Acceleration of randomized Kaczmarz method via the Johnson-Lindenstrauss lemma (Q639988): Difference between revisions

From MaRDI portal
Created claim: Wikidata QID (P12): Q124810364, #quickstatements; #temporary_batch_1712201099914
ReferenceBot (talk | contribs)
Changed an Item
 
(One intermediate revision by one other user not shown)
Property / arXiv ID
 
Property / arXiv ID: 1008.4397 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Database-friendly random projections: Johnson-Lindenstrauss with binary coins. / rank
 
Normal rank
Property / cites work
 
Property / cites work: The Fast Johnson–Lindenstrauss Transform and Approximate Nearest Neighbors / rank
 
Normal rank
Property / cites work
 
Property / cites work: An Almost Optimal Unrestricted Fast Johnson-Lindenstrauss Transform / rank
 
Normal rank
Property / cites work
 
Property / cites work: Analysis of Projection Methods for Solving Linear Systems with Multiple Right-Hand Sides / rank
 
Normal rank
Property / cites work
 
Property / cites work: An elementary proof of a theorem of Johnson and Lindenstrauss / rank
 
Normal rank
Property / cites work
 
Property / cites work: The rate of convergence for the method of alternating projections. II / rank
 
Normal rank
Property / cites work
 
Property / cites work: Sampling algorithms for <i>l</i><sub>2</sub> regression and applications / rank
 
Normal rank
Property / cites work
 
Property / cites work: Relative-Error $CUR$ Matrix Decompositions / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q4313431 / rank
 
Normal rank
Property / cites work
 
Property / cites work: On the rate of convergence of the alternating projection method in finite dimensional spaces / rank
 
Normal rank
Property / cites work
 
Property / cites work: On the acceleration of Kaczmarz's method for inconsistent linear systems / rank
 
Normal rank
Property / cites work
 
Property / cites work: Johnson-Lindenstrauss lemma for circulant matrices** / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q4542583 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Extensions of Lipschitz mappings into a Hilbert space / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q3754464 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Randomized Kaczmarz solver for noisy linear systems / rank
 
Normal rank
Property / cites work
 
Property / cites work: A Levinson--Galerkin Algorithm for Regularized Trigonometric Approximation / rank
 
Normal rank
Property / cites work
 
Property / cites work: A Randomized Solver for Linear Systems with Exponential Convergence / rank
 
Normal rank
Property / cites work
 
Property / cites work: A randomized Kaczmarz algorithm with exponential convergence / rank
 
Normal rank

Latest revision as of 13:14, 4 July 2024

scientific article
Language Label Description Also known as
English
Acceleration of randomized Kaczmarz method via the Johnson-Lindenstrauss lemma
scientific article

    Statements

    Acceleration of randomized Kaczmarz method via the Johnson-Lindenstrauss lemma (English)
    0 references
    0 references
    0 references
    11 October 2011
    0 references
    To solve the system \(Ax=b\), \(A\in{\mathbb R}^{m\times n}\), the Kaczmarz algorithm updates in iteration \(k\) the vector \(x\) by projecting it orthogonally onto the hyperplane \(a_i^Tx=b_i\) where \(a_i\) is the \(i\)th row of \(A\). Hereby \(i\) changes cyclically. In the randomized version, \(i\) is chosen with a probability proportional to the relative norm of \(a_i\) [see \textit{T. Strohmer} and \textit{R. Vershynin}, J. Fourier Anal. Appl. 15, No.~2, 262--278 (2009; Zbl 1169.68052)]. In this paper the same idea is generalized by projecting the rows of \(A\) onto a lower dimensional subspace (Johnson-Lindenstrauss technique), which makes it more efficient to compute the optimal choice of \(i\). It is proved that the convergence rate is expected to be at least as fast as the classical randomized Kaczmarz method and it is experimentally shown to be much faster in most cases [cf. \textit{S. Kaczmarz}, Bull. Int. Acad. Polon. Sci. A 1937, 355--357 (1937; Zbl 0017.31703)].
    0 references
    0 references
    0 references
    0 references
    0 references
    Kaczmarz method
    0 references
    randomized Kaczmarz method
    0 references
    computer tomography
    0 references
    signal processing
    0 references
    numerical examples
    0 references
    Johnson-Lindenstrauss technique
    0 references
    convergence
    0 references
    0 references
    0 references
    0 references