Convergence Analysis of Accelerated Stochastic Gradient Descent Under the Growth Condition
Convergence Analysis of Accelerated Stochastic Gradient Descent Under the Growth Condition
About this item
Full title
Author / Creator
Publisher
INFORMS
Journal title
Language
English
Formats
Publication information
Publisher
INFORMS
Subjects
More information
Scope and Contents
Contents
We study the convergence of accelerated stochastic gradient descent (SGD) for strongly convex objectives under the
growth condition
, which states that the variance of stochastic gradient is bounded by a multiplicative part that grows with the full gradient and a constant additive part. Through the lens of the growth condition, we investigate...
Alternative Titles
Full title
Convergence Analysis of Accelerated Stochastic Gradient Descent Under the Growth Condition
Authors, Artists and Contributors
Author / Creator
Identifiers
Primary Identifiers
Record Identifier
TN_cdi_crossref_primary_10_1287_moor_2021_0293
Permalink
https://devfeature-collection.sl.nsw.gov.au/record/TN_cdi_crossref_primary_10_1287_moor_2021_0293
Other Identifiers
ISSN
0364-765X
E-ISSN
1526-5471
DOI
10.1287/moor.2021.0293
