{"id":"https://openalex.org/W3204244704","doi":"https://doi.org/10.1145/3472456.3472508","title":"CD-SGD: Distributed Stochastic Gradient Descent with Compression and Delay Compensation","display_name":"CD-SGD: Distributed Stochastic Gradient Descent with Compression and Delay Compensation","publication_year":2021,"publication_date":"2021-08-09","ids":{"openalex":"https://openalex.org/W3204244704","doi":"https://doi.org/10.1145/3472456.3472508","mag":"3204244704"},"language":"en","primary_location":{"id":"doi:10.1145/3472456.3472508","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3472456.3472508","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"50th International Conference on Parallel Processing","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/A5080680105","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":true,"raw_author_name":"Enda Yu","raw_affiliation_strings":["National University of Defense Technology, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]},{"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, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042353509","display_name":"Yemao Xu","orcid":"https://orcid.org/0000-0003-3639-5449"},"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":"Yemao Xu","raw_affiliation_strings":["National University of Defense Technology, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007572801","display_name":"Shuo Ouyang","orcid":null},"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":"Shuo Ouyang","raw_affiliation_strings":["National University of Defense Technology, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5000688713","display_name":"Xiangke Liao","orcid":null},"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, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, China","institution_ids":["https://openalex.org/I170215575"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5080680105"],"corresponding_institution_ids":["https://openalex.org/I170215575"],"apc_list":null,"apc_paid":null,"fwci":0.4803,"has_fulltext":false,"cited_by_count":6,"citation_normalized_percentile":{"value":0.65318627,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"10"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9993000030517578,"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":0.9993000030517578,"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/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.9991999864578247,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic 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/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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7276954650878906},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"Stochastic gradient descent","score":0.5005359649658203},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.4960497319698334},{"id":"https://openalex.org/keywords/rate-of-convergence","display_name":"Rate of convergence","score":0.48737603425979614},{"id":"https://openalex.org/keywords/data-compression-ratio","display_name":"Data compression ratio","score":0.479951411485672},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.47595950961112976},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.47315868735313416},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.46639081835746765},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.43565040826797485},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.43196791410446167},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.42402368783950806},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4232000708580017},{"id":"https://openalex.org/keywords/image-compression","display_name":"Image compression","score":0.21205955743789673},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.18042033910751343},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.1562521755695343},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.13041925430297852}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7276954650878906},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.5005359649658203},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.4960497319698334},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.48737603425979614},{"id":"https://openalex.org/C94835093","wikidata":"https://www.wikidata.org/wiki/Q3113333","display_name":"Data compression ratio","level":5,"score":0.479951411485672},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.47595950961112976},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.47315868735313416},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.46639081835746765},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.43565040826797485},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.43196791410446167},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.42402368783950806},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4232000708580017},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.21205955743789673},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.18042033910751343},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.1562521755695343},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.13041925430297852},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"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/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"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/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3472456.3472508","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3472456.3472508","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"50th International Conference on Parallel Processing","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":32,"referenced_works":["https://openalex.org/W2108598243","https://openalex.org/W2405578611","https://openalex.org/W2407022425","https://openalex.org/W2580688187","https://openalex.org/W2606891064","https://openalex.org/W2744599870","https://openalex.org/W2886288284","https://openalex.org/W2904556356","https://openalex.org/W2915943718","https://openalex.org/W2938693278","https://openalex.org/W2963786636","https://openalex.org/W2967558351","https://openalex.org/W2975712713","https://openalex.org/W2985108934","https://openalex.org/W2988070836","https://openalex.org/W2994144272","https://openalex.org/W3016395792","https://openalex.org/W3047357290","https://openalex.org/W3090287616","https://openalex.org/W3099748883","https://openalex.org/W3101036738","https://openalex.org/W3102606411","https://openalex.org/W3102816259","https://openalex.org/W3173941299","https://openalex.org/W6600157417","https://openalex.org/W6600175266","https://openalex.org/W6600281192","https://openalex.org/W6600281463","https://openalex.org/W6600336938","https://openalex.org/W6600351811","https://openalex.org/W6631210761","https://openalex.org/W6702166451"],"related_works":["https://openalex.org/W4206903459","https://openalex.org/W2754816816","https://openalex.org/W4366280654","https://openalex.org/W3160167280","https://openalex.org/W4231621013","https://openalex.org/W4362706668","https://openalex.org/W3008318776","https://openalex.org/W1977633006","https://openalex.org/W1971945429","https://openalex.org/W2041416246"],"abstract_inverted_index":{"Communication":[0],"overhead":[1,106],"is":[2,11,155,206,221],"the":[3,36,64,105,113,127,137,143,148,186],"key":[4],"challenge":[5],"for":[6,136],"distributed":[7,97],"training.":[8],"Gradient":[9],"compression":[10,31,48,55,70,141,154,227],"a":[12,23,74,95,197,229],"widely":[13],"used":[14],"approach":[15],"to":[16,50,73,130,184,208],"reduce":[17],"communication":[18,25,39,114,145],"traffic.":[19],"When":[20],"combining":[21],"with":[22,86],"parallel":[24,84],"mechanism":[26,85],"method":[27,99,160],"like":[28],"pipeline,":[29],"gradient":[30,47,54,69,87,108,140,153,226],"technique":[32,49],"can":[33,103],"greatly":[34],"alleviate":[35],"impact":[37],"of":[38,46,107,112,139,192,204,215],"overhead.":[40],"However,":[41],"there":[42],"exist":[43],"two":[44],"problems":[45],"be":[51,131],"solved.":[52],"Firstly,":[53],"brings":[56],"in":[57,76,124,162],"extra":[58],"computation":[59],"cost,":[60],"which":[61,102],"will":[62],"delay":[63],"next":[65,128],"training":[66],"iteration.":[67],"Secondly,":[68],"usually":[71],"leads":[72],"decrease":[75],"convergence":[77,118,169,176,187,202],"accuracy.":[78,119],"In":[79],"this":[80],"paper,":[81],"we":[82],"combine":[83],"quantization":[88],"and":[89,93,115,142,171,189,217],"delayed":[90],"full-gradient":[91],"compensation,":[92],"propose":[94],"new":[96],"optimization":[98],"named":[100],"CD-SGD,":[101],"hide":[104],"compression,":[109],"overlap":[110],"part":[111],"obtain":[116],"high":[117],"The":[120],"local":[121],"update":[122],"operation":[123],"CD-SGD":[125,167,205],"allows":[126],"iteration":[129],"launched":[132],"quickly":[133],"without":[134],"waiting":[135],"completion":[138],"current":[144],"process.":[146],"Besides,":[147],"accuracy":[149,203],"loss":[150],"caused":[151],"by":[152,157],"solved":[156],"k-step":[158],"correction":[159],"introduced":[161],"CD-SGD.":[163,193],"We":[164,178],"prove":[165],"that":[166,201,214],"has":[168],"guarantee":[170],"it":[172],"achieves":[173],"at":[174],"least":[175],"rate.":[177],"conduct":[179],"extensive":[180],"experiments":[181],"on":[182,196],"MXNet":[183],"verify":[185],"properties":[188],"scaling":[190],"performance":[191],"Experimental":[194],"results":[195],"16-GPU":[198],"cluster":[199],"show":[200],"close":[207],"or":[209],"even":[210],"slightly":[211],"better":[212],"than":[213,224],"S-SGD,":[216],"its":[218],"end-to-end":[219],"time":[220],"30":[222],"less":[223],"2-bit":[225],"under":[228],"56Gbps":[230],"bandwidth":[231],"environment.":[232]},"counts_by_year":[{"year":2025,"cited_by_count":1},{"year":2023,"cited_by_count":3},{"year":2022,"cited_by_count":2}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}