{"id":"https://openalex.org/W2966527647","doi":"https://doi.org/10.24963/ijcai.2019/473","title":"A Convergence Analysis of Distributed SGD with Communication-Efficient Gradient Sparsification","display_name":"A Convergence Analysis of Distributed SGD with Communication-Efficient Gradient Sparsification","publication_year":2019,"publication_date":"2019-07-28","ids":{"openalex":"https://openalex.org/W2966527647","doi":"https://doi.org/10.24963/ijcai.2019/473","mag":"2966527647"},"language":"en","primary_location":{"id":"doi:10.24963/ijcai.2019/473","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2019/473","pdf_url":"https://www.ijcai.org/proceedings/2019/0473.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://www.ijcai.org/proceedings/2019/0473.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5016836702","display_name":"Shaohuai Shi","orcid":"https://orcid.org/0000-0002-1418-5160"},"institutions":[{"id":"https://openalex.org/I141568987","display_name":"Hong Kong Baptist University","ror":"https://ror.org/0145fw131","country_code":"HK","type":"education","lineage":["https://openalex.org/I141568987"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Shaohuai Shi","raw_affiliation_strings":["Department of Computer Science, Hong Kong Baptist University"],"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Hong Kong Baptist University","institution_ids":["https://openalex.org/I141568987"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100675370","display_name":"Kaiyong Zhao","orcid":null},"institutions":[{"id":"https://openalex.org/I141568987","display_name":"Hong Kong Baptist University","ror":"https://ror.org/0145fw131","country_code":"HK","type":"education","lineage":["https://openalex.org/I141568987"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Kaiyong Zhao","raw_affiliation_strings":["Department of Computer Science, Hong Kong Baptist University"],"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Hong Kong Baptist University","institution_ids":["https://openalex.org/I141568987"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018867983","display_name":"Qiang Wang","orcid":"https://orcid.org/0000-0002-2986-967X"},"institutions":[{"id":"https://openalex.org/I141568987","display_name":"Hong Kong Baptist University","ror":"https://ror.org/0145fw131","country_code":"HK","type":"education","lineage":["https://openalex.org/I141568987"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Qiang Wang","raw_affiliation_strings":["Department of Computer Science, Hong Kong Baptist University"],"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Hong Kong Baptist University","institution_ids":["https://openalex.org/I141568987"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048313055","display_name":"Zhenheng Tang","orcid":"https://orcid.org/0000-0001-8769-9974"},"institutions":[{"id":"https://openalex.org/I141568987","display_name":"Hong Kong Baptist University","ror":"https://ror.org/0145fw131","country_code":"HK","type":"education","lineage":["https://openalex.org/I141568987"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Zhenheng Tang","raw_affiliation_strings":["Department of Computer Science, Hong Kong Baptist University"],"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Hong Kong Baptist University","institution_ids":["https://openalex.org/I141568987"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100730785","display_name":"Xiaowen Chu","orcid":"https://orcid.org/0000-0001-9745-4372"},"institutions":[{"id":"https://openalex.org/I141568987","display_name":"Hong Kong Baptist University","ror":"https://ror.org/0145fw131","country_code":"HK","type":"education","lineage":["https://openalex.org/I141568987"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xiaowen Chu","raw_affiliation_strings":["Department of Computer Science, Hong Kong Baptist University"],"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Hong Kong Baptist University","institution_ids":["https://openalex.org/I141568987"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5016836702"],"corresponding_institution_ids":["https://openalex.org/I141568987"],"apc_list":null,"apc_paid":null,"fwci":5.1806,"has_fulltext":true,"cited_by_count":73,"citation_normalized_percentile":{"value":0.963937,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":96,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"3411","last_page":"3417"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9998000264167786,"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":0.9998000264167786,"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/T10500","display_name":"Sparse and Compressive Sensing Techniques","score":0.9987000226974487,"subfield":{"id":"https://openalex.org/subfields/2206","display_name":"Computational Mechanics"},"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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.9962999820709229,"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/scalability","display_name":"Scalability","score":0.7194088697433472},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"Stochastic gradient descent","score":0.6666392087936401},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6662493348121643},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.6650254726409912},{"id":"https://openalex.org/keywords/rate-of-convergence","display_name":"Rate of convergence","score":0.6046538352966309},{"id":"https://openalex.org/keywords/mathematical-proof","display_name":"Mathematical proof","score":0.5876348614692688},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.5492246150970459},{"id":"https://openalex.org/keywords/communication-complexity","display_name":"Communication complexity","score":0.5334793329238892},{"id":"https://openalex.org/keywords/soundness","display_name":"Soundness","score":0.5069120526313782},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.47395849227905273},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.4381639063358307},{"id":"https://openalex.org/keywords/theoretical-computer-science","display_name":"Theoretical computer science","score":0.42125001549720764},{"id":"https://openalex.org/keywords/mathematical-optimization","display_name":"Mathematical optimization","score":0.349259614944458},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.2727562189102173},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.15360718965530396},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.08015471696853638},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.07348161935806274}],"concepts":[{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7194088697433472},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.6666392087936401},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6662493348121643},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.6650254726409912},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.6046538352966309},{"id":"https://openalex.org/C108710211","wikidata":"https://www.wikidata.org/wiki/Q11538","display_name":"Mathematical proof","level":2,"score":0.5876348614692688},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.5492246150970459},{"id":"https://openalex.org/C179145077","wikidata":"https://www.wikidata.org/wiki/Q5154130","display_name":"Communication complexity","level":2,"score":0.5334793329238892},{"id":"https://openalex.org/C39920170","wikidata":"https://www.wikidata.org/wiki/Q693083","display_name":"Soundness","level":2,"score":0.5069120526313782},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.47395849227905273},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.4381639063358307},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.42125001549720764},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.349259614944458},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2727562189102173},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.15360718965530396},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.08015471696853638},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.07348161935806274},{"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/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","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/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"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/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.24963/ijcai.2019/473","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2019/473","pdf_url":"https://www.ijcai.org/proceedings/2019/0473.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-105130","is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-105130","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Conference paper"},{"id":"pmh:oai:repository.ust.hk:1783.1-105130","is_oa":false,"landing_page_url":"http://www.scopus.com/record/display.url?eid=2-s2.0-85074915733&origin=inward","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"Conference paper"}],"best_oa_location":{"id":"doi:10.24963/ijcai.2019/473","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2019/473","pdf_url":"https://www.ijcai.org/proceedings/2019/0473.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320320955","display_name":"Hong Kong Baptist University","ror":"https://ror.org/0145fw131"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W2966527647.pdf","grobid_xml":"https://content.openalex.org/works/W2966527647.grobid-xml"},"referenced_works_count":37,"referenced_works":["https://openalex.org/W1632114991","https://openalex.org/W1686810756","https://openalex.org/W2130062883","https://openalex.org/W2157590573","https://openalex.org/W2168231600","https://openalex.org/W2194775991","https://openalex.org/W2436219157","https://openalex.org/W2606891064","https://openalex.org/W2617766261","https://openalex.org/W2622263826","https://openalex.org/W2626721309","https://openalex.org/W2766140019","https://openalex.org/W2769644379","https://openalex.org/W2771016179","https://openalex.org/W2774000609","https://openalex.org/W2775776326","https://openalex.org/W2803965032","https://openalex.org/W2805997383","https://openalex.org/W2884711234","https://openalex.org/W2891735383","https://openalex.org/W2891970177","https://openalex.org/W2920397365","https://openalex.org/W2946270031","https://openalex.org/W2962758826","https://openalex.org/W2963433607","https://openalex.org/W2963766684","https://openalex.org/W2964163156","https://openalex.org/W2964324519","https://openalex.org/W2982475424","https://openalex.org/W3101036738","https://openalex.org/W4294527468","https://openalex.org/W4297685247","https://openalex.org/W6641578900","https://openalex.org/W6791858558","https://openalex.org/W6863631769","https://openalex.org/W6863994431","https://openalex.org/W6864014924"],"related_works":["https://openalex.org/W1511249877","https://openalex.org/W1578828174","https://openalex.org/W2950376466","https://openalex.org/W2950202488","https://openalex.org/W2604424424","https://openalex.org/W2964170259","https://openalex.org/W3002546633","https://openalex.org/W4206119629","https://openalex.org/W2765682467","https://openalex.org/W4382937879"],"abstract_inverted_index":{"Gradient":[0],"sparsification":[1,24,56,84],"is":[2,37,48,123],"a":[3,30],"promising":[4],"technique":[5],"to":[6,61,68,146],"significantly":[7,72],"reduce":[8,62],"the":[9,38,49,54,63,74,82,99,102,116,125,129,148,151,158,161,165],"communication":[10,31,64],"overhead":[11],"in":[12,88],"decentralized":[13],"synchronous":[14],"stochastic":[15],"gradient":[16,23],"descent":[17],"(S-SGD)":[18],"algorithms.":[19],"Yet,":[20],"many":[21],"existing":[22],"schemes":[25],"(e.g.,":[26],"Top-k":[27],"sparsification)":[28],"have":[29],"complexity":[32,65],"of":[33,40,51,101,119,150,160],"O(kP),":[34],"where":[35],"k":[36],"number":[39,50],"selected":[41],"gradients":[42],"by":[43],"each":[44],"worker":[45],"and":[46,143,153,156],"P":[47],"workers.":[52],"Recently,":[53],"gTop-k":[55,83,103,120],"scheme":[57,85,104],"has":[58],"been":[59],"proposed":[60],"from":[66],"O(kP)":[67],"O(k":[69],"logP),":[70],"which":[71,122],"boosts":[73],"system":[75],"scalability.":[76],"However,":[77],"it":[78],"remains":[79],"unclear":[80],"whether":[81],"can":[86],"converge":[87],"theory.":[89],"In":[90],"this":[91],"paper,":[92],"we":[93,134],"first":[94],"provide":[95],"theoretical":[96,154],"proofs":[97],"on":[98,138,164],"convergence":[100,117,166],"for":[105],"non-convex":[106],"objective":[107],"functions":[108],"under":[109],"certain":[110],"analytic":[111],"assumptions.":[112],"We":[113],"then":[114],"derive":[115],"rate":[118],"S-SGD,":[121],"at":[124],"same":[126],"order":[127],"as":[128],"vanilla":[130],"mini-batch":[131],"SGD.":[132],"Finally,":[133],"conduct":[135],"extensive":[136],"experiments":[137],"different":[139],"machine":[140],"learning":[141],"models":[142],"data":[144],"sets":[145],"verify":[147],"soundness":[149],"assumptions":[152],"results,":[155],"discuss":[157],"impact":[159],"compression":[162],"ratio":[163],"performance.":[167]},"counts_by_year":[{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":11},{"year":2023,"cited_by_count":19},{"year":2022,"cited_by_count":8},{"year":2021,"cited_by_count":16},{"year":2020,"cited_by_count":10},{"year":2019,"cited_by_count":3}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}