SYSTEMS AND METHODS FOR ROBUST LARGE-SCALE MACHINE LEARNING

The present disclosure provides a new scalable coordinate descent (SCD) algorithm and associated system for generalized linear models whose convergence behavior is always the same, regardless of how much SCD is scaled out and regardless of the computing environment. This makes SCD highly robust and...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	SU, Bor-Yiing, RENDLE, Steffen, SHEKITA, Eugene, FETTERLY, Dennis Craig
Format:	Patent
Sprache:	eng ; fre ; ger
Schlagworte:	CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING COUNTING PHYSICS
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	SU, Bor-Yiing RENDLE, Steffen SHEKITA, Eugene FETTERLY, Dennis Craig
description	The present disclosure provides a new scalable coordinate descent (SCD) algorithm and associated system for generalized linear models whose convergence behavior is always the same, regardless of how much SCD is scaled out and regardless of the computing environment. This makes SCD highly robust and enables it to scale to massive datasets on low-cost commodity servers. According to one aspect, by using a natural partitioning of parameters into blocks, updates can be performed in parallel a block at a time without compromising convergence. Experimental results on a real advertising dataset are used to demonstrate SCD's cost effectiveness and scalability.
format	Patent
fullrecord	<record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_EP3380993A1</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>EP3380993A1</sourcerecordid><originalsourceid>FETCH-epo_espacenet_EP3380993A13</originalsourceid><addsrcrecordid>eNrjZLAOjgwOcfUNVnD0c1HwdQ3x8HcJVnDzD1II8ncKDQ5R8HEMcnfVDXZ29HFV8HV09vD0c1XwcXUM8vP0c-dhYE1LzClO5YXS3AwKbq4hzh66qQX58anFBYnJqXmpJfGuAcbGFgaWlsaOhsZEKAEAh9co2g</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>SYSTEMS AND METHODS FOR ROBUST LARGE-SCALE MACHINE LEARNING</title><source>esp@cenet</source><creator>SU, Bor-Yiing ; RENDLE, Steffen ; SHEKITA, Eugene ; FETTERLY, Dennis Craig</creator><creatorcontrib>SU, Bor-Yiing ; RENDLE, Steffen ; SHEKITA, Eugene ; FETTERLY, Dennis Craig</creatorcontrib><description>The present disclosure provides a new scalable coordinate descent (SCD) algorithm and associated system for generalized linear models whose convergence behavior is always the same, regardless of how much SCD is scaled out and regardless of the computing environment. This makes SCD highly robust and enables it to scale to massive datasets on low-cost commodity servers. According to one aspect, by using a natural partitioning of parameters into blocks, updates can be performed in parallel a block at a time without compromising convergence. Experimental results on a real advertising dataset are used to demonstrate SCD's cost effectiveness and scalability.</description><language>eng ; fre ; ger</language><subject>CALCULATING ; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS ; COMPUTING ; COUNTING ; PHYSICS</subject><creationdate>2018</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20181003&DB=EPODOC&CC=EP&NR=3380993A1$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25564,76547</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20181003&DB=EPODOC&CC=EP&NR=3380993A1$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>SU, Bor-Yiing</creatorcontrib><creatorcontrib>RENDLE, Steffen</creatorcontrib><creatorcontrib>SHEKITA, Eugene</creatorcontrib><creatorcontrib>FETTERLY, Dennis Craig</creatorcontrib><title>SYSTEMS AND METHODS FOR ROBUST LARGE-SCALE MACHINE LEARNING</title><description>The present disclosure provides a new scalable coordinate descent (SCD) algorithm and associated system for generalized linear models whose convergence behavior is always the same, regardless of how much SCD is scaled out and regardless of the computing environment. This makes SCD highly robust and enables it to scale to massive datasets on low-cost commodity servers. According to one aspect, by using a natural partitioning of parameters into blocks, updates can be performed in parallel a block at a time without compromising convergence. Experimental results on a real advertising dataset are used to demonstrate SCD's cost effectiveness and scalability.</description><subject>CALCULATING</subject><subject>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2018</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNrjZLAOjgwOcfUNVnD0c1HwdQ3x8HcJVnDzD1II8ncKDQ5R8HEMcnfVDXZ29HFV8HV09vD0c1XwcXUM8vP0c-dhYE1LzClO5YXS3AwKbq4hzh66qQX58anFBYnJqXmpJfGuAcbGFgaWlsaOhsZEKAEAh9co2g</recordid><startdate>20181003</startdate><enddate>20181003</enddate><creator>SU, Bor-Yiing</creator><creator>RENDLE, Steffen</creator><creator>SHEKITA, Eugene</creator><creator>FETTERLY, Dennis Craig</creator><scope>EVB</scope></search><sort><creationdate>20181003</creationdate><title>SYSTEMS AND METHODS FOR ROBUST LARGE-SCALE MACHINE LEARNING</title><author>SU, Bor-Yiing ; RENDLE, Steffen ; SHEKITA, Eugene ; FETTERLY, Dennis Craig</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_EP3380993A13</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>eng ; fre ; ger</language><creationdate>2018</creationdate><topic>CALCULATING</topic><topic>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>SU, Bor-Yiing</creatorcontrib><creatorcontrib>RENDLE, Steffen</creatorcontrib><creatorcontrib>SHEKITA, Eugene</creatorcontrib><creatorcontrib>FETTERLY, Dennis Craig</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>SU, Bor-Yiing</au><au>RENDLE, Steffen</au><au>SHEKITA, Eugene</au><au>FETTERLY, Dennis Craig</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>SYSTEMS AND METHODS FOR ROBUST LARGE-SCALE MACHINE LEARNING</title><date>2018-10-03</date><risdate>2018</risdate><abstract>The present disclosure provides a new scalable coordinate descent (SCD) algorithm and associated system for generalized linear models whose convergence behavior is always the same, regardless of how much SCD is scaled out and regardless of the computing environment. This makes SCD highly robust and enables it to scale to massive datasets on low-cost commodity servers. According to one aspect, by using a natural partitioning of parameters into blocks, updates can be performed in parallel a block at a time without compromising convergence. Experimental results on a real advertising dataset are used to demonstrate SCD's cost effectiveness and scalability.</abstract><oa>free_for_read</oa></addata></record>
fulltext	fulltext_linktorsrc
identifier
ispartof
issn
language	eng ; fre ; ger
recordid	cdi_epo_espacenet_EP3380993A1
source	esp@cenet
subjects	CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING COUNTING PHYSICS
title	SYSTEMS AND METHODS FOR ROBUST LARGE-SCALE MACHINE LEARNING
url	https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-28T03%3A22%3A29IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=SU,%20Bor-Yiing&rft.date=2018-10-03&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3EEP3380993A1%3C/epo_EVB%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