Matches in SemOpenAlex for { <https://semopenalex.org/work/W2808630040> ?p ?o ?g. }
Showing items 1 to 80 of
80
with 100 items per page.
- W2808630040 endingPage "43" @default.
- W2808630040 startingPage "1" @default.
- W2808630040 abstract "In recent years, stochastic gradient descent (SGD) methods and randomized linear algebra (RLA) algorithms have been applied to many large-scale problems in machine learning and data analysis. We aim to bridge the gap between these two methods in solving constrained overdetermined linear regression problems---e.g., $ell_2$ and $ell_1$ regression problems. We propose a hybrid algorithm named pwSGD that uses RLA techniques for preconditioning and constructing an importance sampling distribution, and then performs an SGD-like iterative process with weighted sampling on the preconditioned system. We prove that pwSGD inherits faster convergence rates that only depend on the lower dimension of the linear system, while maintaining low computation complexity. Particularly, when solving $ell_1$ regression with size $n$ by $d$, pwSGD returns an approximate solution with $epsilon$ relative error in the objective value in $mathcal{O}(log n cdot text{nnz}(A) + text{poly}(d)/epsilon^2)$ time. This complexity is uniformly better than that of RLA methods in terms of both $epsilon$ and $d$ when the problem is unconstrained. For $ell_2$ regression, pwSGD returns an approximate solution with $epsilon$ relative error in the objective value and the solution vector measured in prediction norm in $mathcal{O}(log n cdot text{nnz}(A) + text{poly}(d) log(1/epsilon) /epsilon)$ time. We also provide lower bounds on the coreset complexity for more general regression problems, indicating that still new ideas will be needed to extend similar RLA preconditioning ideas to weighted SGD algorithms for more general regression problems. Finally, the effectiveness of such algorithms is illustrated numerically on both synthetic and real datasets." @default.
- W2808630040 created "2018-06-21" @default.
- W2808630040 creator A5033006662 @default.
- W2808630040 creator A5033659558 @default.
- W2808630040 creator A5039948454 @default.
- W2808630040 creator A5070267604 @default.
- W2808630040 date "2018-01-01" @default.
- W2808630040 modified "2023-09-28" @default.
- W2808630040 title "Weighted SGD for $ell_p$ Regression with Randomized Preconditioning" @default.
- W2808630040 hasPublicationYear "2018" @default.
- W2808630040 type Work @default.
- W2808630040 sameAs 2808630040 @default.
- W2808630040 citedByCount "1" @default.
- W2808630040 countsByYear W28086300402020 @default.
- W2808630040 crossrefType "journal-article" @default.
- W2808630040 hasAuthorship W2808630040A5033006662 @default.
- W2808630040 hasAuthorship W2808630040A5033659558 @default.
- W2808630040 hasAuthorship W2808630040A5039948454 @default.
- W2808630040 hasAuthorship W2808630040A5070267604 @default.
- W2808630040 hasConcept C105795698 @default.
- W2808630040 hasConcept C11413529 @default.
- W2808630040 hasConcept C114614502 @default.
- W2808630040 hasConcept C118615104 @default.
- W2808630040 hasConcept C154945302 @default.
- W2808630040 hasConcept C17744445 @default.
- W2808630040 hasConcept C191795146 @default.
- W2808630040 hasConcept C199539241 @default.
- W2808630040 hasConcept C206688291 @default.
- W2808630040 hasConcept C33676613 @default.
- W2808630040 hasConcept C33923547 @default.
- W2808630040 hasConcept C41008148 @default.
- W2808630040 hasConcept C48921125 @default.
- W2808630040 hasConcept C50644808 @default.
- W2808630040 hasConcept C63553672 @default.
- W2808630040 hasConceptScore W2808630040C105795698 @default.
- W2808630040 hasConceptScore W2808630040C11413529 @default.
- W2808630040 hasConceptScore W2808630040C114614502 @default.
- W2808630040 hasConceptScore W2808630040C118615104 @default.
- W2808630040 hasConceptScore W2808630040C154945302 @default.
- W2808630040 hasConceptScore W2808630040C17744445 @default.
- W2808630040 hasConceptScore W2808630040C191795146 @default.
- W2808630040 hasConceptScore W2808630040C199539241 @default.
- W2808630040 hasConceptScore W2808630040C206688291 @default.
- W2808630040 hasConceptScore W2808630040C33676613 @default.
- W2808630040 hasConceptScore W2808630040C33923547 @default.
- W2808630040 hasConceptScore W2808630040C41008148 @default.
- W2808630040 hasConceptScore W2808630040C48921125 @default.
- W2808630040 hasConceptScore W2808630040C50644808 @default.
- W2808630040 hasConceptScore W2808630040C63553672 @default.
- W2808630040 hasIssue "211" @default.
- W2808630040 hasLocation W28086300401 @default.
- W2808630040 hasOpenAccess W2808630040 @default.
- W2808630040 hasPrimaryLocation W28086300401 @default.
- W2808630040 hasRelatedWork W145880085 @default.
- W2808630040 hasRelatedWork W1909276365 @default.
- W2808630040 hasRelatedWork W2223328134 @default.
- W2808630040 hasRelatedWork W2411556044 @default.
- W2808630040 hasRelatedWork W2524580019 @default.
- W2808630040 hasRelatedWork W2752370501 @default.
- W2808630040 hasRelatedWork W2787135684 @default.
- W2808630040 hasRelatedWork W2939588467 @default.
- W2808630040 hasRelatedWork W2949795768 @default.
- W2808630040 hasRelatedWork W2955621017 @default.
- W2808630040 hasRelatedWork W2957233745 @default.
- W2808630040 hasRelatedWork W2962875879 @default.
- W2808630040 hasRelatedWork W3028145528 @default.
- W2808630040 hasRelatedWork W3038605285 @default.
- W2808630040 hasRelatedWork W3080910898 @default.
- W2808630040 hasRelatedWork W3083048200 @default.
- W2808630040 hasRelatedWork W3126892819 @default.
- W2808630040 hasRelatedWork W3208472520 @default.
- W2808630040 hasRelatedWork W3209099313 @default.
- W2808630040 hasRelatedWork W2963821730 @default.
- W2808630040 hasVolume "18" @default.
- W2808630040 isParatext "false" @default.
- W2808630040 isRetracted "false" @default.
- W2808630040 magId "2808630040" @default.
- W2808630040 workType "article" @default.