{"id":"https://openalex.org/W2996693845","doi":null,"title":"SloMo: Improving Communication-Efficient Distributed SGD with Slow Momentum","display_name":"SloMo: Improving Communication-Efficient Distributed SGD with Slow Momentum","publication_year":2020,"publication_date":"2020-04-30","ids":{"openalex":"https://openalex.org/W2996693845","mag":"2996693845"},"language":"en","primary_location":{"id":"mag:2996693845","doi":null,"is_oa":false,"landing_page_url":"https://www.openreview.net/pdf?id=SkxJ8REYPH","pdf_url":null,"source":{"id":"https://openalex.org/S4306419637","display_name":"International Conference on Learning Representations","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"International Conference on Learning Representations","raw_type":null},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100398977","display_name":"Jianyu Wang","orcid":"https://orcid.org/0000-0002-7075-9333","observed_orcids":["https://orcid.org/0000-0002-7075-9333"]},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jianyu Wang","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, United States"],"raw_orcid":"https://orcid.org/0000-0002-7075-9333","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, United States","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074049737","display_name":"Vinayak Tantia","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2252078561","display_name":"Meta (Israel)","ror":"https://ror.org/02388em19","country_code":"IL","type":"company","lineage":["https://openalex.org/I2252078561","https://openalex.org/I4210114444"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Vinayak Tantia","raw_affiliation_strings":["Meta (Israel), Tel Aviv, Israel"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta (Israel), Tel Aviv, Israel","institution_ids":["https://openalex.org/I2252078561"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057065873","display_name":"Nicolas Ballas","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2252078561","display_name":"Meta (Israel)","ror":"https://ror.org/02388em19","country_code":"IL","type":"company","lineage":["https://openalex.org/I2252078561","https://openalex.org/I4210114444"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Nicolas Ballas","raw_affiliation_strings":["Meta (Israel), Tel Aviv, Israel"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta (Israel), Tel Aviv, Israel","institution_ids":["https://openalex.org/I2252078561"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5089819604","display_name":"Michael Rabbat","orcid":"https://orcid.org/0000-0003-0536-7904","observed_orcids":["https://orcid.org/0000-0003-0536-7904"]},"institutions":[{"id":"https://openalex.org/I2252078561","display_name":"Meta (Israel)","ror":"https://ror.org/02388em19","country_code":"IL","type":"company","lineage":["https://openalex.org/I2252078561","https://openalex.org/I4210114444"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Michael Rabbat","raw_affiliation_strings":["Meta (Israel), Tel Aviv, Israel"],"raw_orcid":"https://orcid.org/0000-0003-0536-7904","affiliations":[{"raw_affiliation_string":"Meta (Israel), Tel Aviv, Israel","institution_ids":["https://openalex.org/I2252078561"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.6393,"has_fulltext":false,"cited_by_count":34,"citation_normalized_percentile":{"value":0.921419,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":88,"max":100},"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9990000128746033,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9990000128746033,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10249","display_name":"Distributed Control Multi-Agent Systems","score":0.00039999998989515007,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12535","display_name":"Machine Learning and Data Classification","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/communication-efficiency","display_name":"communication efficiency","score":0.9179999828338623},{"id":"https://openalex.org/keywords/momentum-optimization","display_name":"momentum optimization","score":0.45100000500679016},{"id":"https://openalex.org/keywords/nonconvex-optimization","display_name":"non-convex optimization","score":0.7490000128746033},{"id":"https://openalex.org/keywords/convergence-guarantees","display_name":"convergence guarantees","score":0.968999981880188},{"id":"https://openalex.org/keywords/image-classification","display_name":"image classification","score":0.7950000166893005},{"id":"https://openalex.org/keywords/machine-translation","display_name":"machine translation","score":0.9990000128746033}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7547220587730408},{"id":"https://openalex.org/C162932704","wikidata":"https://www.wikidata.org/wiki/Q1058791","display_name":"Synchronizing","level":3,"score":0.5713366270065308},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5652382969856262},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.5140239000320435},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.4957793653011322},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.44859281182289124},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.4322199821472168},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.43076834082603455},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.4256861209869385},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3816410303115845},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3396783173084259},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.15388363599777222},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.09668770432472229},{"id":"https://openalex.org/C761482","wikidata":"https://www.wikidata.org/wiki/Q118093","display_name":"Transmission (telecommunications)","level":2,"score":0.0},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C50522688","wikidata":"https://www.wikidata.org/wiki/Q189833","display_name":"Economic growth","level":1,"score":0.0},{"id":"https://openalex.org/C13280743","wikidata":"https://www.wikidata.org/wiki/Q131089","display_name":"Geodesy","level":1,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.0},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C205649164","wikidata":"https://www.wikidata.org/wiki/Q1071","display_name":"Geography","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"mag:2996693845","doi":null,"is_oa":false,"landing_page_url":"https://www.openreview.net/pdf?id=SkxJ8REYPH","pdf_url":null,"source":{"id":"https://openalex.org/S4306419637","display_name":"International Conference on Learning Representations","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"International Conference on Learning Representations","raw_type":null}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":41,"referenced_works":["https://openalex.org/W104184427","https://openalex.org/W2138302120","https://openalex.org/W2194775991","https://openalex.org/W2336650964","https://openalex.org/W2405883473","https://openalex.org/W2436219157","https://openalex.org/W2541884796","https://openalex.org/W2617766261","https://openalex.org/W2622263826","https://openalex.org/W2626778328","https://openalex.org/W2744599870","https://openalex.org/W2765170276","https://openalex.org/W2769644379","https://openalex.org/W2780752111","https://openalex.org/W2792630208","https://openalex.org/W2799269579","https://openalex.org/W2803867449","https://openalex.org/W2806311723","https://openalex.org/W2888206291","https://openalex.org/W2888561381","https://openalex.org/W2896101215","https://openalex.org/W2900182564","https://openalex.org/W2902410171","https://openalex.org/W2906936451","https://openalex.org/W2911327153","https://openalex.org/W2911863041","https://openalex.org/W2913357890","https://openalex.org/W2944542720","https://openalex.org/W2946900926","https://openalex.org/W2947027751","https://openalex.org/W2949270253","https://openalex.org/W2963228337","https://openalex.org/W2963712608","https://openalex.org/W2964024720","https://openalex.org/W2964121744","https://openalex.org/W2965373594","https://openalex.org/W2970803838","https://openalex.org/W2970948121","https://openalex.org/W2989361410","https://openalex.org/W2996118901","https://openalex.org/W3038057116"],"related_works":["https://openalex.org/W2977517840","https://openalex.org/W2944542720","https://openalex.org/W4388903782","https://openalex.org/W3206218057","https://openalex.org/W2991209658","https://openalex.org/W2946270031","https://openalex.org/W7102404357","https://openalex.org/W4386804004","https://openalex.org/W3118543361","https://openalex.org/W3113081714"],"abstract_inverted_index":{"Distributed":[0],"optimization":[1,106,123],"is":[2,137,147,175],"essential":[3],"for":[4,193],"training":[5],"large":[6,9],"models":[7,63],"on":[8,109,148],"datasets.":[10],"Multiple":[11],"approaches":[12],"have":[13],"been":[14],"proposed":[15],"to":[16,41,128,165,187],"reduce":[17],"the":[18,61,69,76,129,134,144,153,180,188],"communication":[19,57],"overhead":[20,136],"in":[21,122],"distributed":[22],"training,":[23],"such":[24],"as":[25],"synchronizing":[26],"only":[27],"after":[28,68,100],"performing":[29],"multiple":[30,101],"local":[31],"SGD":[32],"steps,":[33],"and":[34,95,112,124],"decentralized":[35],"methods":[36,48],"(e.g.,":[37],"using":[38],"gossip":[39],"algorithms)":[40],"decouple":[42],"communications":[43],"among":[44],"workers.":[45],"Although":[46],"these":[47],"run":[49],"faster":[50],"than":[51],"AllReduce-based":[52],"methods,":[53],"which":[54],"use":[55],"blocking":[56],"before":[58],"every":[59],"update,":[60,99],"resulting":[62],"may":[64],"be":[65],"less":[66],"accurate":[67],"same":[70],"number":[71],"of":[72,79,103,152,169,179],"updates.":[73],"Inspired":[74],"by":[75],"BMUF":[77,174],"method":[78],"Chen":[80],"&":[81],"Huo":[82],"(2016),":[83],"we":[84],"propose":[85],"a":[86,97,104,166,176],"slow":[87],"momentum":[88,98],"(SloMo)":[89],"framework,":[90,182],"where":[91],"workers":[92],"periodically":[93],"synchronize":[94],"perform":[96],"iterations":[102],"base":[105,130,154],"algorithm.":[107],"Experiments":[108],"image":[110],"classification":[111],"machine":[113],"translation":[114],"tasks":[115],"demonstrate":[116],"that":[117,143,151,162],"SloMo":[118,145,163,181],"consistently":[119],"yields":[120],"improvements":[121],"generalization":[125],"performance":[126],"relative":[127],"optimizer,":[131],"even":[132],"when":[133],"additional":[135],"amortized":[138],"over":[139],"many":[140],"updates":[141],"so":[142],"runtime":[146],"par":[149],"with":[150],"optimizer.":[155],"We":[156],"provide":[157],"theoretical":[158,190],"convergence":[159,191],"guarantees":[160,192],"showing":[161],"converges":[164],"stationary":[167],"point":[168],"smooth":[170],"non-convex":[171],"losses.":[172],"Since":[173],"particular":[177],"instance":[178],"our":[183],"results":[184],"also":[185],"correspond":[186],"first":[189],"BMUF.":[194]},"counts_by_year":[{"year":2022,"cited_by_count":1},{"year":2021,"cited_by_count":20},{"year":2020,"cited_by_count":12},{"year":2019,"cited_by_count":1}],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2019-12-26T00:00:00"}