{"id":"https://openalex.org/W7081987860","doi":"https://doi.org/10.1109/tcsvt.2025.3609549","title":"Multiple-Exit Tuning: Towards Inference-Efficient Adaptation for Vision Transformer","display_name":"Multiple-Exit Tuning: Towards Inference-Efficient Adaptation for Vision Transformer","publication_year":2025,"publication_date":"2025-09-12","ids":{"openalex":"https://openalex.org/W7081987860","doi":"https://doi.org/10.1109/tcsvt.2025.3609549"},"language":"en","primary_location":{"id":"doi:10.1109/tcsvt.2025.3609549","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsvt.2025.3609549","pdf_url":null,"source":{"id":"https://openalex.org/S115173108","display_name":"IEEE Transactions on Circuits and Systems for Video Technology","issn_l":"1051-8215","issn":["1051-8215","1558-2205"],"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 Circuits and Systems for Video Technology","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":null,"display_name":"Zheng Liu","orcid":"https://orcid.org/0000-0003-3913-5482"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zheng Liu","raw_affiliation_strings":["Department of Automation, BNRist, Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0003-3913-5482","affiliations":[{"raw_affiliation_string":"Department of Automation, BNRist, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Jinchao Zhu","orcid":"https://orcid.org/0000-0003-2821-4847"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jinchao Zhu","raw_affiliation_strings":["Department of Automation, BNRist, Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0003-2821-4847","affiliations":[{"raw_affiliation_string":"Department of Automation, BNRist, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Nannan Li","orcid":null},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Nannan Li","raw_affiliation_strings":["Department of Automation, BNRist, Tsinghua University, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Automation, BNRist, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"last","author":{"id":null,"display_name":"Gao Huang","orcid":"https://orcid.org/0000-0002-7251-0988"},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Gao Huang","raw_affiliation_strings":["Department of Automation, BNRist, Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-7251-0988","affiliations":[{"raw_affiliation_string":"Department of Automation, BNRist, Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I99065089"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.5745066,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"36","issue":"2","first_page":"2123","last_page":"2136"},"is_retracted":false,"is_paratext":false,"is_xpac":true,"primary_topic":{"id":"https://openalex.org/T12157","display_name":"Geochemistry and Geologic Mapping","score":0.6657000184059143,"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/T12157","display_name":"Geochemistry and Geologic Mapping","score":0.6657000184059143,"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/T13067","display_name":"Geological Modeling and Analysis","score":0.02459999918937683,"subfield":{"id":"https://openalex.org/subfields/1906","display_name":"Geochemistry and Petrology"},"field":{"id":"https://openalex.org/fields/19","display_name":"Earth and Planetary Sciences"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T14311","display_name":"Electrical and Electromagnetic Research","score":0.019500000402331352,"subfield":{"id":"https://openalex.org/subfields/3107","display_name":"Atomic and Molecular Physics, and Optics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5863000154495239},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.5360000133514404},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.4634999930858612},{"id":"https://openalex.org/keywords/regularization","display_name":"Regularization (linguistics)","score":0.4361000061035156},{"id":"https://openalex.org/keywords/linear-programming","display_name":"Linear programming","score":0.423799991607666},{"id":"https://openalex.org/keywords/computational-complexity-theory","display_name":"Computational complexity theory","score":0.3887999951839447},{"id":"https://openalex.org/keywords/transfer-of-learning","display_name":"Transfer of learning","score":0.37630000710487366}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7853000164031982},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5863000154495239},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5464000105857849},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5360000133514404},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5235000252723694},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.4634999930858612},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.4361000061035156},{"id":"https://openalex.org/C41045048","wikidata":"https://www.wikidata.org/wiki/Q202843","display_name":"Linear programming","level":2,"score":0.423799991607666},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.3887999951839447},{"id":"https://openalex.org/C150899416","wikidata":"https://www.wikidata.org/wiki/Q1820378","display_name":"Transfer of learning","level":2,"score":0.37630000710487366},{"id":"https://openalex.org/C2776434776","wikidata":"https://www.wikidata.org/wiki/Q19246213","display_name":"Domain adaptation","level":3,"score":0.37450000643730164},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.3269999921321869},{"id":"https://openalex.org/C5339829","wikidata":"https://www.wikidata.org/wiki/Q1425977","display_name":"Machine vision","level":2,"score":0.3077000081539154},{"id":"https://openalex.org/C163175372","wikidata":"https://www.wikidata.org/wiki/Q3339222","display_name":"Linear model","level":2,"score":0.3075000047683716},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.3003000020980835},{"id":"https://openalex.org/C88230418","wikidata":"https://www.wikidata.org/wiki/Q131476","display_name":"Graph theory","level":2,"score":0.26820001006126404},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.2678000032901764},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.26339998841285706},{"id":"https://openalex.org/C6802819","wikidata":"https://www.wikidata.org/wiki/Q1072174","display_name":"Linear system","level":2,"score":0.2563000023365021}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tcsvt.2025.3609549","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsvt.2025.3609549","pdf_url":null,"source":{"id":"https://openalex.org/S115173108","display_name":"IEEE Transactions on Circuits and Systems for Video Technology","issn_l":"1051-8215","issn":["1051-8215","1558-2205"],"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 Circuits and Systems for Video Technology","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1050089583","display_name":null,"funder_award_id":"2024M751692","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G3282150264","display_name":null,"funder_award_id":"2023M741952","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G5313418749","display_name":null,"funder_award_id":"42327901","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G690265107","display_name":null,"funder_award_id":"2023M741952","funder_id":"https://openalex.org/F4320321543","funder_display_name":"China Postdoctoral Science Foundation"},{"id":"https://openalex.org/G8589672233","display_name":null,"funder_award_id":"2024M751692","funder_id":"https://openalex.org/F4320321543","funder_display_name":"China Postdoctoral Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320321543","display_name":"China Postdoctoral Science Foundation","ror":"https://ror.org/0426zh255"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":45,"referenced_works":["https://openalex.org/W12634471","https://openalex.org/W2108598243","https://openalex.org/W2791052411","https://openalex.org/W2941129248","https://openalex.org/W2963736817","https://openalex.org/W2981812042","https://openalex.org/W3137609883","https://openalex.org/W3176828726","https://openalex.org/W3204647170","https://openalex.org/W3206816211","https://openalex.org/W4226512186","https://openalex.org/W4290714341","https://openalex.org/W4312429345","https://openalex.org/W4312651322","https://openalex.org/W4321780048","https://openalex.org/W4322766882","https://openalex.org/W4376461217","https://openalex.org/W4378194725","https://openalex.org/W4382317673","https://openalex.org/W4382463911","https://openalex.org/W4386072470","https://openalex.org/W4386075714","https://openalex.org/W4386076397","https://openalex.org/W4386085850","https://openalex.org/W4386917288","https://openalex.org/W4387385613","https://openalex.org/W4387665871","https://openalex.org/W4388266798","https://openalex.org/W4390190100","https://openalex.org/W4390873067","https://openalex.org/W4390873881","https://openalex.org/W4391513848","https://openalex.org/W4393147399","https://openalex.org/W4393148183","https://openalex.org/W4393159845","https://openalex.org/W4399422592","https://openalex.org/W4401109685","https://openalex.org/W4402727193","https://openalex.org/W4402727789","https://openalex.org/W4402728160","https://openalex.org/W4402754189","https://openalex.org/W4403941607","https://openalex.org/W4408858386","https://openalex.org/W4415795672","https://openalex.org/W4415799070"],"related_works":[],"abstract_inverted_index":{"Parameter-efficient":[0],"transfer":[1],"learning":[2],"(PETL)":[3],"has":[4,196],"shown":[5],"great":[6],"potential":[7],"in":[8,204],"adapting":[9],"vision":[10],"transformers":[11],"(ViTs)":[12],"pre-trained":[13,69],"on":[14,25],"large-scale":[15],"datasets":[16],"to":[17,42,46,108,132,170,185],"various":[18],"downstream":[19],"tasks.":[20],"Existing":[21],"studies":[22],"primarily":[23],"focus":[24],"minimizing":[26],"the":[27,68,73,76,109,114,146,154,162,172,176,187,201],"number":[28],"of":[29,75,122,156,189,206],"learnable":[30],"parameters.":[31],"Although":[32],"these":[33],"methods":[34,203],"are":[35,79,130,159],"storage-efficient,":[36],"they":[37],"allocate":[38],"excessive":[39],"computational":[40,115],"resources":[41],"easy":[43,96,118],"samples,":[44,165],"leading":[45],"inefficient":[47],"inference.":[48],"To":[49,139],"address":[50],"this":[51],"issue,":[52],"we":[53,166],"introduce":[54],"an":[55],"inference-efficient":[56],"tuning":[57,61],"method":[58],"termed":[59],"multiple-exit":[60],"(MET).":[62],"MET":[63,120,195],"integrates":[64],"multiple":[65],"exits":[66,101],"into":[67,175],"ViT":[70],"backbone.":[71],"Since":[72],"predictions":[74],"adapted":[77],"ViTs":[78],"typically":[80],"made":[81],"by":[82,161],"linear":[83,91,157],"classifiers,":[84],"each":[85],"exit":[86,98],"is":[87],"equipped":[88],"with":[89],"a":[90,197],"prediction":[92],"head.":[93],"During":[94],"inference,":[95],"samples":[97,106],"at":[99,178],"early":[100,179],"and":[102,126,150,209],"only":[103],"hard":[104],"enough":[105],"flow":[107],"final":[110],"exit,":[111],"thus":[112],"reducing":[113],"cost":[116],"for":[117,136],"samples.":[119],"consists":[121],"exit-specific":[123],"adapters":[124],"(E-adapters)":[125],"graph":[127,168],"regularization.":[128],"E-adapters":[129,144],"designed":[131],"extract":[133],"suitable":[134],"representations":[135,173],"different":[137],"exits.":[138,180],"maintain":[140],"parameter":[141],"efficiency,":[142],"all":[143],"share":[145],"same":[147],"base":[148],"down-projection":[149],"up-projection":[151],"matrices.":[152],"As":[153],"performances":[155],"classifiers":[158,177],"influenced":[160],"relationships":[163],"among":[164],"employ":[167],"regularization":[169],"improve":[171],"fed":[174],"We":[181],"conduct":[182],"extensive":[183],"experiments":[184],"evaluate":[186],"performance":[188],"MET.":[190],"Experimental":[191],"results":[192],"show":[193],"that":[194],"clear":[198],"advantage":[199],"over":[200],"state-of-the-art":[202],"terms":[205],"both":[207],"accuracy":[208],"inference":[210],"efficiency.":[211]},"counts_by_year":[],"updated_date":"2026-02-12T06:13:03.167441","created_date":"2025-10-10T00:00:00"}