{"id":"https://openalex.org/W4405778753","doi":"https://doi.org/10.1109/tpds.2024.3521582","title":"Co-Designing Transformer Architectures for Distributed Inference With Low Communication","display_name":"Co-Designing Transformer Architectures for Distributed Inference With Low Communication","publication_year":2024,"publication_date":"2024-12-25","ids":{"openalex":"https://openalex.org/W4405778753","doi":"https://doi.org/10.1109/tpds.2024.3521582"},"language":"en","primary_location":{"id":"doi:10.1109/tpds.2024.3521582","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2024.3521582","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Parallel and Distributed Systems","raw_type":"journal-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/A5024085713","display_name":"Jiangsu Du","orcid":"https://orcid.org/0000-0003-4707-9492"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiangsu Du","raw_affiliation_strings":["School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0003-4707-9492","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110558003","display_name":"Yuanxin Wei","orcid":null},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yuanxin Wei","raw_affiliation_strings":["School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"],"raw_orcid":"https://orcid.org/0009-0000-6551-9274","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5021061004","display_name":"Shengyuan Ye","orcid":null},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shengyuan Ye","raw_affiliation_strings":["School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0001-8867-0655","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048148542","display_name":"Jiazhi Jiang","orcid":"https://orcid.org/0000-0002-1417-3012"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiazhi Jiang","raw_affiliation_strings":["School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0002-1417-3012","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100385692","display_name":"Xu Chen","orcid":"https://orcid.org/0000-0001-9943-6020"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xu Chen","raw_affiliation_strings":["School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0001-9943-6020","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041534890","display_name":"Dan Huang","orcid":"https://orcid.org/0000-0001-5582-1031"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dan Huang","raw_affiliation_strings":["School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0001-5582-1031","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101633465","display_name":"Yutong Lu","orcid":"https://orcid.org/0000-0001-5315-3375"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yutong Lu","raw_affiliation_strings":["School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0001-5315-3375","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I157773358"],"apc_list":null,"apc_paid":null,"fwci":1.001,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.78156425,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":"36","issue":"4","first_page":"717","last_page":"730"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10904","display_name":"Embedded Systems Design Techniques","score":0.8122000098228455,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.8122000098228455,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10829","display_name":"Interconnection Networks and Systems","score":0.763700008392334,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8178391456604004},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5151909589767456},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.46690186858177185},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.4270270764827728},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.36239364743232727},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.3398836851119995},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.16588667035102844},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.09963753819465637},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.0954708456993103}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8178391456604004},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5151909589767456},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.46690186858177185},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.4270270764827728},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.36239364743232727},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.3398836851119995},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.16588667035102844},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.09963753819465637},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0954708456993103},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tpds.2024.3521582","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2024.3521582","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Parallel and Distributed Systems","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3236153463","display_name":null,"funder_award_id":"62461146204","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G63103377","display_name":null,"funder_award_id":"62402534","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G6651040689","display_name":null,"funder_award_id":"62332021","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8411335299","display_name":null,"funder_award_id":"2023A1515110117","funder_id":"https://openalex.org/F4320337111","funder_display_name":"Basic and Applied Basic Research Foundation of Guangdong Province"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320337111","display_name":"Basic and Applied Basic Research Foundation of Guangdong Province","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":24,"referenced_works":["https://openalex.org/W2896180420","https://openalex.org/W2923014074","https://openalex.org/W2963125010","https://openalex.org/W2963542740","https://openalex.org/W2963748441","https://openalex.org/W2991040477","https://openalex.org/W3108109508","https://openalex.org/W3109233295","https://openalex.org/W3130716829","https://openalex.org/W3193985311","https://openalex.org/W3210103168","https://openalex.org/W4282962596","https://openalex.org/W4285134706","https://openalex.org/W4321636575","https://openalex.org/W4380926962","https://openalex.org/W4385245566","https://openalex.org/W4385571616","https://openalex.org/W4387321091","https://openalex.org/W4387869595","https://openalex.org/W4388937922","https://openalex.org/W4391971727","https://openalex.org/W4401018280","https://openalex.org/W4401508667","https://openalex.org/W4404401018"],"related_works":["https://openalex.org/W2055243143","https://openalex.org/W1986418932","https://openalex.org/W2357796999","https://openalex.org/W4321636575","https://openalex.org/W2111375262","https://openalex.org/W2741131631","https://openalex.org/W4206178588","https://openalex.org/W3094491777","https://openalex.org/W3214715529","https://openalex.org/W4287635093"],"abstract_inverted_index":{"Transformer":[0,60,72,105,117,141,260],"models":[1],"have":[2],"shown":[3],"significant":[4],"success":[5],"in":[6],"a":[7,23,32,102,132,144,171,186,203,207,290],"wide":[8],"range":[9],"of":[10,35,71,111,116,259],"tasks.":[11],"However,":[12],"the":[13,68,114,121,139,148,218,278],"massive":[14],"resources":[15],"required":[16],"for":[17,88,189,253],"its":[18],"inference":[19,61,106,270],"prevent":[20],"deployment":[21,51],"on":[22,46,52,131,230,246,281],"single":[24,145],"device":[25],"with":[26,143,151],"relatively":[27,93],"constrainted":[28],"resources,":[29],"thus":[30],"leaving":[31],"high":[33],"threshold":[34],"integrating":[36],"their":[37],"advancements.":[38],"Observing":[39],"scenarios":[40],"such":[41],"as":[42],"smart":[43],"home":[44],"applications":[45],"edge":[47],"devices":[48,241],"and":[49,178,194,206,256,289],"cloud":[50],"commodity":[53],"hardware,":[54],"it":[55,155],"is":[56,129],"promising":[57],"to":[58,67,81,119,147,239,274,277],"distribute":[59],"across":[62],"multiple":[63,152,182],"devices.":[64,183],"Unfortunately,":[65],"due":[66],"tightly-coupled":[69],"feature":[70],"model,":[73],"existing":[74],"model":[75,158,292],"parallelism":[76,135,159],"approaches":[77],"necessitate":[78],"frequent":[79],"communication":[80,122,174],"resolve":[82],"data":[83],"dependencies,":[84],"making":[85],"them":[86],"unacceptable":[87],"distributed":[89,104,124,269],"inference,":[90],"especially":[91],"under":[92],"weak":[94],"interconnection.":[95],"In":[96,126],"this":[97,234],"paper,":[98],"we":[99],"propose":[100],"DeTransformer,":[101],"communication-efficient":[103],"system.":[107],"The":[108,197,213],"key":[109],"idea":[110],"DeTransformer":[112,128,163,266],"involves":[113],"co-design":[115],"architecture":[118],"reduce":[120,268],"during":[123],"inference.":[125],"detail,":[127],"based":[130,229,245],"novel":[133],"block":[134,146],"approach,":[136],"which":[137],"restructures":[138],"original":[140],"layer":[142,150],"decoupled":[149],"sub-blocks.":[153,161],"Thus,":[154],"can":[156,237,267],"exploit":[157],"between":[160],"Next,":[162],"contains":[164],"an":[165,224],"adaptive":[166],"execution":[167,235],"approach":[168,236,280],"that":[169,265],"strikes":[170],"trade-off":[172],"among":[173],"capability,":[175],"computing":[176,221,248],"power":[177],"memory":[179],"budget":[180],"over":[181],"It":[184],"incorporates":[185],"two-phase":[187],"planning":[188,193,199,215],"execution,":[190],"namely":[191],"static":[192,198],"runtime":[195,214],"planning.":[196],"runs":[200],"offline,":[201],"containing":[202],"profiling":[204],"procedure":[205],"weight":[208],"placement":[209],"strategy":[210,222],"before":[211],"execution.":[212],"dynamically":[216],"determines":[217],"optimal":[219],"parallel":[220],"from":[223],"expertly":[225],"crafted":[226],"search":[227],"space":[228],"real-time":[231],"requests.":[232],"Notably,":[233],"adapt":[238],"heterogeneous":[240],"by":[242,272],"distributing":[243],"workload":[244],"devices\u2019":[247],"capabilities.":[249],"We":[250],"conduct":[251],"experiments":[252],"both":[254],"auto-regressive":[255],"auto-encoder":[257],"tasks":[258],"models.":[261],"Experimental":[262],"results":[263],"show":[264],"latency":[271],"up":[273],"2.81\u00d7":[275],"compared":[276],"SOTA":[279],"4":[282],"devices,":[283],"while":[284],"effectively":[285],"maintaining":[286],"task":[287],"accuracy":[288],"consistent":[291],"size.":[293]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":2}],"updated_date":"2026-07-15T18:14:33.161393","created_date":"2025-10-10T00:00:00"}