{"id":"https://openalex.org/W3113081714","doi":"https://doi.org/10.1145/3563038","title":"SSD-SGD: Communication Sparsification for Distributed Deep Learning Training","display_name":"SSD-SGD: Communication Sparsification for Distributed Deep Learning Training","publication_year":2022,"publication_date":"2022-09-14","ids":{"openalex":"https://openalex.org/W3113081714","doi":"https://doi.org/10.1145/3563038","mag":"3113081714"},"language":"en","primary_location":{"id":"doi:10.1145/3563038","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3563038","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3563038","source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3563038","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5042353509","display_name":"Yemao Xu","orcid":"https://orcid.org/0000-0003-3639-5449"},"institutions":[{"id":"https://openalex.org/I2802541053","display_name":"Academy of Military Medical Sciences","ror":"https://ror.org/02bv3c993","country_code":"CN","type":"facility","lineage":["https://openalex.org/I2802541053"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Yemao Xu","raw_affiliation_strings":["Defense Innovation Institute, Academy of Military Sciences, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0003-3639-5449","affiliations":[{"raw_affiliation_string":"Defense Innovation Institute, Academy of Military Sciences, Beijing, China","institution_ids":["https://openalex.org/I2802541053"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006729432","display_name":"Dezun Dong","orcid":"https://orcid.org/0000-0001-6243-8479"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dezun Dong","raw_affiliation_strings":["National University of Defense Technology, Changsha, Hunan, China"],"raw_orcid":"https://orcid.org/0000-0001-6243-8479","affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, Hunan, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Dongsheng Wang","orcid":"https://orcid.org/0000-0003-3192-7493"},"institutions":[{"id":"https://openalex.org/I2802541053","display_name":"Academy of Military Medical Sciences","ror":"https://ror.org/02bv3c993","country_code":"CN","type":"facility","lineage":["https://openalex.org/I2802541053"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dongsheng Wang","raw_affiliation_strings":["Defense Innovation Institute, Academy of Military Sciences, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0003-3192-7493","affiliations":[{"raw_affiliation_string":"Defense Innovation Institute, Academy of Military Sciences, Beijing, China","institution_ids":["https://openalex.org/I2802541053"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102534602","display_name":"Weixia Xu","orcid":null},"institutions":[{"id":"https://openalex.org/I2802541053","display_name":"Academy of Military Medical Sciences","ror":"https://ror.org/02bv3c993","country_code":"CN","type":"facility","lineage":["https://openalex.org/I2802541053"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shi Xu","raw_affiliation_strings":["Defense Innovation Institute, Academy of Military Sciences, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0003-3866-9060","affiliations":[{"raw_affiliation_string":"Defense Innovation Institute, Academy of Military Sciences, Beijing, China","institution_ids":["https://openalex.org/I2802541053"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Enda Yu","orcid":"https://orcid.org/0000-0003-2661-0889"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Enda Yu","raw_affiliation_strings":["National University of Defense Technology, Changsha, Hunan, China"],"raw_orcid":"https://orcid.org/0000-0003-2661-0889","affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, Hunan, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Weixia Xu","orcid":"https://orcid.org/0000-0003-2412-6591"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Weixia Xu","raw_affiliation_strings":["National University of Defense Technology, Changsha, Hunan, China"],"raw_orcid":"https://orcid.org/0000-0003-2412-6591","affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, Hunan, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"last","author":{"id":null,"display_name":"Xiangke Liao","orcid":"https://orcid.org/0000-0002-5437-0088"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiangke Liao","raw_affiliation_strings":["National University of Defense Technology, Changsha, Hunan, China"],"raw_orcid":"https://orcid.org/0000-0002-5437-0088","affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, Hunan, China","institution_ids":["https://openalex.org/I170215575"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5042353509"],"corresponding_institution_ids":["https://openalex.org/I2802541053"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.00251742,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"20","issue":"1","first_page":"1","last_page":"25"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9977999925613403,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9977999925613403,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9970999956130981,"subfield":{"id":"https://openalex.org/subfields/2808","display_name":"Neurology"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9965000152587891,"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/bottleneck","display_name":"Bottleneck","score":0.8528536558151245},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8222968578338623},{"id":"https://openalex.org/keywords/asynchronous-communication","display_name":"Asynchronous communication","score":0.7456871271133423},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.6809656023979187},{"id":"https://openalex.org/keywords/synchronization","display_name":"Synchronization (alternating current)","score":0.604396641254425},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.5372846722602844},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.43646419048309326},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.43481940031051636},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.37511545419692993},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.18382114171981812},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.18238642811775208},{"id":"https://openalex.org/keywords/channel","display_name":"Channel (broadcasting)","score":0.09126746654510498},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.08428969979286194}],"concepts":[{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.8528536558151245},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8222968578338623},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.7456871271133423},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.6809656023979187},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.604396641254425},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.5372846722602844},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.43646419048309326},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.43481940031051636},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37511545419692993},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.18382114171981812},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.18238642811775208},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.09126746654510498},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.08428969979286194},{"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/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","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}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3563038","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3563038","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3563038","source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"},{"id":"mag:3113081714","is_oa":true,"landing_page_url":"https://arxiv.org/pdf/2012.05396","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"arXiv (Cornell University)","raw_type":null}],"best_oa_location":{"id":"doi:10.1145/3563038","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3563038","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3563038","source":{"id":"https://openalex.org/S26056741","display_name":"ACM Transactions on Architecture and Code Optimization","issn_l":"1544-3566","issn":["1544-3566","1544-3973"],"is_oa":true,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ACM Transactions on Architecture and Code Optimization","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3113081714.pdf","grobid_xml":"https://content.openalex.org/works/W3113081714.grobid-xml"},"referenced_works_count":28,"referenced_works":["https://openalex.org/W2108598243","https://openalex.org/W2194775991","https://openalex.org/W2579247884","https://openalex.org/W2580688187","https://openalex.org/W2626721309","https://openalex.org/W2744599870","https://openalex.org/W2794670651","https://openalex.org/W2894576937","https://openalex.org/W2902410171","https://openalex.org/W2942889782","https://openalex.org/W2944542720","https://openalex.org/W2944793600","https://openalex.org/W2949161920","https://openalex.org/W2963917928","https://openalex.org/W2965658867","https://openalex.org/W2969388332","https://openalex.org/W2973606585","https://openalex.org/W2975712713","https://openalex.org/W2982664135","https://openalex.org/W2994144272","https://openalex.org/W2997300524","https://openalex.org/W3016395792","https://openalex.org/W3037871107","https://openalex.org/W3038104246","https://openalex.org/W3043522163","https://openalex.org/W3047002910","https://openalex.org/W3090287616","https://openalex.org/W6703420464"],"related_works":["https://openalex.org/W2990061623","https://openalex.org/W3032687999","https://openalex.org/W2963179579","https://openalex.org/W3191152329","https://openalex.org/W3205436659","https://openalex.org/W3195071629","https://openalex.org/W2746114819","https://openalex.org/W3106387899","https://openalex.org/W3041891006","https://openalex.org/W2886404071","https://openalex.org/W2982408624","https://openalex.org/W3175155662","https://openalex.org/W2931501180","https://openalex.org/W2529853123","https://openalex.org/W2900182564","https://openalex.org/W3204244704","https://openalex.org/W2990352720","https://openalex.org/W3204684378","https://openalex.org/W2891735383","https://openalex.org/W3016599188"],"abstract_inverted_index":{"Intensive":[0],"communication":[1,55,58,111],"and":[2,7,70,83,92,110,113,119,135,173,180],"synchronization":[3,85,108],"cost":[4],"for":[5,142,149,161],"gradients":[6],"parameters":[8],"is":[9,147],"the":[10,20,54,67,75,97,104,115,123,136,162,204],"well-known":[11],"bottleneck":[12,56],"of":[13,99,126,139,203],"distributed":[14,190],"deep":[15],"learning":[16],"training.":[17],"Based":[18],"on":[19,170],"observations":[21],"that":[22,186],"Synchronous":[23],"SGD":[24,32,45],"(SSGD)":[25],"obtains":[26],"good":[27,89,209],"convergence":[28,90,210],"accuracy":[29,91,118],"while":[30,207],"asynchronous":[31,71],"(ASGD)":[33],"delivers":[34],"a":[35],"faster":[36],"raw":[37],"training":[38,94,120,191],"speed,":[39],"we":[40,102,167],"propose":[41],"Several":[42],"Steps":[43],"Delay":[44],"(SSD-SGD)":[46],"to":[47,158,199],"combine":[48],"their":[49],"merits,":[50],"aiming":[51],"at":[52],"tackling":[53],"via":[57],"sparsification.":[59],"SSD-SGD":[60,87,127,169,187],"explores":[61],"both":[62],"global":[63,140,155],"synchronous":[64],"updates":[65,73],"in":[66,74,77],"parameter":[68],"servers":[69],"local":[72,143,150,164],"workers":[76],"each":[78],"periodic":[79,82],"iteration.":[80],"The":[81],"flexible":[84],"makes":[86],"achieve":[88],"fast":[93],"speed.":[95,121],"To":[96],"best":[98],"our":[100],"knowledge,":[101],"strike":[103],"new":[105],"balance":[106],"between":[107,117],"quality":[109],"sparsification,":[112],"improve":[114],"tradeoff":[116],"Specifically,":[122],"core":[124],"components":[125],"include":[128],"proper":[129],"warm-up":[130],"stage,":[131,134],"steps":[132],"delay":[133],"novel":[137],"algorithm":[138],"gradient":[141,156],"update":[144,151],"(GLU).":[145],"GLU":[146],"critical":[148],"operations":[152],"by":[153,197],"using":[154],"information":[157],"effectively":[159],"compensate":[160],"delayed":[163],"weights.":[165],"Furthermore,":[166],"implement":[168],"MXNet":[171],"framework":[172],"comprehensively":[174],"evaluate":[175],"its":[176],"performance":[177],"with":[178],"CIFAR-10":[179],"ImageNet":[181],"datasets.":[182],"Experimental":[183],"results":[184],"show":[185],"can":[188],"accelerate":[189],"speed":[192],"under":[193],"different":[194],"experimental":[195],"configurations,":[196],"up":[198],"110%":[200],"(or":[201],"2.1\u00d7":[202],"original":[205],"speed),":[206],"achieving":[208],"accuracy.":[211]},"counts_by_year":[],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}