{"id":"https://openalex.org/W4408323713","doi":"https://doi.org/10.1109/access.2025.3550161","title":"Dimension-Specific Margins and Element-Wise Gradient Scaling for Enhanced Matryoshka Speaker Embedding","display_name":"Dimension-Specific Margins and Element-Wise Gradient Scaling for Enhanced Matryoshka Speaker Embedding","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4408323713","doi":"https://doi.org/10.1109/access.2025.3550161"},"language":"en","primary_location":{"id":"doi:10.1109/access.2025.3550161","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2025.3550161","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1109/access.2025.3550161","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5077698897","display_name":"Sunchan Park","orcid":"https://orcid.org/0009-0004-2145-819X"},"institutions":[{"id":"https://openalex.org/I4921948","display_name":"Pusan National University","ror":"https://ror.org/01an57a31","country_code":"KR","type":"education","lineage":["https://openalex.org/I4921948"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Sunchan Park","raw_affiliation_strings":["Department of Electronics Engineering, Pusan National University, Busan, South Korea"],"raw_orcid":"https://orcid.org/0009-0004-2145-819X","affiliations":[{"raw_affiliation_string":"Department of Electronics Engineering, Pusan National University, Busan, South Korea","institution_ids":["https://openalex.org/I4921948"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5073733529","display_name":"Hyung Soon Kim","orcid":"https://orcid.org/0000-0001-5070-897X"},"institutions":[{"id":"https://openalex.org/I4921948","display_name":"Pusan National University","ror":"https://ror.org/01an57a31","country_code":"KR","type":"education","lineage":["https://openalex.org/I4921948"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Hyung Soon Kim","raw_affiliation_strings":["Department of Electronics Engineering, Pusan National University, Busan, South Korea"],"raw_orcid":"https://orcid.org/0000-0001-5070-897X","affiliations":[{"raw_affiliation_string":"Department of Electronics Engineering, Pusan National University, Busan, South Korea","institution_ids":["https://openalex.org/I4921948"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":1850,"currency":"USD","value_usd":1850},"apc_paid":{"value":1850,"currency":"USD","value_usd":1850},"fwci":3.1921,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.90446117,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":91,"max":99},"biblio":{"volume":"13","issue":null,"first_page":"45473","last_page":"45487"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10860","display_name":"Speech and Audio Processing","score":0.9958000183105469,"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"}},"topics":[{"id":"https://openalex.org/T10860","display_name":"Speech and Audio Processing","score":0.9958000183105469,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9915000200271606,"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/T10901","display_name":"Advanced Data Compression Techniques","score":0.960099995136261,"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/embedding","display_name":"Embedding","score":0.6995601654052734},{"id":"https://openalex.org/keywords/dimension","display_name":"Dimension (graph theory)","score":0.687210738658905},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.669204592704773},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5229378938674927},{"id":"https://openalex.org/keywords/element","display_name":"Element (criminal law)","score":0.4465581178665161},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.2390902042388916},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.20344436168670654},{"id":"https://openalex.org/keywords/geometry","display_name":"Geometry","score":0.16948360204696655},{"id":"https://openalex.org/keywords/combinatorics","display_name":"Combinatorics","score":0.10891318321228027}],"concepts":[{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.6995601654052734},{"id":"https://openalex.org/C33676613","wikidata":"https://www.wikidata.org/wiki/Q13415176","display_name":"Dimension (graph theory)","level":2,"score":0.687210738658905},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.669204592704773},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5229378938674927},{"id":"https://openalex.org/C200288055","wikidata":"https://www.wikidata.org/wiki/Q2621792","display_name":"Element (criminal law)","level":2,"score":0.4465581178665161},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2390902042388916},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.20344436168670654},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.16948360204696655},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.10891318321228027},{"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/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/access.2025.3550161","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2025.3550161","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},{"id":"pmh:oai:doaj.org/article:9dbfcee9266c4a91861446bd36958ecf","is_oa":true,"landing_page_url":"https://doaj.org/article/9dbfcee9266c4a91861446bd36958ecf","pdf_url":null,"source":{"id":"https://openalex.org/S4306401280","display_name":"DOAJ (DOAJ: Directory of Open Access Journals)","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":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"IEEE Access, Vol 13, Pp 45473-45487 (2025)","raw_type":"article"}],"best_oa_location":{"id":"doi:10.1109/access.2025.3550161","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2025.3550161","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":40,"referenced_works":["https://openalex.org/W2096733369","https://openalex.org/W2157161740","https://openalex.org/W2157364932","https://openalex.org/W2194775991","https://openalex.org/W2402146185","https://openalex.org/W2696967604","https://openalex.org/W2726515241","https://openalex.org/W2784163702","https://openalex.org/W2802973008","https://openalex.org/W2808631503","https://openalex.org/W2890964092","https://openalex.org/W2962898354","https://openalex.org/W2963386851","https://openalex.org/W2963466847","https://openalex.org/W2969985801","https://openalex.org/W2972986505","https://openalex.org/W2973190269","https://openalex.org/W3010925296","https://openalex.org/W3013020904","https://openalex.org/W3024869864","https://openalex.org/W3095079700","https://openalex.org/W3096431026","https://openalex.org/W3141035251","https://openalex.org/W3169129566","https://openalex.org/W4283030875","https://openalex.org/W4297841827","https://openalex.org/W4312402191","https://openalex.org/W4319301109","https://openalex.org/W4372267288","https://openalex.org/W4372340947","https://openalex.org/W4372341266","https://openalex.org/W4375869323","https://openalex.org/W4386608742","https://openalex.org/W4388452001","https://openalex.org/W4391021719","https://openalex.org/W4393058549","https://openalex.org/W4402111801","https://openalex.org/W6688816777","https://openalex.org/W6845038366","https://openalex.org/W6873523219"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2081900870","https://openalex.org/W2390279801","https://openalex.org/W2013985456","https://openalex.org/W2951564084","https://openalex.org/W4302984940","https://openalex.org/W2776807270","https://openalex.org/W2392685819"],"abstract_inverted_index":{"Speaker":[0],"embeddings":[1,9,58,172,188],"trained":[2,59],"with":[3,13,60],"Matryoshka":[4],"Representation":[5],"Learning":[6],"(MRL)":[7],"provide":[8],"of":[10,56,146,182,186],"various":[11],"dimensions":[12],"minimal":[14],"overhead,":[15],"adapting":[16],"to":[17,24,75,115,135,164],"different":[18],"computational":[19],"and":[20,78,87,100,143],"storage":[21],"constraints.":[22],"Compared":[23],"single-dimensional":[25],"models,":[26],"MRL-based":[27],"models":[28],"show":[29],"improved":[30],"speaker":[31,57,80,167],"verification":[32,81,168],"performance":[33,89,169,175],"in":[34,42,52,140,170,176],"lower":[35],"dimensions,":[36,93],"but":[37],"there":[38],"is":[39],"some":[40],"degradation":[41],"higher":[43],"dimensions.":[44,125],"Analyzing":[45],"learned":[46,187],"embeddings,":[47],"we":[48,94],"observe":[49],"an":[50,180],"imbalance":[51,86,130],"the":[53,63,110,144,154,159,183,197],"element-wise":[54,101,184,192],"magnitudes":[55,185],"MRL.":[61],"Specifically,":[62],"higher-dimensional":[64,171],"elements":[65],"exhibit":[66],"extremely":[67],"small":[68],"values,":[69],"which":[70],"could":[71],"reduce":[72],"their":[73],"contribution":[74],"cosine":[76],"similarity":[77],"degrade":[79],"performance.":[82],"To":[83],"address":[84],"this":[85],"improve":[88,166],"consistency":[90],"across":[91],"all":[92],"propose":[95],"two":[96],"methods:":[97],"dimension-specific":[98],"margins":[99,105],"gradient":[102,127,193],"scaling.":[103],"Dimension-specific":[104],"stabilize":[106],"training":[107],"by":[108,119,131],"adjusting":[109],"margin":[111],"for":[112,123],"each":[113,136,150],"dimension":[114],"mitigate":[116],"instability":[117],"caused":[118],"excessively":[120],"high":[121],"values":[122],"given":[124],"Element-wise":[126],"scaling":[128,132,194],"mitigates":[129,196],"gradients":[133],"propagated":[134],"element,":[137],"considering":[138],"differences":[139],"embedding":[141],"dimensionality":[142],"number":[145],"loss":[147],"functions":[148],"influencing":[149],"element.":[151],"Evaluation":[152],"on":[153],"VoxCeleb":[155],"benchmark":[156],"shows":[157],"that":[158,191],"proposed":[160],"methods,":[161],"when":[162],"applied":[163],"MRL,":[165],"while":[173],"maintaining":[174],"lower-dimensional":[177],"embeddings.":[178],"Additionally,":[179],"analysis":[181],"visually":[189],"demonstrates":[190],"effectively":[195],"magnitude":[198],"imbalance.":[199]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}