{"id":"https://openalex.org/W4414360991","doi":"https://doi.org/10.24963/ijcai.2025/4","title":"Optimal Distributed Training With Co-Adaptive Data Parallelism in Heterogeneous Environments","display_name":"Optimal Distributed Training With Co-Adaptive Data Parallelism in Heterogeneous Environments","publication_year":2025,"publication_date":"2025-09-01","ids":{"openalex":"https://openalex.org/W4414360991","doi":"https://doi.org/10.24963/ijcai.2025/4"},"language":"en","primary_location":{"id":"doi:10.24963/ijcai.2025/4","is_oa":false,"landing_page_url":"https://doi.org/10.24963/ijcai.2025/4","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","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/A5062143141","display_name":"Lifang Chen","orcid":"https://orcid.org/0000-0002-2375-1902"},"institutions":[{"id":"https://openalex.org/I50760025","display_name":"Hangzhou Dianzi University","ror":"https://ror.org/0576gt767","country_code":"CN","type":"education","lineage":["https://openalex.org/I50760025"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lifang Chen","raw_affiliation_strings":["hangzhou dianzi university"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"hangzhou dianzi university","institution_ids":["https://openalex.org/I50760025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100619477","display_name":"Zhichao Chen","orcid":"https://orcid.org/0000-0002-8776-3793"},"institutions":[{"id":"https://openalex.org/I50760025","display_name":"Hangzhou Dianzi University","ror":"https://ror.org/0576gt767","country_code":"CN","type":"education","lineage":["https://openalex.org/I50760025"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhichao Chen","raw_affiliation_strings":["hangzhou dianzi university"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"hangzhou dianzi university","institution_ids":["https://openalex.org/I50760025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051997091","display_name":"Liqi Yan","orcid":"https://orcid.org/0000-0002-7077-4947"},"institutions":[{"id":"https://openalex.org/I50760025","display_name":"Hangzhou Dianzi University","ror":"https://ror.org/0576gt767","country_code":"CN","type":"education","lineage":["https://openalex.org/I50760025"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Liqi Yan","raw_affiliation_strings":["hangzhou dianzi university"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"hangzhou dianzi university","institution_ids":["https://openalex.org/I50760025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073957898","display_name":"Yanyu Cheng","orcid":"https://orcid.org/0000-0001-9104-9376"},"institutions":[{"id":"https://openalex.org/I50760025","display_name":"Hangzhou Dianzi University","ror":"https://ror.org/0576gt767","country_code":"CN","type":"education","lineage":["https://openalex.org/I50760025"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yanyu Cheng","raw_affiliation_strings":["hangzhou dianzi university"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"hangzhou dianzi university","institution_ids":["https://openalex.org/I50760025"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5012288206","display_name":"Fangli Guan","orcid":"https://orcid.org/0000-0001-7409-2129"},"institutions":[{"id":"https://openalex.org/I50760025","display_name":"Hangzhou Dianzi University","ror":"https://ror.org/0576gt767","country_code":"CN","type":"education","lineage":["https://openalex.org/I50760025"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fangli Guan","raw_affiliation_strings":["hangzhou dianzi university"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"hangzhou dianzi university","institution_ids":["https://openalex.org/I50760025"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100455219","display_name":"Pan Li","orcid":"https://orcid.org/0000-0003-3742-0845"},"institutions":[{"id":"https://openalex.org/I50760025","display_name":"Hangzhou Dianzi University","ror":"https://ror.org/0576gt767","country_code":"CN","type":"education","lineage":["https://openalex.org/I50760025"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Pan Li","raw_affiliation_strings":["hangzhou dianzi university"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"hangzhou dianzi university","institution_ids":["https://openalex.org/I50760025"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I50760025"],"apc_list":null,"apc_paid":null,"fwci":1.396,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.86036428,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":null,"issue":null,"first_page":"29","last_page":"37"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.9429000020027161,"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/T10320","display_name":"Neural Networks and Applications","score":0.9429000020027161,"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/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9056000113487244,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.6204000115394592},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.590499997138977},{"id":"https://openalex.org/keywords/data-parallelism","display_name":"Data parallelism","score":0.5133000016212463},{"id":"https://openalex.org/keywords/distributed-algorithm","display_name":"Distributed algorithm","score":0.4586000144481659},{"id":"https://openalex.org/keywords/synchronization","display_name":"Synchronization (alternating current)","score":0.4562999904155731},{"id":"https://openalex.org/keywords/symmetric-multiprocessor-system","display_name":"Symmetric multiprocessor system","score":0.4496999979019165},{"id":"https://openalex.org/keywords/distributed-database","display_name":"Distributed database","score":0.4189999997615814},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.3862000107765198},{"id":"https://openalex.org/keywords/distributed-learning","display_name":"Distributed learning","score":0.38530001044273376}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8515999913215637},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.6960999965667725},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.6204000115394592},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.590499997138977},{"id":"https://openalex.org/C61483411","wikidata":"https://www.wikidata.org/wiki/Q3124522","display_name":"Data parallelism","level":3,"score":0.5133000016212463},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4722999930381775},{"id":"https://openalex.org/C130120984","wikidata":"https://www.wikidata.org/wiki/Q2835898","display_name":"Distributed algorithm","level":2,"score":0.4586000144481659},{"id":"https://openalex.org/C2778562939","wikidata":"https://www.wikidata.org/wiki/Q1298791","display_name":"Synchronization (alternating current)","level":3,"score":0.4562999904155731},{"id":"https://openalex.org/C172430144","wikidata":"https://www.wikidata.org/wiki/Q17111997","display_name":"Symmetric multiprocessor system","level":2,"score":0.4496999979019165},{"id":"https://openalex.org/C70061542","wikidata":"https://www.wikidata.org/wiki/Q989016","display_name":"Distributed database","level":2,"score":0.4189999997615814},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.3862000107765198},{"id":"https://openalex.org/C2779582901","wikidata":"https://www.wikidata.org/wiki/Q21013010","display_name":"Distributed learning","level":2,"score":0.38530001044273376},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.38359999656677246},{"id":"https://openalex.org/C2781172179","wikidata":"https://www.wikidata.org/wiki/Q853109","display_name":"Parallelism (grammar)","level":2,"score":0.35350000858306885},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.34220001101493835},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.34139999747276306},{"id":"https://openalex.org/C75684735","wikidata":"https://www.wikidata.org/wiki/Q858810","display_name":"Big data","level":2,"score":0.33399999141693115},{"id":"https://openalex.org/C66882249","wikidata":"https://www.wikidata.org/wiki/Q169336","display_name":"Homogeneous","level":2,"score":0.3206000030040741},{"id":"https://openalex.org/C24885549","wikidata":"https://www.wikidata.org/wiki/Q339678","display_name":"Distributed data store","level":2,"score":0.3183000087738037},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3073999881744385},{"id":"https://openalex.org/C3739613","wikidata":"https://www.wikidata.org/wiki/Q679003","display_name":"Distributed Computing Environment","level":2,"score":0.298799991607666},{"id":"https://openalex.org/C120373497","wikidata":"https://www.wikidata.org/wiki/Q1087987","display_name":"Parallel algorithm","level":2,"score":0.29429998993873596},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.2930999994277954},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.28940001130104065},{"id":"https://openalex.org/C93996380","wikidata":"https://www.wikidata.org/wiki/Q44127","display_name":"Server","level":2,"score":0.2775999903678894},{"id":"https://openalex.org/C1668388","wikidata":"https://www.wikidata.org/wiki/Q1149776","display_name":"Data management","level":2,"score":0.2689000070095062},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.26510000228881836},{"id":"https://openalex.org/C55416958","wikidata":"https://www.wikidata.org/wiki/Q6206757","display_name":"Job shop scheduling","level":3,"score":0.26089999079704285},{"id":"https://openalex.org/C854659","wikidata":"https://www.wikidata.org/wiki/Q1859284","display_name":"Message passing","level":2,"score":0.25850000977516174}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.24963/ijcai.2025/4","is_oa":false,"landing_page_url":"https://doi.org/10.24963/ijcai.2025/4","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","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":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"computational":[1,39],"power":[2],"required":[3],"for":[4,43,61,74],"training":[5,52,95,128,149,193,218],"deep":[6],"learning":[7,119],"models":[8],"has":[9,23],"been":[10],"skyrocketing":[11],"in":[12,175,190,211,217],"the":[13,117,122,148,170,176,181],"past":[14],"decade":[15],"as":[16],"they":[17],"scale":[18],"with":[19,77],"big":[20],"data,":[21],"and":[22,28,56,67,70,80,108,115,129,166,201,207,214,223],"become":[24],"a":[25,89,105,140,215],"very":[26],"expensive":[27],"scarce":[29],"resource.":[30],"Therefore,":[31],"distributed":[32,37,51,94,192],"training,":[33],"which":[34],"can":[35],"leverage":[36],"available":[38],"power,":[40],"is":[41,173],"vital":[42],"efficient":[44,159],"large-scale":[45],"model":[46],"training.":[47],"However,":[48],"most":[49],"previous":[50],"frameworks":[53],"like":[54],"DDP":[55],"DeepSpeed":[57],"are":[58],"primarily":[59],"designed":[60],"co-located":[62],"clusters":[63,76],"under":[64],"homogeneous":[65],"computing":[66,79,123],"communication":[68,81,135],"conditions,":[69],"hence":[71],"cannot":[72],"account":[73],"geo-distributed":[75],"both":[78],"heterogeneity.":[82],"To":[83],"address":[84],"this":[85,163],"challenge,":[86],"we":[87,103,138,156],"develop":[88],"new":[90],"data":[91,106,113,141,153],"parallel":[92,142],"based":[93],"framework":[96],"called":[97],"Co-Adaptive":[98],"Data":[99,198],"Parallelism":[100],"(C-ADP).":[101],"First,":[102],"consider":[104],"owner":[107],"parameter":[109,131],"server":[110],"that":[111,169,185],"distributes":[112],"to":[114,133,146,161,196],"coordinates":[116],"collaborative":[118],"across":[120],"all":[121],"devices.":[124],"We":[125],"employ":[126],"local":[127],"delayed":[130],"synchronization":[132],"reduce":[134],"costs.":[136],"Second,":[137],"formulate":[139],"scheduling":[143,164],"optimization":[144],"problem":[145],"minimize":[147],"time":[150,219],"by":[151],"optimizing":[152],"distribution.":[154],"Third,":[155],"devise":[157],"an":[158],"algorithm":[160],"solve":[162],"problem,":[165],"formally":[167],"prove":[168],"obtained":[171],"solution":[172],"optimal":[174],"asymptotic":[177],"sense.":[178],"Experiments":[179],"on":[180],"ImageNet100":[182],"dataset":[183],"demonstrate":[184],"C-ADP":[186,203],"achieves":[187,204],"fast":[188],"convergence":[189],"heterogeneous":[191],"environments.":[194],"Compared":[195],"Distributed":[197],"Parallel":[199],"(DDP)":[200],"DeepSpeed,":[202],"21.6":[205],"times":[206,209],"26.3":[208],"improvements":[210],"FLOPS,":[212],"respectively,":[213],"reduction":[216],"of":[220],"about":[221],"72%":[222],"47%,":[224],"respectively.":[225]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}