{"id":"https://openalex.org/W4321021726","doi":"https://doi.org/10.1109/tcsvt.2023.3245584","title":"Understanding and Mitigating Overfitting in Prompt Tuning for Vision-Language Models","display_name":"Understanding and Mitigating Overfitting in Prompt Tuning for Vision-Language Models","publication_year":2023,"publication_date":"2023-02-16","ids":{"openalex":"https://openalex.org/W4321021726","doi":"https://doi.org/10.1109/tcsvt.2023.3245584"},"language":"en","primary_location":{"id":"doi:10.1109/tcsvt.2023.3245584","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsvt.2023.3245584","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":"https://openalex.org/A5101950904","display_name":"Chengcheng Ma","orcid":"https://orcid.org/0000-0002-0502-3960"},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210112150","display_name":"Institute of Automation","ror":"https://ror.org/022c3hy66","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210112150"]},{"id":"https://openalex.org/I4210165038","display_name":"University of Chinese Academy of Sciences","ror":"https://ror.org/05qbk4x57","country_code":"CN","type":"education","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chengcheng Ma","raw_affiliation_strings":["National Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences (CASIA), Beijing, China","School of Artificial Intelligence, University of Chinese Academy of Sciences (UCAS), Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-0502-3960","affiliations":[{"raw_affiliation_string":"National Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences (CASIA), Beijing, China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210112150"]},{"raw_affiliation_string":"School of Artificial Intelligence, University of Chinese Academy of Sciences (UCAS), Beijing, China","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100355884","display_name":"Yang Liu","orcid":"https://orcid.org/0000-0002-4259-3882"},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang Liu","raw_affiliation_strings":["Alibaba DAMO Academy, Hangzhou, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba DAMO Academy, Hangzhou, China","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073110676","display_name":"Jiankang Deng","orcid":"https://orcid.org/0000-0002-3709-6216"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiankang Deng","raw_affiliation_strings":["Huawei Inc., Shenzhen, China"],"raw_orcid":"https://orcid.org/0000-0002-3709-6216","affiliations":[{"raw_affiliation_string":"Huawei Inc., Shenzhen, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5075290241","display_name":"Lingxi Xie","orcid":"https://orcid.org/0000-0003-4831-9451"},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lingxi Xie","raw_affiliation_strings":["Huawei Inc., Shenzhen, China"],"raw_orcid":"https://orcid.org/0000-0003-4831-9451","affiliations":[{"raw_affiliation_string":"Huawei Inc., Shenzhen, China","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069298091","display_name":"Weiming Dong","orcid":"https://orcid.org/0000-0001-6502-145X"},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210112150","display_name":"Institute of Automation","ror":"https://ror.org/022c3hy66","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210112150"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Weiming Dong","raw_affiliation_strings":["National Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences (CASIA), Beijing, China"],"raw_orcid":"https://orcid.org/0000-0001-6502-145X","affiliations":[{"raw_affiliation_string":"National Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences (CASIA), Beijing, China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210112150"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5022636178","display_name":"Changsheng Xu","orcid":"https://orcid.org/0000-0001-8343-9665"},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210112150","display_name":"Institute of Automation","ror":"https://ror.org/022c3hy66","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210112150"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Changsheng Xu","raw_affiliation_strings":["National Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences (CASIA), Beijing, China"],"raw_orcid":"https://orcid.org/0000-0001-8343-9665","affiliations":[{"raw_affiliation_string":"National Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences (CASIA), Beijing, China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210112150"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":6.4478,"has_fulltext":false,"cited_by_count":66,"citation_normalized_percentile":{"value":0.97939037,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":"33","issue":"9","first_page":"4616","last_page":"4629"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9998999834060669,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9998999834060669,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9994999766349792,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9915000200271606,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/overfitting","display_name":"Overfitting","score":0.9736685752868652},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8079475164413452},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.7010748982429504},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.6469066739082336},{"id":"https://openalex.org/keywords/test-set","display_name":"Test set","score":0.46359461545944214},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.4474239945411682},{"id":"https://openalex.org/keywords/segmentation","display_name":"Segmentation","score":0.4328042268753052},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.41853955388069153},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.4178139865398407},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.22141724824905396}],"concepts":[{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.9736685752868652},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8079475164413452},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.7010748982429504},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.6469066739082336},{"id":"https://openalex.org/C169903167","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Test set","level":2,"score":0.46359461545944214},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.4474239945411682},{"id":"https://openalex.org/C89600930","wikidata":"https://www.wikidata.org/wiki/Q1423946","display_name":"Segmentation","level":2,"score":0.4328042268753052},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.41853955388069153},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.4178139865398407},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.22141724824905396},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"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/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tcsvt.2023.3245584","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsvt.2023.3245584","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":[{"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4","score":0.7699999809265137}],"awards":[{"id":"https://openalex.org/G2203821447","display_name":null,"funder_award_id":"U20B2070","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G5049913043","display_name":null,"funder_award_id":"L221013","funder_id":"https://openalex.org/F4320322919","funder_display_name":"Natural Science Foundation of Beijing Municipality"},{"id":"https://openalex.org/G7596092009","display_name":"\u57fa\u4e8e\u89c6\u89c9\u8ba4\u77e5\u7684\u53ef\u89c6\u5a92\u4f53\u5408\u6210\u4e0e\u8bc4\u4ef7","funder_award_id":"61832016","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320322919","display_name":"Natural Science Foundation of Beijing Municipality","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":72,"referenced_works":["https://openalex.org/W12634471","https://openalex.org/W1846799578","https://openalex.org/W1977295328","https://openalex.org/W2017814585","https://openalex.org/W2047643928","https://openalex.org/W2100483895","https://openalex.org/W2108598243","https://openalex.org/W2138011018","https://openalex.org/W2155904486","https://openalex.org/W2194775991","https://openalex.org/W2519388618","https://openalex.org/W2533598788","https://openalex.org/W2561196672","https://openalex.org/W2901114541","https://openalex.org/W2948672349","https://openalex.org/W2963150697","https://openalex.org/W2964194231","https://openalex.org/W2981165461","https://openalex.org/W3027585699","https://openalex.org/W3037492894","https://openalex.org/W3106542916","https://openalex.org/W3107094551","https://openalex.org/W3133542152","https://openalex.org/W3169422999","https://openalex.org/W3177096435","https://openalex.org/W3177206657","https://openalex.org/W3195577433","https://openalex.org/W3198377975","https://openalex.org/W3204250462","https://openalex.org/W3204762109","https://openalex.org/W3213454282","https://openalex.org/W3213647938","https://openalex.org/W4205991051","https://openalex.org/W4207055979","https://openalex.org/W4224311887","https://openalex.org/W4225495512","https://openalex.org/W4229453513","https://openalex.org/W4282939141","https://openalex.org/W4289639938","https://openalex.org/W4297697565","https://openalex.org/W4312310776","https://openalex.org/W4312338297","https://openalex.org/W4312458986","https://openalex.org/W4312581490","https://openalex.org/W4312651322","https://openalex.org/W4385567175","https://openalex.org/W4390873714","https://openalex.org/W6638677478","https://openalex.org/W6736966227","https://openalex.org/W6752842830","https://openalex.org/W6763468762","https://openalex.org/W6764632358","https://openalex.org/W6764990469","https://openalex.org/W6771876938","https://openalex.org/W6786219303","https://openalex.org/W6786798588","https://openalex.org/W6788840927","https://openalex.org/W6790019176","https://openalex.org/W6791353385","https://openalex.org/W6797716411","https://openalex.org/W6797827564","https://openalex.org/W6798837711","https://openalex.org/W6800751262","https://openalex.org/W6802347785","https://openalex.org/W6803476161","https://openalex.org/W6807286953","https://openalex.org/W6810158684","https://openalex.org/W6811028123","https://openalex.org/W6838789006","https://openalex.org/W6838789509","https://openalex.org/W6839015040","https://openalex.org/W6948022635"],"related_works":["https://openalex.org/W2989932438","https://openalex.org/W4387297750","https://openalex.org/W2186333919","https://openalex.org/W3208882810","https://openalex.org/W2972862903","https://openalex.org/W3099765033","https://openalex.org/W2921259037","https://openalex.org/W2970699417","https://openalex.org/W4288419306","https://openalex.org/W2987302549"],"abstract_inverted_index":{"Pretrained":[0],"vision-language":[1],"models":[2],"(VLMs)":[3],"such":[4,94],"as":[5],"CLIP":[6],"have":[7,49],"shown":[8],"impressive":[9],"generalization":[10,189],"capability":[11],"in":[12,58,122,151],"downstream":[13,45,234],"vision":[14,233],"tasks":[15],"with":[16,180],"appropriate":[17],"text":[18],"prompts.":[19],"Instead":[20],"of":[21,86,105,191,203,221,248],"designing":[22],"prompts":[23,35,194],"manually,":[24],"Context":[25],"Optimization":[26],"(CoOp)":[27],"has":[28],"been":[29],"recently":[30],"proposed":[31,250],"to":[32,131,147,186],"learn":[33],"continuous":[34],"using":[36],"task-specific":[37],"training":[38,127,167,200,205],"data.":[39,206],"Despite":[40],"the":[41,55,62,76,87,103,109,123,132,149,154,159,165,172,188,192,199,219,225,246,249],"performance":[42,220],"improvements":[43],"on":[44,65,79,209,230],"tasks,":[46,235],"several":[47],"studies":[48,89],"reported":[50],"that":[51,115,214],"CoOp":[52,116,179,222],"suffers":[53],"from":[54],"overfitting":[56,95,106,135,173],"issue":[57],"two":[59],"aspects:":[60],"(i)":[61],"test":[63,77],"accuracy":[64,78],"base":[66],"classes":[67,81],"first":[68,101],"improves":[69],"and":[70,92,119,125,134,169,223,240],"then":[71],"worsens":[72],"during":[73,164],"training;":[74],"(ii)":[75],"novel":[80,196],"keeps":[82],"decreasing.":[83],"However,":[84],"none":[85],"existing":[88],"can":[90,253],"understand":[91],"mitigate":[93],"problems.":[96],"In":[97,175],"this":[98],"study,":[99],"we":[100,140,177],"explore":[102],"cause":[104],"by":[107,158],"analyzing":[108],"gradient":[110,161],"flow.":[111],"Comparative":[112],"experiments":[113,208],"reveal":[114],"favors":[117],"generalizable":[118],"spurious":[120],"features":[121],"early":[124],"later":[126],"stages,":[128],"respectively,":[129],"leading":[130],"non-overfitting":[133],"phenomena.":[136],"Given":[137],"those":[138],"observations,":[139],"propose":[141],"Subspace":[142],"Prompt":[143],"Tuning":[144],"(Sub":[145],"PT)":[146],"project":[148],"gradients":[150],"back-propagation":[152],"onto":[153,195],"low-rank":[155],"subspace":[156],"spanned":[157],"early-stage":[160],"flow":[162],"eigenvectors":[163],"entire":[166],"process":[168],"successfully":[170],"eliminate":[171],"problem.":[174],"addition,":[176],"equip":[178],"a":[181],"Novel":[182],"Feature":[183],"Learner":[184],"(NFL)":[185],"enhance":[187],"ability":[190],"learned":[193],"categories":[197],"beyond":[198],"set,":[201],"needless":[202],"image":[204],"Extensive":[207],"11":[210],"classification":[211],"datasets":[212],"demonstrate":[213],"Sub":[215],"PT+NFL":[216],"consistently":[217],"boost":[218],"outperform":[224],"state-of-the-art":[226],"CoCoOp":[227],"approach.":[228],"Experiments":[229],"more":[231],"challenging":[232],"including":[236],"open-vocabulary":[237],"object":[238],"detection":[239],"zero-shot":[241],"semantic":[242],"segmentation,":[243],"also":[244],"verify":[245],"effectiveness":[247],"method.":[251],"Codes":[252],"be":[254],"found":[255],"at":[256],"<uri":[257],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[258],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">https://tinyurl.com/mpe64f89</uri>":[259],".":[260]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":19},{"year":2024,"cited_by_count":34},{"year":2023,"cited_by_count":10}],"updated_date":"2026-07-22T07:51:19.307946","created_date":"2025-10-10T00:00:00"}