eprintid: 10140922
rev_number: 15
eprint_status: archive
userid: 608
dir: disk0/10/14/09/22
datestamp: 2022-01-05 14:52:49
lastmod: 2023-01-05 07:10:25
status_changed: 2022-01-05 14:52:49
type: article
metadata_visibility: show
creators_name: Jin, B
creators_name: Zhou, Z
creators_name: Zou, J
title: An analysis of stochastic variance reduced gradient for linear inverse problems
ispublished: pub
divisions: UCL
divisions: B04
divisions: C05
divisions: F48
note: This version is the author accepted manuscript. For information on re-use, please refer to the publisher's terms and conditions.
abstract: Stochastic variance reduced gradient (SVRG) is a popular variance reduction technique for accelerating stochastic gradient descent (SGD). We provide a first analysis of the method for solving a class of linear inverse problems in the lens of the classical regularization theory. We prove that for a suitable constant step size schedule, the method can achieve an optimal convergence rate in terms of the noise level (under suitable regularity condition) and the variance of the SVRG iterate error is smaller than that by SGD. These theoretical findings are corroborated by a set of numerical experiments.
date: 2021-01-04
publisher: IOP Publishing
official_url: https://doi.org/10.1088/1361-6420/ac4428
oa_status: green
full_text_type: other
language: eng
primo: open
primo_central: open_green
verified: verified_manual
elements_id: 1911441
doi: 10.1088/1361-6420/ac4428
lyricists_name: Jin, Bangti
lyricists_id: BJINX59
actors_name: Flynn, Bernadette
actors_id: BFFLY94
actors_role: owner
full_text_status: public
publication: Inverse Problems
volume: 38
number: 2
article_number: 025009
citation:        Jin, B;    Zhou, Z;    Zou, J;      (2021)    An analysis of stochastic variance reduced gradient for linear inverse problems.                   Inverse Problems , 38  (2)    , Article 025009.  10.1088/1361-6420/ac4428 <https://doi.org/10.1088/1361-6420%2Fac4428>.       Green open access   
 
document_url: https://discovery.ucl.ac.uk/id/eprint/10140922/1/Jin%2Bet%2Bal_2021_Inverse_Problems_10.1088_1361-6420_ac4428.pdf