Matches in SemOpenAlex for { <https://semopenalex.org/work/W3105816804> ?p ?o ?g. }
- W3105816804 abstract "We study decentralized non-convex finite-sum minimization problems described over a network of nodes, where each node possesses a local batch of data samples. In this context, we analyze a single-timescale randomized incremental gradient method, called GT-SAGA. GT-SAGA is computationally efficient as it evaluates one component gradient per node per iteration and achieves provably fast and robust performance by leveraging node-level variance reduction and network-level gradient tracking. For general smooth non-convex problems, we show the almost sure and mean-squared convergence of GT-SAGA to a first-order stationary point and further describe regimes of practical significance where it outperforms the existing approaches and achieves a network topology-independent iteration complexity respectively. When the global function satisfies the Polyak-Lojaciewisz condition, we show that GT-SAGA exhibits linear convergence to an optimal solution in expectation and describe regimes of practical interest where the performance is network topology-independent and improves upon the existing methods. Numerical experiments are included to highlight the main convergence aspects of GT-SAGA in non-convex settings." @default.
- W3105816804 created "2020-11-23" @default.
- W3105816804 creator A5025641667 @default.
- W3105816804 creator A5033024261 @default.
- W3105816804 creator A5077268766 @default.
- W3105816804 date "2020-11-07" @default.
- W3105816804 modified "2023-10-18" @default.
- W3105816804 title "A fast randomized incremental gradient method for decentralized non-convex optimization" @default.
- W3105816804 cites W1553365534 @default.
- W3105816804 cites W1568288633 @default.
- W3105816804 cites W1571416372 @default.
- W3105816804 cites W1602844557 @default.
- W3105816804 cites W1616857247 @default.
- W3105816804 cites W2029080014 @default.
- W3105816804 cites W2047152541 @default.
- W3105816804 cites W2066332749 @default.
- W3105816804 cites W2107438106 @default.
- W3105816804 cites W2108306501 @default.
- W3105816804 cites W2135482703 @default.
- W3105816804 cites W2151532749 @default.
- W3105816804 cites W2154834860 @default.
- W3105816804 cites W2216483973 @default.
- W3105816804 cites W2266148951 @default.
- W3105816804 cites W2291434103 @default.
- W3105816804 cites W2556623528 @default.
- W3105816804 cites W2570983198 @default.
- W3105816804 cites W2609549560 @default.
- W3105816804 cites W2610857016 @default.
- W3105816804 cites W2763081248 @default.
- W3105816804 cites W2782375002 @default.
- W3105816804 cites W2789243870 @default.
- W3105816804 cites W2806414180 @default.
- W3105816804 cites W2896243341 @default.
- W3105816804 cites W2902410171 @default.
- W3105816804 cites W2942536919 @default.
- W3105816804 cites W2949099147 @default.
- W3105816804 cites W2953881522 @default.
- W3105816804 cites W2962696654 @default.
- W3105816804 cites W2962744561 @default.
- W3105816804 cites W2962771678 @default.
- W3105816804 cites W2962913067 @default.
- W3105816804 cites W2963087817 @default.
- W3105816804 cites W2963228337 @default.
- W3105816804 cites W2963248893 @default.
- W3105816804 cites W2963649943 @default.
- W3105816804 cites W2964075583 @default.
- W3105816804 cites W2979725263 @default.
- W3105816804 cites W2991772480 @default.
- W3105816804 cites W3003613721 @default.
- W3105816804 cites W3007193855 @default.
- W3105816804 cites W3009274618 @default.
- W3105816804 cites W3009875021 @default.
- W3105816804 cites W3010627535 @default.
- W3105816804 cites W3011252985 @default.
- W3105816804 cites W3012206388 @default.
- W3105816804 cites W3021559198 @default.
- W3105816804 cites W3023944502 @default.
- W3105816804 cites W3033819030 @default.
- W3105816804 cites W3035647850 @default.
- W3105816804 cites W3038605285 @default.
- W3105816804 cites W3042491641 @default.
- W3105816804 cites W3043048378 @default.
- W3105816804 cites W3043309535 @default.
- W3105816804 cites W3047780503 @default.
- W3105816804 cites W3048505912 @default.
- W3105816804 cites W3084132440 @default.
- W3105816804 cites W3086080324 @default.
- W3105816804 cites W3089227111 @default.
- W3105816804 cites W3091831107 @default.
- W3105816804 cites W3097497066 @default.
- W3105816804 cites W3100513551 @default.
- W3105816804 cites W3104398353 @default.
- W3105816804 cites W3104871901 @default.
- W3105816804 cites W3129321571 @default.
- W3105816804 cites W3134022240 @default.
- W3105816804 cites W3166071591 @default.
- W3105816804 doi "https://doi.org/10.48550/arxiv.2011.03853" @default.
- W3105816804 hasPublicationYear "2020" @default.
- W3105816804 type Work @default.
- W3105816804 sameAs 3105816804 @default.
- W3105816804 citedByCount "1" @default.
- W3105816804 countsByYear W31058168042021 @default.
- W3105816804 crossrefType "posted-content" @default.
- W3105816804 hasAuthorship W3105816804A5025641667 @default.
- W3105816804 hasAuthorship W3105816804A5033024261 @default.
- W3105816804 hasAuthorship W3105816804A5077268766 @default.
- W3105816804 hasBestOaLocation W31058168041 @default.
- W3105816804 hasConcept C10494615 @default.
- W3105816804 hasConcept C105795698 @default.
- W3105816804 hasConcept C111335779 @default.
- W3105816804 hasConcept C111919701 @default.
- W3105816804 hasConcept C112680207 @default.
- W3105816804 hasConcept C114614502 @default.
- W3105816804 hasConcept C126255220 @default.
- W3105816804 hasConcept C127413603 @default.
- W3105816804 hasConcept C14036430 @default.
- W3105816804 hasConcept C145446738 @default.
- W3105816804 hasConcept C151730666 @default.
- W3105816804 hasConcept C157972887 @default.
- W3105816804 hasConcept C162324750 @default.