{"id":"https://openalex.org/W3025125834","doi":"https://doi.org/10.21437/odyssey.2020-50","title":"Dropping Classes for Deep Speaker Representation Learning","display_name":"Dropping Classes for Deep Speaker Representation Learning","publication_year":2020,"publication_date":"2020-05-15","ids":{"openalex":"https://openalex.org/W3025125834","doi":"https://doi.org/10.21437/odyssey.2020-50","mag":"3025125834"},"language":"en","primary_location":{"id":"doi:10.21437/odyssey.2020-50","is_oa":false,"landing_page_url":"https://doi.org/10.21437/odyssey.2020-50","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"The Speaker and Language Recognition Workshop (Odyssey 2020)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.research.ed.ac.uk/en/publications/1921aabc-9b66-44ba-804a-7f12bbb6f986","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5020396429","display_name":"Chau Luu","orcid":null},"institutions":[{"id":"https://openalex.org/I98677209","display_name":"University of Edinburgh","ror":"https://ror.org/01nrxwf90","country_code":"GB","type":"education","lineage":["https://openalex.org/I98677209"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Chau Luu","raw_affiliation_strings":["Centre for Speech Technology Research University of Edinburgh, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Centre for Speech Technology Research University of Edinburgh, UK","institution_ids":["https://openalex.org/I98677209"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102911387","display_name":"Peter Bell","orcid":"https://orcid.org/0000-0002-9597-9615"},"institutions":[{"id":"https://openalex.org/I98677209","display_name":"University of Edinburgh","ror":"https://ror.org/01nrxwf90","country_code":"GB","type":"education","lineage":["https://openalex.org/I98677209"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Peter Bell","raw_affiliation_strings":["Centre for Speech Technology Research University of Edinburgh, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Centre for Speech Technology Research University of Edinburgh, UK","institution_ids":["https://openalex.org/I98677209"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5027442277","display_name":"Steve Renals","orcid":"https://orcid.org/0000-0002-8790-3389"},"institutions":[{"id":"https://openalex.org/I98677209","display_name":"University of Edinburgh","ror":"https://ror.org/01nrxwf90","country_code":"GB","type":"education","lineage":["https://openalex.org/I98677209"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Steve Renals","raw_affiliation_strings":["Centre for Speech Technology Research University of Edinburgh, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Centre for Speech Technology Research University of Edinburgh, UK","institution_ids":["https://openalex.org/I98677209"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I98677209"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":4,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"357","last_page":"364"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9998999834060669,"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":0.9998999834060669,"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":0.9986000061035156,"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"}},{"id":"https://openalex.org/T11309","display_name":"Music and Audio Processing","score":0.996399998664856,"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/discriminative-model","display_name":"Discriminative model","score":0.8031277656555176},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.726828932762146},{"id":"https://openalex.org/keywords/word-error-rate","display_name":"Word error rate","score":0.6674622893333435},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.6054974794387817},{"id":"https://openalex.org/keywords/margin","display_name":"Margin (machine learning)","score":0.5738734006881714},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5666401386260986},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.5407962203025818},{"id":"https://openalex.org/keywords/extractor","display_name":"Extractor","score":0.5398887991905212},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.5389028787612915},{"id":"https://openalex.org/keywords/feature-learning","display_name":"Feature learning","score":0.48790594935417175},{"id":"https://openalex.org/keywords/feature-extraction","display_name":"Feature extraction","score":0.4711015820503235},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.47096922993659973},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.46132248640060425},{"id":"https://openalex.org/keywords/layer","display_name":"Layer (electronics)","score":0.44644662737846375},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.36073291301727295}],"concepts":[{"id":"https://openalex.org/C97931131","wikidata":"https://www.wikidata.org/wiki/Q5282087","display_name":"Discriminative model","level":2,"score":0.8031277656555176},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.726828932762146},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.6674622893333435},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.6054974794387817},{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.5738734006881714},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5666401386260986},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.5407962203025818},{"id":"https://openalex.org/C117978034","wikidata":"https://www.wikidata.org/wiki/Q5422192","display_name":"Extractor","level":2,"score":0.5398887991905212},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.5389028787612915},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.48790594935417175},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.4711015820503235},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.47096922993659973},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.46132248640060425},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.44644662737846375},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.36073291301727295},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C21880701","wikidata":"https://www.wikidata.org/wiki/Q2144042","display_name":"Process engineering","level":1,"score":0.0},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0},{"id":"https://openalex.org/C94625758","wikidata":"https://www.wikidata.org/wiki/Q7163","display_name":"Politics","level":2,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C178790620","wikidata":"https://www.wikidata.org/wiki/Q11351","display_name":"Organic chemistry","level":1,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.0},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.21437/odyssey.2020-50","is_oa":false,"landing_page_url":"https://doi.org/10.21437/odyssey.2020-50","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"The Speaker and Language Recognition Workshop (Odyssey 2020)","raw_type":"proceedings-article"},{"id":"pmh:oai:pure.ed.ac.uk:openaire/1921aabc-9b66-44ba-804a-7f12bbb6f986","is_oa":true,"landing_page_url":"https://www.research.ed.ac.uk/en/publications/1921aabc-9b66-44ba-804a-7f12bbb6f986","pdf_url":null,"source":{"id":"https://openalex.org/S4406922455","display_name":"Edinburgh Research Explorer","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Luu, C, Bell, P & Renals, S 2020, DropClass and DropAdapt: Dropping classes for deep speaker representation learning. in Proceedings of Odyssey 2020 The Speaker and Language Recognition Workshop. International Speech Communication Association, pp. 357-364, Odyssey 2020 The Speaker and Language Recognition Workshop , Tokyo, Japan, 1/11/20. https://doi.org/10.21437/Odyssey.2020-50","raw_type":"contributionToPeriodical"},{"id":"pmh:oai:pure.ed.ac.uk:publications/1921aabc-9b66-44ba-804a-7f12bbb6f986","is_oa":false,"landing_page_url":"https://www.research.ed.ac.uk/portal/en/publications/dropclass-and-dropadapt-dropping-classes-for-deep-speaker-representation-learning(1921aabc-9b66-44ba-804a-7f12bbb6f986).html","pdf_url":null,"source":{"id":"https://openalex.org/S4406922455","display_name":"Edinburgh Research Explorer","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":""}],"best_oa_location":{"id":"pmh:oai:pure.ed.ac.uk:openaire/1921aabc-9b66-44ba-804a-7f12bbb6f986","is_oa":true,"landing_page_url":"https://www.research.ed.ac.uk/en/publications/1921aabc-9b66-44ba-804a-7f12bbb6f986","pdf_url":null,"source":{"id":"https://openalex.org/S4406922455","display_name":"Edinburgh Research Explorer","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Luu, C, Bell, P & Renals, S 2020, DropClass and DropAdapt: Dropping classes for deep speaker representation learning. in Proceedings of Odyssey 2020 The Speaker and Language Recognition Workshop. International Speech Communication Association, pp. 357-364, Odyssey 2020 The Speaker and Language Recognition Workshop , Tokyo, Japan, 1/11/20. https://doi.org/10.21437/Odyssey.2020-50","raw_type":"contributionToPeriodical"},"sustainable_development_goals":[{"score":0.7099999785423279,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":36,"referenced_works":["https://openalex.org/W1485147275","https://openalex.org/W1936725236","https://openalex.org/W1982696459","https://openalex.org/W2046056978","https://openalex.org/W2060277733","https://openalex.org/W2062227835","https://openalex.org/W2095705004","https://openalex.org/W2114925438","https://openalex.org/W2150769028","https://openalex.org/W2164370980","https://openalex.org/W2516764878","https://openalex.org/W2604763608","https://openalex.org/W2726515241","https://openalex.org/W2748488820","https://openalex.org/W2767106145","https://openalex.org/W2794506738","https://openalex.org/W2808631503","https://openalex.org/W2889418727","https://openalex.org/W2890964092","https://openalex.org/W2916104401","https://openalex.org/W2916301830","https://openalex.org/W2962788625","https://openalex.org/W2962898354","https://openalex.org/W2963212406","https://openalex.org/W2963466847","https://openalex.org/W2963609956","https://openalex.org/W2963775347","https://openalex.org/W2963814162","https://openalex.org/W2964050365","https://openalex.org/W2964067969","https://openalex.org/W2964099985","https://openalex.org/W2969985801","https://openalex.org/W2972680151","https://openalex.org/W2979593053","https://openalex.org/W2995049146","https://openalex.org/W3103152812"],"related_works":["https://openalex.org/W4389116644","https://openalex.org/W2153315159","https://openalex.org/W3103844505","https://openalex.org/W2162582511","https://openalex.org/W3119773509","https://openalex.org/W3208297503","https://openalex.org/W2889153461","https://openalex.org/W2964117661","https://openalex.org/W4388405611","https://openalex.org/W2619127353"],"abstract_inverted_index":{"Many":[0],"recent":[1],"works":[2,89],"on":[3,12,63,114,143,175,212],"deep":[4],"speaker":[5,81],"embeddings":[6,50],"train":[7],"their":[8],"feature":[9,111],"extraction":[10],"networks":[11],"large":[13,206],"classification":[14,117],"tasks,":[15],"distinguishing":[16],"between":[17],"all":[18],"speakers":[19,163,187],"in":[20,80,109,134,164,210],"a":[21,93,110,130,140,151,155,159,173,205],"training":[22,49,100],"set.":[23],"Empirically,":[24],"this":[25,43,126,217],"has":[26,219],"been":[27,220],"shown":[28],"to":[29,59,158],"produce":[30,180],"speaker-discriminative":[31],"embeddings,":[32,61],"even":[33],"for":[34,216],"unseen":[35],"speakers.":[36],"However,":[37],"it":[38],"is":[39,44,150,169],"not":[40],"clear":[41],"that":[42,51,73],"the":[45,64,99,103,185,195,199],"optimal":[46],"means":[47,152],"of":[48,66,96,153,161],"generalize":[52],"well.":[53],"This":[54,168,202],"work":[55],"proposes":[56],"two":[57],"approaches":[58,75],"learning":[60],"based":[62],"notion":[65],"dropping":[67,92,194],"classes":[68,97,178],"during":[69],"training.":[70],"We":[71],"demonstrate":[72],"both":[74],"can":[76,128],"yield":[77,129],"performance":[78],"gains":[79],"verification":[82],"tasks.":[83,118],"The":[84,145,214],"first":[85],"proposed":[86,147],"method,":[87,148],"DropClass,":[88],"via":[90],"periodically":[91],"random":[94],"subset":[95],"from":[98,198],"data":[101],"and":[102],"output":[104,200],"layer":[105],"throughout":[106],"training,":[107],"resulting":[108],"extractor":[112],"trained":[113,156],"many":[115],"different":[116],"Combined":[119],"with":[120],"an":[121,165],"additive":[122],"angular":[123],"margin":[124],"loss,":[125],"method":[127,203],"7.9%":[131],"relative":[132,208],"improvement":[133,209],"equal":[135],"error":[136],"rate":[137],"(EER)":[138],"over":[139],"strong":[141],"baseline":[142],"VoxCeleb.":[144,213],"second":[146],"DropAdapt,":[149],"adapting":[154],"model":[157,174],"set":[160],"enrolment":[162,186],"unsupervised":[166],"manner.":[167],"performed":[170],"by":[171],"fine-tuning":[172],"only":[176],"those":[177],"which":[179],"high":[181],"probability":[182],"predictions":[183],"when":[184],"are":[188],"used":[189],"as":[190],"input,":[191],"again":[192],"also":[193],"relevant":[196],"rows":[197],"layer.":[201],"yields":[204],"13.2%":[207],"EER":[211],"code":[215],"paper":[218],"made":[221],"publicly":[222],"available.":[223]},"counts_by_year":[{"year":2022,"cited_by_count":1},{"year":2021,"cited_by_count":1},{"year":2020,"cited_by_count":2}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}