{"id":"https://openalex.org/W4407169754","doi":"https://doi.org/10.1109/taslpro.2025.3539005","title":"UniPET-SPK: A Unified Framework for Parameter-Efficient Tuning of Pre-Trained Speech Models for Robust Speaker Verification","display_name":"UniPET-SPK: A Unified Framework for Parameter-Efficient Tuning of Pre-Trained Speech Models for Robust Speaker Verification","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4407169754","doi":"https://doi.org/10.1109/taslpro.2025.3539005"},"language":"en","primary_location":{"id":"doi:10.1109/taslpro.2025.3539005","doi":"https://doi.org/10.1109/taslpro.2025.3539005","is_oa":false,"landing_page_url":"https://doi.org/10.1109/taslpro.2025.3539005","pdf_url":null,"source":{"id":"https://openalex.org/S5407042752","display_name":"IEEE Transactions on Audio Speech and Language Processing","issn_l":"2998-4173","issn":["2998-4173"],"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":["jufo-3"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Audio, Speech and Language Processing","raw_type":"journal-article"},"volume":null,"type":"article","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2501.16542","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5005211900","display_name":"Mufan Sang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I162577319","display_name":"The University of Texas at Dallas","ror":"https://ror.org/049emcs32","country_code":"US","type":"education","lineage":["https://openalex.org/I162577319"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mufan Sang","raw_affiliation_strings":["Center for Robust Speech Systems (CRSS), Erik Jonsson School of Engineering & Computer Science, University of Texas at Dallas, Richardson, TX, USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Center for Robust Speech Systems (CRSS), Erik Jonsson School of Engineering & Computer Science, University of Texas at Dallas, Richardson, TX, USA","institution_ids":["https://openalex.org/I162577319"]}]},{"author_position":"last","author":{"id":null,"display_name":"John H. L. Hansen","orcid":"https://orcid.org/0000-0003-1382-9929","observed_orcids":["https://orcid.org/0000-0003-1382-9929"]},"institutions":[{"id":"https://openalex.org/I162577319","display_name":"The University of Texas at Dallas","ror":"https://ror.org/049emcs32","country_code":"US","type":"education","lineage":["https://openalex.org/I162577319"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"John H. L. Hansen","raw_affiliation_strings":["Center for Robust Speech Systems (CRSS), Erik Jonsson School of Engineering & Computer Science, University of Texas at Dallas, Richardson, TX, USA"],"raw_orcid":"https://orcid.org/0000-0003-1382-9929","affiliations":[{"raw_affiliation_string":"Center for Robust Speech Systems (CRSS), Erik Jonsson School of Engineering & Computer Science, University of Texas at Dallas, Richardson, TX, USA","institution_ids":["https://openalex.org/I162577319"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I162577319"],"apc_list":{"value":2645,"currency":"USD","value_usd":2645},"apc_paid":null,"fwci":1.9984,"has_fulltext":true,"cited_by_count":5,"citation_normalized_percentile":{"value":0.86060737,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":98},"biblio":{"volume":"33","issue":null,"first_page":"1402","last_page":"1414"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9945999979972839,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9945999979972839,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.002099999925121665,"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/T10860","display_name":"Speech and Audio Processing","score":0.0013000000035390258,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/parameter-efficient-fine-tuning","display_name":"parameter-efficient fine-tuning","score":0.9300000071525574},{"id":"https://openalex.org/keywords/speaker-verification","display_name":"speaker verification","score":0.9210000038146973},{"id":"https://openalex.org/keywords/self-supervised-speech-models","display_name":"self-supervised speech models","score":0.8980000019073486},{"id":"https://openalex.org/keywords/adapter-tuning","display_name":"adapter tuning","score":0.6299999952316284},{"id":"https://openalex.org/keywords/prompt-tuning","display_name":"prompt tuning","score":0.9879999756813049}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7982559204101562},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6662611961364746},{"id":"https://openalex.org/C2982762665","wikidata":"https://www.wikidata.org/wiki/Q1145189","display_name":"Speaker verification","level":3,"score":0.5735053420066833},{"id":"https://openalex.org/C61328038","wikidata":"https://www.wikidata.org/wiki/Q3358061","display_name":"Speech processing","level":2,"score":0.5316399335861206},{"id":"https://openalex.org/C133892786","wikidata":"https://www.wikidata.org/wiki/Q1145189","display_name":"Speaker recognition","level":2,"score":0.48594892024993896},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.4246850609779358},{"id":"https://openalex.org/C14999030","wikidata":"https://www.wikidata.org/wiki/Q16346","display_name":"Speech synthesis","level":2,"score":0.41540223360061646},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3754483759403229},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/taslpro.2025.3539005","doi":"https://doi.org/10.1109/taslpro.2025.3539005","is_oa":false,"landing_page_url":"https://doi.org/10.1109/taslpro.2025.3539005","pdf_url":null,"source":{"id":"https://openalex.org/S5407042752","display_name":"IEEE Transactions on Audio Speech and Language Processing","issn_l":"2998-4173","issn":["2998-4173"],"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":["jufo-3"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Audio, Speech and Language Processing","raw_type":"journal-article"},{"id":"pmh:oai:arXiv.org:2501.16542","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.16542","pdf_url":"https://arxiv.org/pdf/2501.16542","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2501.16542","doi":"https://doi.org/10.48550/arxiv.2501.16542","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2501.16542","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2501.16542","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.16542","pdf_url":"https://arxiv.org/pdf/2501.16542","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320327708","display_name":"University of Texas at Dallas","ror":"https://ror.org/049emcs32"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4407169754.pdf","grobid_xml":"https://content.openalex.org/works/W4407169754.grobid-xml"},"referenced_works_count":60,"referenced_works":["https://openalex.org/W1522301498","https://openalex.org/W1533861849","https://openalex.org/W1936725236","https://openalex.org/W2626778328","https://openalex.org/W2726515241","https://openalex.org/W2760938034","https://openalex.org/W2784874046","https://openalex.org/W2808631503","https://openalex.org/W2890964092","https://openalex.org/W2900022717","https://openalex.org/W2913946806","https://openalex.org/W2939710050","https://openalex.org/W2971840980","https://openalex.org/W2972552635","https://openalex.org/W2981005517","https://openalex.org/W3015598461","https://openalex.org/W3024869864","https://openalex.org/W3030163527","https://openalex.org/W3096730275","https://openalex.org/W3097470351","https://openalex.org/W3110524561","https://openalex.org/W3135367836","https://openalex.org/W3139080614","https://openalex.org/W3152956381","https://openalex.org/W3160397447","https://openalex.org/W3163829335","https://openalex.org/W3165404421","https://openalex.org/W3169320628","https://openalex.org/W3185341429","https://openalex.org/W3196974791","https://openalex.org/W3197580070","https://openalex.org/W3198377975","https://openalex.org/W3205949070","https://openalex.org/W3206252155","https://openalex.org/W3206287456","https://openalex.org/W3206816211","https://openalex.org/W3209984917","https://openalex.org/W3217778739","https://openalex.org/W4200629441","https://openalex.org/W4200633018","https://openalex.org/W4205234379","https://openalex.org/W4221154745","https://openalex.org/W4225274946","https://openalex.org/W4225747866","https://openalex.org/W4226162428","https://openalex.org/W4283073456","https://openalex.org/W4285247752","https://openalex.org/W4296068991","https://openalex.org/W4320561328","https://openalex.org/W4372266902","https://openalex.org/W4372270069","https://openalex.org/W4375869415","https://openalex.org/W4385484947","https://openalex.org/W4392902886","https://openalex.org/W4392904027","https://openalex.org/W6769178842","https://openalex.org/W6780218876","https://openalex.org/W6796581206","https://openalex.org/W6850477478","https://openalex.org/W7133219196"],"related_works":["https://openalex.org/W4392904027","https://openalex.org/W4372266902","https://openalex.org/W4392902886","https://openalex.org/W4319862271","https://openalex.org/W7151888762","https://openalex.org/W4415433842","https://openalex.org/W4408325495","https://openalex.org/W7223032420","https://openalex.org/W4385823373","https://openalex.org/W4402112384"],"abstract_inverted_index":{"With":[0],"excellent":[1],"generalization":[2,283],"ability,":[3],"self-supervised":[4,63],"speech":[5,14,64],"models":[6,28,65,161],"have":[7],"shown":[8],"impressive":[9],"performance":[10,260],"on":[11,53,216,231,272],"various":[12],"downstream":[13],"tasks":[15],"in":[16,292],"the":[17,24,44,67,106,126,145,156,168,197,221,224,232,236,242,248,266,273,280,286],"pre-training":[18],"and":[19,38,86,95,130,207,235,253,275,282,295],"fine-tuning":[20,30],"paradigm.":[21],"However,":[22],"as":[23,41,43],"size":[25],"of":[26,46,114,122,159,200,214,223,265,285],"pre-trained":[27,62,117,160],"grows,":[29],"becomes":[31],"practically":[32],"unfeasible":[33],"due":[34],"to":[35,66,162,195,203,219,278],"expanding":[36],"computation":[37],"storage":[39],"requirements,":[40],"well":[42],"risk":[45],"overfitting.":[47],"In":[48],"this":[49],"study,":[50],"we":[51,72,104,143,166],"concentrate":[52],"exploring":[54],"parameter-efficient":[55,75,226,255],"tuning":[56,76,227,256],"(PET)":[57],"methods":[58,180,202,288],"for":[59,120,289],"adapting":[60],"large-scale":[61],"speaker":[68,290],"verification":[69,291],"task.":[70],"Correspondingly,":[71],"propose":[73,105,144,167],"three":[74],"methods:":[77],"i)":[78],"an":[79],"adapter-tuning":[80,94],"method,":[81,85],"ii)":[82],"a":[83,88,98,138,170,182,186,211],"prompt-tuning":[84,96],"iii)":[87],"unified":[89,171],"framework":[90,172,184],"that":[91,150,173,241],"effectively":[92,174],"incorporates":[93,175],"with":[97,185],"dynamically":[99],"learnable":[100],"gating":[101,189],"mechanism.":[102,190],"First,":[103],"Inner+Inter":[107],"Adapter":[108],"framework,":[109],"which":[110],"inserts":[111],"two":[112,177,249],"types":[113],"adapters":[115],"into":[116,155,181],"models,":[118],"allowing":[119],"adaptation":[121],"latent":[123],"features":[124],"within":[125],"intermediate":[127],"Transformer":[128,135],"layers":[129],"output":[131],"embeddings":[132],"from":[133],"all":[134],"layers,":[136],"through":[137],"parallel":[139],"adapter":[140],"design.":[141],"Second,":[142],"Deep":[146],"Speaker":[147],"Prompting":[148],"method":[149],"concatenates":[151],"trainable":[152,188],"prompt":[153],"tokens":[154],"input":[157],"space":[158],"guide":[163],"adaptation.":[164],"Lastly,":[165],"UniPET-SPK,":[169],"these":[176],"alternate":[178],"PET":[179,201,250],"single":[183],"dynamic":[187],"The":[191],"proposed":[192,225,243,287],"UniPET-SPK":[193,244],"learns":[194],"find":[196],"optimal":[198],"mixture":[199],"match":[204],"different":[205,293],"datasets":[206,218,239,277],"scenarios.":[208,298],"We":[209,268],"conduct":[210,270],"comprehensive":[212],"set":[213],"experiments":[215,271],"several":[217],"validate":[220],"effectiveness":[222],"methods.":[228],"Experimental":[229],"results":[230],"VoxCeleb,":[233],"CN-Celeb,":[234],"1$^{\\mathrm{{st}}}$48-UTD":[237,276],"forensic":[238],"demonstrate":[240,279],"can":[245],"consistently":[246],"outperform":[247],"methods,":[251,257],"fine-tuning,":[252],"other":[254],"achieving":[258],"superior":[259],"while":[261],"updating":[262],"only":[263],"5.4%":[264],"parameters.":[267],"further":[269],"CN-Celeb":[274],"robustness":[281],"ability":[284],"languages":[294],"more":[296],"challenging":[297]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":2}],"updated_date":"2026-10-29T09:56:32.521200","created_date":"2025-02-06T00:00:00"}