{"id":"https://openalex.org/W2914138317","doi":"https://doi.org/10.48550/arxiv.1901.09269","title":"Distributed Learning with Compressed Gradient Differences","display_name":"Distributed Learning with Compressed Gradient Differences","publication_year":2019,"publication_date":"2019-01-26","ids":{"openalex":"https://openalex.org/W2914138317","doi":"https://doi.org/10.48550/arxiv.1901.09269","mag":"2914138317"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:1901.09269","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/1901.09269","pdf_url":"https://arxiv.org/pdf/1901.09269","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"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":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/1901.09269","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5015808280","display_name":"Konstantin Mishchenko","orcid":"https://orcid.org/0000-0002-5241-7292","observed_orcids":["https://orcid.org/0000-0002-5241-7292"]},"institutions":[{"id":"https://openalex.org/I1326498283","display_name":"Institut national de recherche en sciences et technologies du num\u00e9rique","ror":"https://ror.org/02kvxyf05","country_code":"FR","type":"government","lineage":["https://openalex.org/I1326498283"]},{"id":"https://openalex.org/I71920554","display_name":"King Abdullah University of Science and Technology","ror":"https://ror.org/01q3tbs38","country_code":"SA","type":"education","lineage":["https://openalex.org/I71920554"]}],"countries":["FR","SA"],"is_corresponding":false,"raw_author_name":"Mishchenko, Konstantin","raw_affiliation_strings":["INRIA , Paris , France;","KING ABDULLAH UNIVERSITY OF SCIENCE AND TECHNOLOGY"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"INRIA , Paris , France;","institution_ids":["https://openalex.org/I1326498283"]},{"raw_affiliation_string":"KING ABDULLAH UNIVERSITY OF SCIENCE AND TECHNOLOGY","institution_ids":["https://openalex.org/I71920554"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087594198","display_name":"Eduard Gorbunov","orcid":"https://orcid.org/0000-0002-3370-4130","observed_orcids":["https://orcid.org/0000-0002-3370-4130"]},"institutions":[{"id":"https://openalex.org/I153845743","display_name":"Moscow Institute of Physics and Technology","ror":"https://ror.org/00v0z9322","country_code":"RU","type":"education","lineage":["https://openalex.org/I153845743"]},{"id":"https://openalex.org/I4210113480","display_name":"Mohamed bin Zayed University of Artificial Intelligence","ror":"https://ror.org/0258gkt32","country_code":"AE","type":"education","lineage":["https://openalex.org/I4210113480"]}],"countries":["AE","RU"],"is_corresponding":false,"raw_author_name":"Gorbunov, Eduard","raw_affiliation_strings":["Mohamed bin Zayed University of Artificial Intelligence , Abu Dhabi , UAE;","Moscow Institute of Physics and Technology , Moscow , Russia","Moscow Institute of Physics and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Mohamed bin Zayed University of Artificial Intelligence , Abu Dhabi , UAE;","institution_ids":["https://openalex.org/I4210113480"]},{"raw_affiliation_string":"Moscow Institute of Physics and Technology , Moscow , Russia","institution_ids":["https://openalex.org/I153845743"]},{"raw_affiliation_string":"Moscow Institute of Physics and Technology","institution_ids":["https://openalex.org/I153845743"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070679093","display_name":"Martin Tak\u00e1\u010d","orcid":"https://orcid.org/0000-0001-7455-2025","observed_orcids":["https://orcid.org/0000-0001-7455-2025"]},"institutions":[{"id":"https://openalex.org/I186143895","display_name":"Lehigh University","ror":"https://ror.org/012afjb06","country_code":"US","type":"education","lineage":["https://openalex.org/I186143895"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tak\u00e1\u010d, Martin","raw_affiliation_strings":["Lehigh University,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Lehigh University,","institution_ids":["https://openalex.org/I186143895"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5036598221","display_name":"Peter Richt\u00e1rik","orcid":"https://orcid.org/0000-0003-4380-5848","observed_orcids":["https://orcid.org/0000-0003-4380-5848"]},"institutions":[{"id":"https://openalex.org/I71920554","display_name":"King Abdullah University of Science and Technology","ror":"https://ror.org/01q3tbs38","country_code":"SA","type":"education","lineage":["https://openalex.org/I71920554"]}],"countries":["SA"],"is_corresponding":false,"raw_author_name":"Richt\u00e1rik, Peter","raw_affiliation_strings":["KING ABDULLAH UNIVERSITY OF SCIENCE AND TECHNOLOGY"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"KING ABDULLAH UNIVERSITY OF SCIENCE AND TECHNOLOGY","institution_ids":["https://openalex.org/I71920554"]}]}],"institutions":[],"countries_distinct_count":5,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":98,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"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":0.0005000000237487257,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10764","display_name":"Privacy-Preserving Technologies in Data","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/distributed-learning","display_name":"distributed learning","score":0.8320000171661377},{"id":"https://openalex.org/keywords/gradient-compression","display_name":"gradient compression","score":0.8659999966621399},{"id":"https://openalex.org/keywords/gradient-difference","display_name":"gradient difference","score":0.7720000147819519},{"id":"https://openalex.org/keywords/diana","display_name":"Diana","score":0.9729999899864197},{"id":"https://openalex.org/keywords/communication-efficiency","display_name":"communication efficiency","score":0.5590000152587891},{"id":"https://openalex.org/keywords/quantization","display_name":"quantization","score":0.6069999933242798},{"id":"https://openalex.org/keywords/sparsification","display_name":"sparsification","score":0.8090000152587891},{"id":"https://openalex.org/keywords/convergence-rate","display_name":"convergence rate","score":0.6299999952316284},{"id":"https://openalex.org/keywords/strongly-convex-optimization","display_name":"strongly convex optimization","score":0.8330000042915344},{"id":"https://openalex.org/keywords/nonconvex-optimization","display_name":"non-convex optimization","score":0.9779999852180481},{"id":"https://openalex.org/keywords/signsgd","display_name":"signSGD","score":0.9700000286102295}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.8199840188026428},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.706642746925354},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6318563222885132},{"id":"https://openalex.org/C112680207","wikidata":"https://www.wikidata.org/wiki/Q714886","display_name":"Regular polygon","level":2,"score":0.5495365262031555},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.5302015542984009},{"id":"https://openalex.org/C2779582901","wikidata":"https://www.wikidata.org/wiki/Q21013010","display_name":"Distributed learning","level":2,"score":0.5053413510322571},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4688141345977783},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3582877218723297},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.32577353715896606},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.30288803577423096},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.07852676510810852},{"id":"https://openalex.org/C19417346","wikidata":"https://www.wikidata.org/wiki/Q7922","display_name":"Pedagogy","level":1,"score":0.0},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","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/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.0},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"pmh:oai:arXiv.org:1901.09269","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/1901.09269","pdf_url":"https://arxiv.org/pdf/1901.09269","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"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":"text"},{"id":"pmh:oai:repository.kaust.edu.sa:10754/653106","doi":null,"is_oa":true,"landing_page_url":"http://hdl.handle.net/10754/653106","pdf_url":"http://hdl.handle.net/10754/653106","source":{"id":"https://openalex.org/S4306401596","display_name":"King Abdullah University of Science and Technology Repository (King Abdullah University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I71920554","host_organization_name":"King Abdullah University of Science and Technology","host_organization_lineage":["https://openalex.org/I71920554"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"Preprint"},{"id":"doi:10.48550/arxiv.1901.09269","doi":"https://doi.org/10.48550/arxiv.1901.09269","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.1901.09269","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:1901.09269","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/1901.09269","pdf_url":"https://arxiv.org/pdf/1901.09269","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"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":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G2928192979","display_name":null,"funder_award_id":"000000D730321P5Q0002","funder_id":"https://openalex.org/F4320319755","funder_display_name":"Analytical Center for the Government of the Russian Federation"},{"id":"https://openalex.org/G3041188381","display_name":"Collaborative Research:  TRIPODS Institute for Optimization and Learning","funder_award_id":"1740796","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G3736275146","display_name":null,"funder_award_id":"70-2021-00138","funder_id":"https://openalex.org/F4320319755","funder_display_name":"Analytical Center for the Government of the Russian Federation"},{"id":"https://openalex.org/G5479629224","display_name":"GOALI: Machine Learning Approaches for Supply Chain Decision-Making","funder_award_id":"1663256","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G5927019889","display_name":null,"funder_award_id":"70-2021-00138","funder_id":"https://openalex.org/F4320330091","funder_display_name":"Moscow Institute of Physics and Technology"},{"id":"https://openalex.org/G8084227899","display_name":"AF: Small: New classes of optimization methods for nonconvex large scale machine learning models.","funder_award_id":"1618717","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320319755","display_name":"Analytical Center for the Government of the Russian Federation","ror":"https://ror.org/01zwzc219"},{"id":"https://openalex.org/F4320330091","display_name":"Moscow Institute of Physics and Technology","ror":"https://ror.org/00v0z9322"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W2914138317.pdf","grobid_xml":"https://content.openalex.org/works/W2914138317.grobid-xml"},"referenced_works_count":18,"referenced_works":["https://openalex.org/W1697545848","https://openalex.org/W1845160311","https://openalex.org/W2143021454","https://openalex.org/W2165905842","https://openalex.org/W2199097987","https://openalex.org/W2407022425","https://openalex.org/W2510078447","https://openalex.org/W2535838896","https://openalex.org/W2617766261","https://openalex.org/W2769644379","https://openalex.org/W2808495300","https://openalex.org/W2898280316","https://openalex.org/W2950308151","https://openalex.org/W2952986481","https://openalex.org/W2962741697","https://openalex.org/W2963664311","https://openalex.org/W2963992805","https://openalex.org/W2966396637"],"related_works":["https://openalex.org/W2595172197","https://openalex.org/W2084856301","https://openalex.org/W2127970246","https://openalex.org/W4382618745","https://openalex.org/W2885125400","https://openalex.org/W1001352512","https://openalex.org/W1989889224","https://openalex.org/W1973775000","https://openalex.org/W2114711060","https://openalex.org/W2790129348"],"abstract_inverted_index":{"Training":[0],"large":[1],"machine":[2],"learning":[3,92],"models":[4],"requires":[5],"a":[6,89,108],"distributed":[7,91],"computing":[8],"approach,":[9],"with":[10],"communication":[11],"of":[12,32,61,74,103,144],"the":[13,17,26,68,77,81,112,137,154,170],"model":[14],"updates":[15,33],"being":[16],"bottleneck.":[18],"For":[19],"this":[20,85,99,175],"reason,":[21],"several":[22],"methods":[23,63],"based":[24],"on":[25],"compression":[27,102],"(e.g.,":[28],"sparsification":[29],"and/or":[30],"quantization)":[31],"were":[34],"recently":[35],"proposed,":[36],"including":[37],"QSGD":[38],"(Alistarh":[39],"et":[40,45,50,56],"al.,":[41,46,51,57],"2017),":[42,47],"TernGrad":[43],"(Wen":[44],"SignSGD":[48],"(Bernstein":[49],"2018),":[52],"and":[53,115,118,146,150,158],"DQGD":[54],"(Khirirat":[55],"2018).":[58],"However,":[59],"none":[60],"these":[62],"are":[64,123],"able":[65],"to":[66,76,125,132,166],"learn":[67],"gradients,":[69],"which":[70,97],"renders":[71],"them":[72],"incapable":[73],"converging":[75],"true":[78],"optimum":[79],"in":[80,111,156],"batch":[82],"mode.":[83],"In":[84],"work":[86],"we":[87,168],"propose":[88],"new":[90],"method":[93],"--":[94,96],"DIANA":[95],"resolves":[98],"issue":[100],"via":[101],"gradient":[104],"differences.":[105],"We":[106,128],"perform":[107],"theoretical":[109],"analysis":[110,143,164],"strongly":[113],"convex":[114],"nonconvex":[116],"settings":[117],"show":[119],"that":[120],"our":[121,163],"rates":[122],"superior":[124],"existing":[126],"rates.":[127],"also":[129],"provide":[130],"theory":[131,157],"support":[133],"non-smooth":[134],"regularizers":[135],"study":[136],"difference":[138],"between":[139,148],"quantization":[140,152],"schemes.":[141],"Our":[142],"block-quantization":[145],"differences":[147],"$\\ell_2$":[149],"$\\ell_{\\infty}$":[151],"closes":[153],"gaps":[155],"practice.":[159],"Finally,":[160],"by":[161],"applying":[162],"technique":[165],"TernGrad,":[167],"establish":[169],"first":[171],"convergence":[172],"rate":[173],"for":[174],"method.":[176]},"counts_by_year":[{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":12},{"year":2023,"cited_by_count":12},{"year":2022,"cited_by_count":10},{"year":2021,"cited_by_count":29},{"year":2020,"cited_by_count":18},{"year":2019,"cited_by_count":13}],"updated_date":"2026-10-16T09:59:22.235173","created_date":"2025-10-10T00:00:00"}