Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizers

We present a new, stochastic variant of the projective splitting (PS) family of algorithms for monotone inclusion problems. It can solve min-max and noncooperative game formulations arising in applications such as robust ML without the convergence issues associated with gradient descent-ascent, the...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Johnstone, Patrick R, Eckstein, Jonathan, Flynn, Thomas, Yoo, Shinjae
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Johnstone, Patrick R
Eckstein, Jonathan
Flynn, Thomas
Yoo, Shinjae
description We present a new, stochastic variant of the projective splitting (PS) family of algorithms for monotone inclusion problems. It can solve min-max and noncooperative game formulations arising in applications such as robust ML without the convergence issues associated with gradient descent-ascent, the current de facto standard approach in such situations. Our proposal is the first version of PS able to use stochastic (as opposed to deterministic) gradient oracles. It is also the first stochastic method that can solve min-max games while easily handling multiple constraints and nonsmooth regularizers via projection and proximal operators. We close with numerical experiments on a distributionally robust sparse logistic regression problem.
doi_str_mv 10.48550/arxiv.2106.13067
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2106_13067</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2106_13067</sourcerecordid><originalsourceid>FETCH-LOGICAL-a677-5e369405ca1f00a788db2770f69938cee4f9c5c5c1d25820a4934ffcc4c3ae473</originalsourceid><addsrcrecordid>eNotj8tOwzAURL1hgQofwKr-gQQntmObHap4SUVUpJuuolvnujVym8hxw-PraQuaxcziaKRDyE3BcqGlZLcQv_yYlwWr8oKzSl2SVZ06u4UheUsXsftAm_yItO6DT8nvN3e07sJ4HLSGtg2YLTq_Tyd0HXA30E-ftvT1EJLvA9J33BwCRP-DcbgiFw7CgNf_PSHLx4fl7Dmbvz29zO7nGVRKZRJ5ZQSTFgrHGCit23WpFHOVMVxbROGMlccUbSl1yUAYLpyzVlgOKBSfkOnf7dmt6aPfQfxuTo7N2ZH_AuvWTcA</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizers</title><source>arXiv.org</source><creator>Johnstone, Patrick R ; Eckstein, Jonathan ; Flynn, Thomas ; Yoo, Shinjae</creator><creatorcontrib>Johnstone, Patrick R ; Eckstein, Jonathan ; Flynn, Thomas ; Yoo, Shinjae</creatorcontrib><description>We present a new, stochastic variant of the projective splitting (PS) family of algorithms for monotone inclusion problems. It can solve min-max and noncooperative game formulations arising in applications such as robust ML without the convergence issues associated with gradient descent-ascent, the current de facto standard approach in such situations. Our proposal is the first version of PS able to use stochastic (as opposed to deterministic) gradient oracles. It is also the first stochastic method that can solve min-max games while easily handling multiple constraints and nonsmooth regularizers via projection and proximal operators. We close with numerical experiments on a distributionally robust sparse logistic regression problem.</description><identifier>DOI: 10.48550/arxiv.2106.13067</identifier><language>eng</language><subject>Computer Science - Learning ; Mathematics - Optimization and Control</subject><creationdate>2021-06</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,781,886</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2106.13067$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2106.13067$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Johnstone, Patrick R</creatorcontrib><creatorcontrib>Eckstein, Jonathan</creatorcontrib><creatorcontrib>Flynn, Thomas</creatorcontrib><creatorcontrib>Yoo, Shinjae</creatorcontrib><title>Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizers</title><description>We present a new, stochastic variant of the projective splitting (PS) family of algorithms for monotone inclusion problems. It can solve min-max and noncooperative game formulations arising in applications such as robust ML without the convergence issues associated with gradient descent-ascent, the current de facto standard approach in such situations. Our proposal is the first version of PS able to use stochastic (as opposed to deterministic) gradient oracles. It is also the first stochastic method that can solve min-max games while easily handling multiple constraints and nonsmooth regularizers via projection and proximal operators. We close with numerical experiments on a distributionally robust sparse logistic regression problem.</description><subject>Computer Science - Learning</subject><subject>Mathematics - Optimization and Control</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj8tOwzAURL1hgQofwKr-gQQntmObHap4SUVUpJuuolvnujVym8hxw-PraQuaxcziaKRDyE3BcqGlZLcQv_yYlwWr8oKzSl2SVZ06u4UheUsXsftAm_yItO6DT8nvN3e07sJ4HLSGtg2YLTq_Tyd0HXA30E-ftvT1EJLvA9J33BwCRP-DcbgiFw7CgNf_PSHLx4fl7Dmbvz29zO7nGVRKZRJ5ZQSTFgrHGCit23WpFHOVMVxbROGMlccUbSl1yUAYLpyzVlgOKBSfkOnf7dmt6aPfQfxuTo7N2ZH_AuvWTcA</recordid><startdate>20210624</startdate><enddate>20210624</enddate><creator>Johnstone, Patrick R</creator><creator>Eckstein, Jonathan</creator><creator>Flynn, Thomas</creator><creator>Yoo, Shinjae</creator><scope>AKY</scope><scope>AKZ</scope><scope>GOX</scope></search><sort><creationdate>20210624</creationdate><title>Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizers</title><author>Johnstone, Patrick R ; Eckstein, Jonathan ; Flynn, Thomas ; Yoo, Shinjae</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a677-5e369405ca1f00a788db2770f69938cee4f9c5c5c1d25820a4934ffcc4c3ae473</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Computer Science - Learning</topic><topic>Mathematics - Optimization and Control</topic><toplevel>online_resources</toplevel><creatorcontrib>Johnstone, Patrick R</creatorcontrib><creatorcontrib>Eckstein, Jonathan</creatorcontrib><creatorcontrib>Flynn, Thomas</creatorcontrib><creatorcontrib>Yoo, Shinjae</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv Mathematics</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Johnstone, Patrick R</au><au>Eckstein, Jonathan</au><au>Flynn, Thomas</au><au>Yoo, Shinjae</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizers</atitle><date>2021-06-24</date><risdate>2021</risdate><abstract>We present a new, stochastic variant of the projective splitting (PS) family of algorithms for monotone inclusion problems. It can solve min-max and noncooperative game formulations arising in applications such as robust ML without the convergence issues associated with gradient descent-ascent, the current de facto standard approach in such situations. Our proposal is the first version of PS able to use stochastic (as opposed to deterministic) gradient oracles. It is also the first stochastic method that can solve min-max games while easily handling multiple constraints and nonsmooth regularizers via projection and proximal operators. We close with numerical experiments on a distributionally robust sparse logistic regression problem.</abstract><doi>10.48550/arxiv.2106.13067</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2106.13067
ispartof
issn
language eng
recordid cdi_arxiv_primary_2106_13067
source arXiv.org
subjects Computer Science - Learning
Mathematics - Optimization and Control
title Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizers
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-16T16%3A02%3A31IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Stochastic%20Projective%20Splitting:%20Solving%20Saddle-Point%20Problems%20with%20Multiple%20Regularizers&rft.au=Johnstone,%20Patrick%20R&rft.date=2021-06-24&rft_id=info:doi/10.48550/arxiv.2106.13067&rft_dat=%3Carxiv_GOX%3E2106_13067%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true