Stochastic line search methods via Martingale analysis

Line search methods have represented an important avenue for guaranteeing global convergence of algorithms in nonlinear optimization. In this review, we discuss a recent effort to develop counterparts that can contend with expectation-valued functions via Martingale analysis.

  • Advised by: Dr. Uday V. Shanbhag
  • Reviewed by: Dr. Eunhye Song

Files

Metadata

Work Title Stochastic line search methods via Martingale analysis
Access
Open Access
Creators
  1. Joonha Chang
Keyword
  1. Line search
  2. Stochastic gradient descent
License CC0 1.0 (Public Domain Dedication)
Work Type Research Paper
Publication Date 2021
Deposited July 16, 2021

Versions

Analytics

Collections

This resource is currently not in any collection.

Work History

Version 1
published

  • Created

    July 16, 2021 11:19 by juc1609

  • Added Creator Joonha Chang

    July 16, 2021 11:19 by juc1609

  • Added Joonha Chang — PSU MS Paper, Signed.pdf

    July 16, 2021 11:19 by juc1609

  • Updated Keyword, Description, Publication Date, and 1 more Show Changes

    July 16, 2021 11:23 by juc1609

    Keyword
    • Line search, Stochastic gradient descent
    Description
    • Line search methods have represented an important avenue for guaranteeing global convergence of algorithms in nonlinear optimization. In this review, we discuss a recent effort to develop counterparts that can contend with expectation-valued functions via Martingale analysis.
    • - Advised by: Dr. Uday V. Shanbhag
    • - Reviewed by: Dr. Eunhye Song
    Publication Date
    • 2021
    License
    • http://creativecommons.org/publicdomain/zero/1.0/
  • Published

    July 16, 2021 11:23 by juc1609

  • Updated

    March 22, 2022 16:17 by [unknown user]

Version 2
published

  • Created

    July 19, 2021 23:08 by juc1609

  • Deleted Joonha Chang — PSU MS Paper, Signed.pdf

    July 19, 2021 23:08 by juc1609

  • Added Joonha Chang — PSU MS Paper, Signed.pdf

    July 19, 2021 23:08 by juc1609

  • Published

    July 19, 2021 23:09 by juc1609

  • Updated

    March 22, 2022 16:17 by [unknown user]