{"id":"https://openalex.org/W7123874558","doi":"https://doi.org/10.1145/3772052.3772230","title":"Symbiosis: Multi-Adapter Inference and Fine-Tuning","display_name":"Symbiosis: Multi-Adapter Inference and Fine-Tuning","publication_year":2025,"publication_date":"2025-11-19","ids":{"openalex":"https://openalex.org/W7123874558","doi":"https://doi.org/10.1145/3772052.3772230"},"language":null,"primary_location":{"id":"doi:10.1145/3772052.3772230","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3772052.3772230","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 ACM Symposium on Cloud Computing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3772052.3772230","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5122915953","display_name":"Saransh Gupta","orcid":null},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Saransh Gupta","raw_affiliation_strings":["IBM Research, San Jose, USA"],"raw_orcid":"https://orcid.org/0000-0001-5814-3934","affiliations":[{"raw_affiliation_string":"IBM Research, San Jose, USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5035364139","display_name":"U. D. Deshpande","orcid":"https://orcid.org/0000-0001-9506-0003"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Umesh Deshpande","raw_affiliation_strings":["IBM Research, San Jose, USA"],"raw_orcid":"https://orcid.org/0000-0001-9506-0003","affiliations":[{"raw_affiliation_string":"IBM Research, San Jose, USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076308023","display_name":"Travis Janssen","orcid":"https://orcid.org/0009-0007-3033-3964"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Travis Janssen","raw_affiliation_strings":["IBM Research, San Jose, USA"],"raw_orcid":"https://orcid.org/0009-0007-3033-3964","affiliations":[{"raw_affiliation_string":"IBM Research, San Jose, USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5099587073","display_name":"Swaminathan Sundararaman","orcid":"https://orcid.org/0000-0003-4468-3061"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Swaminathan Sundararaman","raw_affiliation_strings":["IBM Research, San Jose, USA"],"raw_orcid":"https://orcid.org/0000-0003-4468-3061","affiliations":[{"raw_affiliation_string":"IBM Research, San Jose, USA","institution_ids":["https://openalex.org/I1341412227"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.65907473,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"776","last_page":"789"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.21529999375343323,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.21529999375343323,"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/T12127","display_name":"Software System Performance and Reliability","score":0.11410000175237656,"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"}},{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.09350000321865082,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.7193999886512756},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.6220999956130981},{"id":"https://openalex.org/keywords/flexibility","display_name":"Flexibility (engineering)","score":0.5440000295639038},{"id":"https://openalex.org/keywords/base","display_name":"Base (topology)","score":0.4641000032424927},{"id":"https://openalex.org/keywords/resource-consumption","display_name":"Resource consumption","score":0.414000004529953},{"id":"https://openalex.org/keywords/resource","display_name":"Resource (disambiguation)","score":0.3702999949455261}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8476999998092651},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7193999886512756},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.6220999956130981},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.5440000295639038},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.4641000032424927},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.4456999897956848},{"id":"https://openalex.org/C2777480716","wikidata":"https://www.wikidata.org/wiki/Q23582796","display_name":"Resource consumption","level":2,"score":0.414000004529953},{"id":"https://openalex.org/C206345919","wikidata":"https://www.wikidata.org/wiki/Q20380951","display_name":"Resource (disambiguation)","level":2,"score":0.3702999949455261},{"id":"https://openalex.org/C46743427","wikidata":"https://www.wikidata.org/wiki/Q1341685","display_name":"Inference engine","level":3,"score":0.3294999897480011},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.31839999556541443},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3149000108242035},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3012000024318695},{"id":"https://openalex.org/C4554734","wikidata":"https://www.wikidata.org/wiki/Q593744","display_name":"Knowledge base","level":2,"score":0.2969000041484833},{"id":"https://openalex.org/C115903868","wikidata":"https://www.wikidata.org/wiki/Q80993","display_name":"Software engineering","level":1,"score":0.28940001130104065},{"id":"https://openalex.org/C2780378061","wikidata":"https://www.wikidata.org/wiki/Q25351891","display_name":"Service (business)","level":2,"score":0.2791999876499176},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.27639999985694885},{"id":"https://openalex.org/C2780586970","wikidata":"https://www.wikidata.org/wiki/Q1357284","display_name":"Popularity","level":2,"score":0.2637999951839447},{"id":"https://openalex.org/C68649174","wikidata":"https://www.wikidata.org/wiki/Q1379116","display_name":"Base station","level":2,"score":0.25270000100135803}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3772052.3772230","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3772052.3772230","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 ACM Symposium on Cloud Computing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3772052.3772230","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3772052.3772230","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 ACM Symposium on Cloud Computing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":8,"referenced_works":["https://openalex.org/W2798291715","https://openalex.org/W3119841746","https://openalex.org/W3205803342","https://openalex.org/W4321636575","https://openalex.org/W4387321091","https://openalex.org/W4388514131","https://openalex.org/W4404367639","https://openalex.org/W4409048416"],"related_works":[],"abstract_inverted_index":{"Parameter-efficient":[0],"fine-tuning":[1,30,57,168,197],"(PEFT)":[2],"allows":[3],"model":[4,76,159,185],"builders":[5],"to":[6,33,71,127,133,138,190,194,199,208,226],"capture":[7],"the":[8,18,21,34,62,145,150,154,174,182,217,222],"task-specific":[9],"parameters":[10,137],"into":[11],"adapters,":[12,98],"which":[13,78],"are":[14],"a":[15,37],"fraction":[16],"of":[17,20,26,36,40,108,118,153,176,224],"size":[19],"original":[22],"base":[23,75,155,158,184],"model.":[24,156],"Popularity":[25],"PEFT":[27,97,110],"technique":[28,172],"for":[29,42,213],"has":[31],"led":[32],"creation":[35],"large":[38],"number":[39],"adapters":[41,60,178,232],"popular":[43,91],"Large":[44],"Language":[45],"Models":[46],"(LLMs).":[47],"However,":[48],"existing":[49],"frameworks":[50],"fall":[51],"short":[52],"in":[53,61,80,216],"supporting":[54],"inference":[55,92,166],"or":[56,106,167],"with":[58],"multiple":[59,96,165],"following":[63],"ways.":[64],"1)":[65],"For":[66],"fine-tuning,":[67],"each":[68],"job":[69],"needs":[70],"deploy":[72],"its":[73],"dedicated":[74],"instance,":[77],"results":[79],"excessive":[81],"GPU":[82,87],"memory":[83],"consumption":[84],"and":[85,179,210],"poor":[86],"utilization.":[88],"2)":[89],"While":[90],"platforms":[93],"can":[94,161],"serve":[95],"they":[99],"do":[100,123],"not":[101,124,131],"allow":[102],"independent":[103],"resource":[104],"management":[105],"mixing":[107],"different":[109],"methods.":[111],"3)":[112],"They":[113,122],"cannot":[114],"make":[115],"effective":[116],"use":[117,223],"heterogeneous":[119],"accelerators.":[120],"4)":[121],"provide":[125],"privacy":[126],"users":[128],"who":[129],"may":[130],"wish":[132],"expose":[134],"their":[135,192,196,201],"fine-tuned":[136],"service":[139],"providers.":[140],"In":[141],"Symbiosis,":[142],"we":[143],"address":[144],"above":[146],"problems":[147],"by":[148],"enabling":[149],"as-a-service":[151],"deployment":[152],"The":[157],"layers":[160,180,186],"be":[162],"shared":[163],"across":[164],"processes.":[169],"Our":[170,204],"split-execution":[171],"decouples":[173],"execution":[175],"client-specific":[177],"from":[181],"frozen":[183],"offering":[187],"them":[188],"flexibility":[189],"manage":[191],"resources,":[193],"select":[195],"method,":[198],"achieve":[200],"performance":[202],"goals.":[203],"approach":[205],"is":[206],"transparent":[207],"models":[209,215],"works":[211],"out-of-the-box":[212],"most":[214],"transformers":[218],"library.":[219],"We":[220],"demonstrate":[221],"Symbiosis":[225],"simultaneously":[227],"fine-tune":[228],"20":[229],"Gemma2-27B":[230],"LoRA":[231],"on":[233],"8":[234],"GPUs.":[235]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-01-14T00:00:00"}