Matches in SemOpenAlex for { <https://semopenalex.org/work/W4297438049> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W4297438049 abstract "Distributed machine learning enables scalability and computational offloading, but requires significant levels of communication. Consequently, communication efficiency in distributed learning settings is an important consideration, especially when the communications are wireless and battery-driven devices are employed. In this paper we develop a censoring-based heavy ball (CHB) method for distributed learning in a server-worker architecture. Each worker self-censors unless its local gradient is sufficiently different from the previously transmitted one. The significant practical advantages of the HB method for learning problems are well known, but the question of reducing communications has not been addressed. CHB takes advantage of the HB smoothing to eliminate reporting small changes, and provably achieves a linear convergence rate equivalent to that of the classical HB method for smooth and strongly convex objective functions. The convergence guarantee of CHB is theoretically justified for both convex and nonconvex cases. In addition we prove that, under some conditions, at least half of all communications can be eliminated without any impact on convergence rate. Extensive numerical results validate the communication efficiency of CHB on both synthetic and real datasets, for convex, nonconvex, and nondifferentiable cases. Given a target accuracy, CHB can significantly reduce the number of communications compared to existing algorithms, achieving the same accuracy without slowing down the optimization process." @default.
- W4297438049 created "2022-09-28" @default.
- W4297438049 creator A5003777876 @default.
- W4297438049 creator A5042577942 @default.
- W4297438049 creator A5083294740 @default.
- W4297438049 date "2022-09-24" @default.
- W4297438049 modified "2023-10-06" @default.
- W4297438049 title "Communication-Efficient {Federated} Learning Using Censored Heavy Ball Descent" @default.
- W4297438049 doi "https://doi.org/10.48550/arxiv.2209.11944" @default.
- W4297438049 hasPublicationYear "2022" @default.
- W4297438049 type Work @default.
- W4297438049 citedByCount "0" @default.
- W4297438049 crossrefType "posted-content" @default.
- W4297438049 hasAuthorship W4297438049A5003777876 @default.
- W4297438049 hasAuthorship W4297438049A5042577942 @default.
- W4297438049 hasAuthorship W4297438049A5083294740 @default.
- W4297438049 hasBestOaLocation W42974380491 @default.
- W4297438049 hasConcept C112680207 @default.
- W4297438049 hasConcept C120314980 @default.
- W4297438049 hasConcept C122041747 @default.
- W4297438049 hasConcept C126255220 @default.
- W4297438049 hasConcept C127162648 @default.
- W4297438049 hasConcept C134306372 @default.
- W4297438049 hasConcept C145446738 @default.
- W4297438049 hasConcept C15744967 @default.
- W4297438049 hasConcept C162324750 @default.
- W4297438049 hasConcept C19417346 @default.
- W4297438049 hasConcept C2524010 @default.
- W4297438049 hasConcept C2777303404 @default.
- W4297438049 hasConcept C2779582901 @default.
- W4297438049 hasConcept C2992525071 @default.
- W4297438049 hasConcept C31258907 @default.
- W4297438049 hasConcept C31972630 @default.
- W4297438049 hasConcept C33923547 @default.
- W4297438049 hasConcept C3770464 @default.
- W4297438049 hasConcept C41008148 @default.
- W4297438049 hasConcept C48044578 @default.
- W4297438049 hasConcept C50522688 @default.
- W4297438049 hasConcept C555944384 @default.
- W4297438049 hasConcept C57869625 @default.
- W4297438049 hasConcept C76155785 @default.
- W4297438049 hasConcept C77088390 @default.
- W4297438049 hasConceptScore W4297438049C112680207 @default.
- W4297438049 hasConceptScore W4297438049C120314980 @default.
- W4297438049 hasConceptScore W4297438049C122041747 @default.
- W4297438049 hasConceptScore W4297438049C126255220 @default.
- W4297438049 hasConceptScore W4297438049C127162648 @default.
- W4297438049 hasConceptScore W4297438049C134306372 @default.
- W4297438049 hasConceptScore W4297438049C145446738 @default.
- W4297438049 hasConceptScore W4297438049C15744967 @default.
- W4297438049 hasConceptScore W4297438049C162324750 @default.
- W4297438049 hasConceptScore W4297438049C19417346 @default.
- W4297438049 hasConceptScore W4297438049C2524010 @default.
- W4297438049 hasConceptScore W4297438049C2777303404 @default.
- W4297438049 hasConceptScore W4297438049C2779582901 @default.
- W4297438049 hasConceptScore W4297438049C2992525071 @default.
- W4297438049 hasConceptScore W4297438049C31258907 @default.
- W4297438049 hasConceptScore W4297438049C31972630 @default.
- W4297438049 hasConceptScore W4297438049C33923547 @default.
- W4297438049 hasConceptScore W4297438049C3770464 @default.
- W4297438049 hasConceptScore W4297438049C41008148 @default.
- W4297438049 hasConceptScore W4297438049C48044578 @default.
- W4297438049 hasConceptScore W4297438049C50522688 @default.
- W4297438049 hasConceptScore W4297438049C555944384 @default.
- W4297438049 hasConceptScore W4297438049C57869625 @default.
- W4297438049 hasConceptScore W4297438049C76155785 @default.
- W4297438049 hasConceptScore W4297438049C77088390 @default.
- W4297438049 hasLocation W42974380491 @default.
- W4297438049 hasOpenAccess W4297438049 @default.
- W4297438049 hasPrimaryLocation W42974380491 @default.
- W4297438049 hasRelatedWork W1596201972 @default.
- W4297438049 hasRelatedWork W2137731592 @default.
- W4297438049 hasRelatedWork W2364921833 @default.
- W4297438049 hasRelatedWork W2385146268 @default.
- W4297438049 hasRelatedWork W2952586095 @default.
- W4297438049 hasRelatedWork W3129682320 @default.
- W4297438049 hasRelatedWork W4297438049 @default.
- W4297438049 hasRelatedWork W4298016328 @default.
- W4297438049 hasRelatedWork W4312926285 @default.
- W4297438049 hasRelatedWork W4320067866 @default.
- W4297438049 isParatext "false" @default.
- W4297438049 isRetracted "false" @default.
- W4297438049 workType "article" @default.