Matches in SemOpenAlex for { <https://semopenalex.org/work/W4366149060> ?p ?o ?g. }
- W4366149060 endingPage "5276" @default.
- W4366149060 startingPage "5253" @default.
- W4366149060 abstract "We study linear regression from data distributed over a network of agents (with no server node) by means of LASSO estimation, in high-dimension, which allows the ambient dimension to grow faster than the sample size. While there is a vast literature of distributed algorithms applicable to the problem, statistical and computational guarantees of most of them remain unclear in high dimension. This paper provides a first statistical study of the Distributed Gradient Descent (DGD) in the Adapt-Then-Combine (ATC) form. Our theory shows that, under standard notions of restricted strong convexity and smoothness of the loss functions–which hold with high probability for standard data generation models–suitable conditions on the network connectivity and algorithm tuning, DGD-ATC converges globally at a linear rate to an estimate that is within the <italic xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>centralized</i> statistical precision of the model. In the worst-case scenario, the total number of communications to statistical optimality grows logarithmically with the ambient dimension, which improves on the communication complexity of DGD in the Combine-Then-Adapt (CTA) form, scaling linearly with the dimension. This reveals that mixing gradient information among agents, as DGD-ATC does, is critical in high-dimensions to obtain favorable rate scalings." @default.
- W4366149060 created "2023-04-19" @default.
- W4366149060 creator A5023307885 @default.
- W4366149060 creator A5041765559 @default.
- W4366149060 creator A5046695536 @default.
- W4366149060 creator A5070127492 @default.
- W4366149060 date "2023-08-01" @default.
- W4366149060 modified "2023-09-25" @default.
- W4366149060 title "Distributed (ATC) Gradient Descent for High Dimension Sparse Regression" @default.
- W4366149060 cites W1523985187 @default.
- W4366149060 cites W1571416372 @default.
- W4366149060 cites W1616857247 @default.
- W4366149060 cites W1627780903 @default.
- W4366149060 cites W1984305442 @default.
- W4366149060 cites W2012445782 @default.
- W4366149060 cites W2015418199 @default.
- W4366149060 cites W2024307985 @default.
- W4366149060 cites W2029080014 @default.
- W4366149060 cites W2030449718 @default.
- W4366149060 cites W2044212084 @default.
- W4366149060 cites W2057873962 @default.
- W4366149060 cites W2066544664 @default.
- W4366149060 cites W2114791779 @default.
- W4366149060 cites W2123705108 @default.
- W4366149060 cites W2129122308 @default.
- W4366149060 cites W2142682063 @default.
- W4366149060 cites W2151685478 @default.
- W4366149060 cites W2153635508 @default.
- W4366149060 cites W2266148951 @default.
- W4366149060 cites W2566240941 @default.
- W4366149060 cites W2737743075 @default.
- W4366149060 cites W2782375002 @default.
- W4366149060 cites W2801571965 @default.
- W4366149060 cites W2962771678 @default.
- W4366149060 cites W2964026281 @default.
- W4366149060 cites W2964231067 @default.
- W4366149060 cites W2964256184 @default.
- W4366149060 cites W2965497096 @default.
- W4366149060 cites W3004725330 @default.
- W4366149060 cites W3007193855 @default.
- W4366149060 cites W3016223826 @default.
- W4366149060 cites W3042491641 @default.
- W4366149060 cites W3092262213 @default.
- W4366149060 cites W3093673559 @default.
- W4366149060 cites W3100513551 @default.
- W4366149060 doi "https://doi.org/10.1109/tit.2023.3267742" @default.
- W4366149060 hasPublicationYear "2023" @default.
- W4366149060 type Work @default.
- W4366149060 citedByCount "0" @default.
- W4366149060 crossrefType "journal-article" @default.
- W4366149060 hasAuthorship W4366149060A5023307885 @default.
- W4366149060 hasAuthorship W4366149060A5041765559 @default.
- W4366149060 hasAuthorship W4366149060A5046695536 @default.
- W4366149060 hasAuthorship W4366149060A5070127492 @default.
- W4366149060 hasConcept C102634674 @default.
- W4366149060 hasConcept C105795698 @default.
- W4366149060 hasConcept C106159729 @default.
- W4366149060 hasConcept C11413529 @default.
- W4366149060 hasConcept C114614502 @default.
- W4366149060 hasConcept C126255220 @default.
- W4366149060 hasConcept C127413603 @default.
- W4366149060 hasConcept C134306372 @default.
- W4366149060 hasConcept C136764020 @default.
- W4366149060 hasConcept C153258448 @default.
- W4366149060 hasConcept C154945302 @default.
- W4366149060 hasConcept C162324750 @default.
- W4366149060 hasConcept C2524010 @default.
- W4366149060 hasConcept C33676613 @default.
- W4366149060 hasConcept C33923547 @default.
- W4366149060 hasConcept C37616216 @default.
- W4366149060 hasConcept C39927690 @default.
- W4366149060 hasConcept C41008148 @default.
- W4366149060 hasConcept C48921125 @default.
- W4366149060 hasConcept C50644808 @default.
- W4366149060 hasConcept C62611344 @default.
- W4366149060 hasConcept C66938386 @default.
- W4366149060 hasConcept C72134830 @default.
- W4366149060 hasConcept C99844830 @default.
- W4366149060 hasConceptScore W4366149060C102634674 @default.
- W4366149060 hasConceptScore W4366149060C105795698 @default.
- W4366149060 hasConceptScore W4366149060C106159729 @default.
- W4366149060 hasConceptScore W4366149060C11413529 @default.
- W4366149060 hasConceptScore W4366149060C114614502 @default.
- W4366149060 hasConceptScore W4366149060C126255220 @default.
- W4366149060 hasConceptScore W4366149060C127413603 @default.
- W4366149060 hasConceptScore W4366149060C134306372 @default.
- W4366149060 hasConceptScore W4366149060C136764020 @default.
- W4366149060 hasConceptScore W4366149060C153258448 @default.
- W4366149060 hasConceptScore W4366149060C154945302 @default.
- W4366149060 hasConceptScore W4366149060C162324750 @default.
- W4366149060 hasConceptScore W4366149060C2524010 @default.
- W4366149060 hasConceptScore W4366149060C33676613 @default.
- W4366149060 hasConceptScore W4366149060C33923547 @default.
- W4366149060 hasConceptScore W4366149060C37616216 @default.
- W4366149060 hasConceptScore W4366149060C39927690 @default.
- W4366149060 hasConceptScore W4366149060C41008148 @default.
- W4366149060 hasConceptScore W4366149060C48921125 @default.
- W4366149060 hasConceptScore W4366149060C50644808 @default.