Matches in SemOpenAlex for { <https://semopenalex.org/work/W4319069175> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4319069175 endingPage "18" @default.
- W4319069175 startingPage "1" @default.
- W4319069175 abstract "Federated learning allows edge devices to collaboratively train a global model without sharing their local private data. Yet, with limited network bandwidth at the edge, communication often becomes a severe bottleneck. In this paper, we find that it is unnecessary to always synchronize the full model in the entire training process, because many parameters already become mature (i.e., stable) prior to model convergence, and can thus be excluded from later synchronizations. This allows us to reduce the communication overhead without compromising the model accuracy. However, challenges are that the local parameters excluded from global synchronization may diverge on different clients, and meanwhile some parameters may stabilize only temporally. To address these challenges, we propose a novel scheme called Adaptive Parameter Freezing (APF), which fixes (freezes) the non-synchronized stable parameters in intermittent periods. Specifically, the freezing periods are tentatively adjusted in an additively-increase and multiplicatively-decrease manner—depending on whether the previously-frozen parameters remain stable in subsequent iterations. We also extend APF into APF# and APF++, which freeze parameters in a more aggressive manner to achieve larger performance benefit for large complex models. We implemented APF and its variants as Python modules with PyTorch, and extensive experiments show that APF can reduce data transfer amount by over 60%." @default.
- W4319069175 created "2023-02-04" @default.
- W4319069175 creator A5000042335 @default.
- W4319069175 creator A5031383697 @default.
- W4319069175 creator A5037989560 @default.
- W4319069175 creator A5039003558 @default.
- W4319069175 creator A5062210904 @default.
- W4319069175 creator A5067697790 @default.
- W4319069175 creator A5089736472 @default.
- W4319069175 date "2023-01-01" @default.
- W4319069175 modified "2023-10-02" @default.
- W4319069175 title "Synchronize Only the Immature Parameters: Communication-Efficient Federated Learning By Freezing Parameters Adaptively" @default.
- W4319069175 doi "https://doi.org/10.1109/tpds.2023.3241965" @default.
- W4319069175 hasPublicationYear "2023" @default.
- W4319069175 type Work @default.
- W4319069175 citedByCount "1" @default.
- W4319069175 countsByYear W43190691752023 @default.
- W4319069175 crossrefType "journal-article" @default.
- W4319069175 hasAuthorship W4319069175A5000042335 @default.
- W4319069175 hasAuthorship W4319069175A5031383697 @default.
- W4319069175 hasAuthorship W4319069175A5037989560 @default.
- W4319069175 hasAuthorship W4319069175A5039003558 @default.
- W4319069175 hasAuthorship W4319069175A5062210904 @default.
- W4319069175 hasAuthorship W4319069175A5067697790 @default.
- W4319069175 hasAuthorship W4319069175A5089736472 @default.
- W4319069175 hasConcept C111919701 @default.
- W4319069175 hasConcept C120314980 @default.
- W4319069175 hasConcept C127162648 @default.
- W4319069175 hasConcept C138236772 @default.
- W4319069175 hasConcept C149635348 @default.
- W4319069175 hasConcept C154945302 @default.
- W4319069175 hasConcept C162307627 @default.
- W4319069175 hasConcept C2776257435 @default.
- W4319069175 hasConcept C2778562939 @default.
- W4319069175 hasConcept C2779960059 @default.
- W4319069175 hasConcept C2780513914 @default.
- W4319069175 hasConcept C31258907 @default.
- W4319069175 hasConcept C41008148 @default.
- W4319069175 hasConcept C519991488 @default.
- W4319069175 hasConcept C79974875 @default.
- W4319069175 hasConcept C98045186 @default.
- W4319069175 hasConceptScore W4319069175C111919701 @default.
- W4319069175 hasConceptScore W4319069175C120314980 @default.
- W4319069175 hasConceptScore W4319069175C127162648 @default.
- W4319069175 hasConceptScore W4319069175C138236772 @default.
- W4319069175 hasConceptScore W4319069175C149635348 @default.
- W4319069175 hasConceptScore W4319069175C154945302 @default.
- W4319069175 hasConceptScore W4319069175C162307627 @default.
- W4319069175 hasConceptScore W4319069175C2776257435 @default.
- W4319069175 hasConceptScore W4319069175C2778562939 @default.
- W4319069175 hasConceptScore W4319069175C2779960059 @default.
- W4319069175 hasConceptScore W4319069175C2780513914 @default.
- W4319069175 hasConceptScore W4319069175C31258907 @default.
- W4319069175 hasConceptScore W4319069175C41008148 @default.
- W4319069175 hasConceptScore W4319069175C519991488 @default.
- W4319069175 hasConceptScore W4319069175C79974875 @default.
- W4319069175 hasConceptScore W4319069175C98045186 @default.
- W4319069175 hasLocation W43190691751 @default.
- W4319069175 hasOpenAccess W4319069175 @default.
- W4319069175 hasPrimaryLocation W43190691751 @default.
- W4319069175 hasRelatedWork W1597592026 @default.
- W4319069175 hasRelatedWork W1990379658 @default.
- W4319069175 hasRelatedWork W1998323989 @default.
- W4319069175 hasRelatedWork W2124870959 @default.
- W4319069175 hasRelatedWork W2131630752 @default.
- W4319069175 hasRelatedWork W2157013742 @default.
- W4319069175 hasRelatedWork W2350861899 @default.
- W4319069175 hasRelatedWork W2391167130 @default.
- W4319069175 hasRelatedWork W4320067866 @default.
- W4319069175 hasRelatedWork W94000989 @default.
- W4319069175 isParatext "false" @default.
- W4319069175 isRetracted "false" @default.
- W4319069175 workType "article" @default.