through uniform sampling by weighting device updates according to their gradient information. FOLB is able to handle both communication and computation heterogeneity of devices by adapting the aggregations according to estimates of device's capabilities of contributing to the updates. We evaluate FOLB in comparison with existing federated learning algorithms and experimentally show its improvement in trained model accuracy, convergence speed, and/or model stability across various machine learning tasks and datasets." @default.
• W4287704986 created "2022-07-26" @default.
• W4287704986 creator A5011437254 @default.
• W4287704986 creator A5020399355 @default.
• W4287704986 creator A5042307561 @default.
• W4287704986 creator A5052709754 @default.
• W4287704986 creator A5057294414 @default.
• W4287704986 creator A5074403000 @default.
• W4287704986 date "2020-07-26" @default.
• W4287704986 modified "2023-10-18" @default.
• W4287704986 title "Fast-Convergent Federated Learning" @default.
• W4287704986 doi "https://doi.org/10.48550/arxiv.2007.13137" @default.
• W4287704986 hasPublicationYear "2020" @default.
• W4287704986 type Work @default.
• W4287704986 citedByCount "0" @default.
• W4287704986 crossrefType "posted-content" @default.
• W4287704986 hasAuthorship W4287704986A5011437254 @default.
• W4287704986 hasAuthorship W4287704986A5020399355 @default.
• W4287704986 hasAuthorship W4287704986A5042307561 @default.
• W4287704986 hasAuthorship W4287704986A5052709754 @default.
• W4287704986 hasAuthorship W4287704986A5057294414 @default.
• W4287704986 hasAuthorship W4287704986A5074403000 @default.
• W4287704986 hasBestOaLocation W42877049861 @default.
• W4287704986 hasConcept C111919701 @default.
• W4287704986 hasConcept C112972136 @default.
• W4287704986 hasConcept C11413529 @default.
• W4287704986 hasConcept C119857082 @default.
• W4287704986 hasConcept C120314980 @default.
• W4287704986 hasConcept C126838900 @default.
• W4287704986 hasConcept C134306372 @default.
• W4287704986 hasConcept C140779682 @default.
• W4287704986 hasConcept C154945302 @default.
• W4287704986 hasConcept C162324750 @default.
• W4287704986 hasConcept C183115368 @default.
• W4287704986 hasConcept C186967261 @default.
• W4287704986 hasConcept C2777303404 @default.
• W4287704986 hasConcept C33923547 @default.
• W4287704986 hasConcept C41008148 @default.
• W4287704986 hasConcept C45374587 @default.
• W4287704986 hasConcept C50522688 @default.
• W4287704986 hasConcept C71924100 @default.
• W4287704986 hasConcept C76155785 @default.
• W4287704986 hasConcept C77553402 @default.
• W4287704986 hasConcept C94915269 @default.
• W4287704986 hasConceptScore W4287704986C111919701 @default.
• W4287704986 hasConceptScore W4287704986C112972136 @default.
• W4287704986 hasConceptScore W4287704986C11413529 @default.
• W4287704986 hasConceptScore W4287704986C119857082 @default.
• W4287704986 hasConceptScore W4287704986C120314980 @default.
• W4287704986 hasConceptScore W4287704986C126838900 @default.
• W4287704986 hasConceptScore W4287704986C134306372 @default.
• W4287704986 hasConceptScore W4287704986C140779682 @default.
• W4287704986 hasConceptScore W4287704986C154945302 @default.
• W4287704986 hasConceptScore W4287704986C162324750 @default.
• W4287704986 hasConceptScore W4287704986C183115368 @default.
• W4287704986 hasConceptScore W4287704986C186967261 @default.
• W4287704986 hasConceptScore W4287704986C2777303404 @default.
• W4287704986 hasConceptScore W4287704986C33923547 @default.
• W4287704986 hasConceptScore W4287704986C41008148 @default.
• W4287704986 hasConceptScore W4287704986C45374587 @default.
• W4287704986 hasConceptScore W4287704986C50522688 @default.
• W4287704986 hasConceptScore W4287704986C71924100 @default.
• W4287704986 hasConceptScore W4287704986C76155785 @default.
• W4287704986 hasConceptScore W4287704986C77553402 @default.
• W4287704986 hasConceptScore W4287704986C94915269 @default.
• W4287704986 hasLocation W42877049861 @default.
• W4287704986 hasOpenAccess W4287704986 @default.
• W4287704986 hasPrimaryLocation W42877049861 @default.
• W4287704986 hasRelatedWork W1577085395 @default.
• W4287704986 hasRelatedWork W1991928096 @default.
• W4287704986 hasRelatedWork W2035435054 @default.
• W4287704986 hasRelatedWork W2100179986 @default.
• W4287704986 hasRelatedWork W2137675151 @default.
• W4287704986 hasRelatedWork W2182183508 @default.
• W4287704986 hasRelatedWork W2241340767 @default.
• W4287704986 hasRelatedWork W2363892110 @default.
• W4287704986 hasRelatedWork W2371614152 @default.
• W4287704986 hasRelatedWork W2381210555 @default.
• W4287704986 isParatext "false" @default.
• W4287704986 isRetracted "false" @default.
• W4287704986 workType "article" @default.