{"id":"https://openalex.org/W3016159127","doi":"https://doi.org/10.1109/icassp40776.2020.9054065","title":"Improving Efficiency in Large-Scale Decentralized Distributed Training","display_name":"Improving Efficiency in Large-Scale Decentralized Distributed Training","publication_year":2020,"publication_date":"2020-04-09","ids":{"openalex":"https://openalex.org/W3016159127","doi":"https://doi.org/10.1109/icassp40776.2020.9054065","mag":"3016159127"},"language":"en","primary_location":{"id":"doi:10.1109/icassp40776.2020.9054065","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp40776.2020.9054065","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"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/A5101541045","display_name":"Wei Zhang","orcid":"https://orcid.org/0000-0001-6763-8146"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Wei Zhang","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102014291","display_name":"Xiaodong Cui","orcid":"https://orcid.org/0000-0003-4865-1307"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaodong Cui","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043483012","display_name":"Abdullah Kayi","orcid":"https://orcid.org/0000-0001-5909-9891"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Abdullah Kayi","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101839271","display_name":"Mingrui Liu","orcid":"https://orcid.org/0000-0002-6499-7212"},"institutions":[{"id":"https://openalex.org/I126307644","display_name":"University of Iowa","ror":"https://ror.org/036jqmy94","country_code":"US","type":"education","lineage":["https://openalex.org/I126307644"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mingrui Liu","raw_affiliation_strings":["University of Iowa"],"affiliations":[{"raw_affiliation_string":"University of Iowa","institution_ids":["https://openalex.org/I126307644"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000286196","display_name":"Ulrich Finkler","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ulrich Finkler","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003725957","display_name":"Brian Kingsbury","orcid":"https://orcid.org/0000-0002-1343-6837"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Brian Kingsbury","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5079994647","display_name":"George Saon","orcid":"https://orcid.org/0009-0004-6837-5009"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"George Saon","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064636177","display_name":"Youssef Mroueh","orcid":"https://orcid.org/0000-0001-8798-1267"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Youssef Mroueh","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5044824009","display_name":"Alper Buyuktosunoglu","orcid":"https://orcid.org/0000-0002-5341-8916"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Alper Buyuktosunoglu","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064335115","display_name":"Payel Das","orcid":"https://orcid.org/0000-0002-7288-0516"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Payel Das","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5063522852","display_name":"David S. Kung","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"David Kung","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5034529775","display_name":"Michael Picheny","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Michael Picheny","raw_affiliation_strings":["IBM Research"],"affiliations":[{"raw_affiliation_string":"IBM Research","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":12,"corresponding_author_ids":["https://openalex.org/A5101541045"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.3256,"has_fulltext":false,"cited_by_count":13,"citation_normalized_percentile":{"value":0.84482393,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"3022","last_page":"3026"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9983000159263611,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9983000159263611,"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/T10860","display_name":"Speech and Audio Processing","score":0.9965999722480774,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/T12676","display_name":"Machine Learning and ELM","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/computer-science","display_name":"Computer science","score":0.8390027284622192},{"id":"https://openalex.org/keywords/asynchronous-communication","display_name":"Asynchronous communication","score":0.8248833417892456},{"id":"https://openalex.org/keywords/supercomputer","display_name":"Supercomputer","score":0.6670546531677246},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6280051469802856},{"id":"https://openalex.org/keywords/ibm","display_name":"IBM","score":0.5457819700241089},{"id":"https://openalex.org/keywords/test-set","display_name":"Test set","score":0.5304538607597351},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.5176406502723694},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.5048235058784485},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.4975006878376007},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3965589702129364},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.39558374881744385},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.33943408727645874},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.08765101432800293}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8390027284622192},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.8248833417892456},{"id":"https://openalex.org/C83283714","wikidata":"https://www.wikidata.org/wiki/Q121117","display_name":"Supercomputer","level":2,"score":0.6670546531677246},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6280051469802856},{"id":"https://openalex.org/C70388272","wikidata":"https://www.wikidata.org/wiki/Q5968558","display_name":"IBM","level":2,"score":0.5457819700241089},{"id":"https://openalex.org/C169903167","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Test set","level":2,"score":0.5304538607597351},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.5176406502723694},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.5048235058784485},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.4975006878376007},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3965589702129364},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.39558374881744385},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.33943408727645874},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.08765101432800293},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C187736073","wikidata":"https://www.wikidata.org/wiki/Q2920921","display_name":"Management","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/C50522688","wikidata":"https://www.wikidata.org/wiki/Q189833","display_name":"Economic growth","level":1,"score":0.0},{"id":"https://openalex.org/C171250308","wikidata":"https://www.wikidata.org/wiki/Q11468","display_name":"Nanotechnology","level":1,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp40776.2020.9054065","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp40776.2020.9054065","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":31,"referenced_works":["https://openalex.org/W1442374986","https://openalex.org/W2057332538","https://openalex.org/W2094233035","https://openalex.org/W2168231600","https://openalex.org/W2193413348","https://openalex.org/W2194775991","https://openalex.org/W2336650964","https://openalex.org/W2405883473","https://openalex.org/W2407022425","https://openalex.org/W2617960902","https://openalex.org/W2622263826","https://openalex.org/W2902410171","https://openalex.org/W2936012302","https://openalex.org/W2962863496","https://openalex.org/W2962907457","https://openalex.org/W2962950660","https://openalex.org/W2963228337","https://openalex.org/W2973188344","https://openalex.org/W4299802343","https://openalex.org/W6628377381","https://openalex.org/W6684859321","https://openalex.org/W6687483927","https://openalex.org/W6687566353","https://openalex.org/W6692488971","https://openalex.org/W6703420464","https://openalex.org/W6738250615","https://openalex.org/W6739622702","https://openalex.org/W6745458143","https://openalex.org/W6756976665","https://openalex.org/W6760911985","https://openalex.org/W6780043775"],"related_works":["https://openalex.org/W3126131865","https://openalex.org/W2384867379","https://openalex.org/W2329539859","https://openalex.org/W4253186488","https://openalex.org/W2227905990","https://openalex.org/W2765823764","https://openalex.org/W3214280620","https://openalex.org/W2044344400","https://openalex.org/W2251285835","https://openalex.org/W1599154403"],"abstract_inverted_index":{"Decentralized":[0],"Parallel":[1,9],"SGD":[2,10],"(D-PSGD)":[3],"and":[4,96,130,143,152],"its":[5],"asynchronous":[6],"variant":[7],"Asynchronous":[8],"(AD-PSGD)":[11],"is":[12,35,109],"a":[13],"family":[14],"of":[15,33,40,48,82],"distributed":[16],"learning":[17,29],"algorithms":[18],"that":[19,36],"have":[20],"been":[21],"demonstrated":[22],"to":[23,63,111,166],"perform":[24],"well":[25],"for":[26],"large-scale":[27],"deep":[28],"tasks.":[30],"One":[31],"drawback":[32],"(A)D-PSGD":[34,65],"the":[37,41,46,51,70,75,80,90,97,124,134,161],"spectral":[38,71],"gap":[39,72],"mixing":[42],"matrix":[43],"decreases":[44],"when":[45],"number":[47],"learners":[49],"in":[50,117,144],"system":[52,108],"increases,":[53],"which":[54],"hampers":[55],"convergence.":[56],"In":[57],"this":[58],"paper,":[59],"we":[60],"investigate":[61],"techniques":[62,85],"accelerate":[64],"based":[66],"training":[67,163],"by":[68,86],"improving":[69],"while":[73],"minimizing":[74],"communication":[76],"cost.":[77],"We":[78],"demonstrate":[79],"effectiveness":[81],"our":[83,107],"proposed":[84],"running":[87],"experiments":[88],"on":[89,123,133,150,155],"2000-hour":[91],"Switchboard":[92,126],"speech":[93],"recognition":[94],"task":[95],"ImageNet":[98],"computer":[99],"vision":[100],"task.":[101],"On":[102],"an":[103,113],"IBM":[104],"P9":[105],"supercomputer,":[106],"able":[110],"train":[112],"LSTM":[114],"acoustic":[115],"model":[116],"2.28":[118],"hours":[119,146],"with":[120,147],"7.5%":[121],"WER":[122,132,149,154],"Hub5-2000":[125],"(SWB)":[127],"test":[128,137],"set":[129,138],"13.3%":[131,153],"CallHome":[135],"(CH)":[136],"using":[139,157],"64":[140],"V100":[141,159],"GPUs":[142],"1.98":[145],"7.7%":[148],"SWB":[151],"CH":[156],"128":[158],"GPUs,":[160],"fastest":[162],"time":[164],"reported":[165],"date.":[167]},"counts_by_year":[{"year":2024,"cited_by_count":2},{"year":2022,"cited_by_count":3},{"year":2021,"cited_by_count":3},{"year":2020,"cited_by_count":4},{"year":2019,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}