{"id":"https://openalex.org/W4408355292","doi":"https://doi.org/10.1109/icassp49660.2025.10887836","title":"Keep what you need : extracting efficient subnetworks from large audio representation models","display_name":"Keep what you need : extracting efficient subnetworks from large audio representation models","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408355292","doi":"https://doi.org/10.1109/icassp49660.2025.10887836"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10887836","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10887836","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5041371281","display_name":"David Genova","orcid":null},"institutions":[{"id":"https://openalex.org/I1294671590","display_name":"Centre National de la Recherche Scientifique","ror":"https://ror.org/02feahw73","country_code":"FR","type":"government","lineage":["https://openalex.org/I1294671590"]},{"id":"https://openalex.org/I35345632","display_name":"Institut de Recherche et Coordination Acoustique Musique","ror":"https://ror.org/0121jnt59","country_code":"FR","type":"education","lineage":["https://openalex.org/I35345632"]},{"id":"https://openalex.org/I39804081","display_name":"Sorbonne Universit\u00e9","ror":"https://ror.org/02en5vm52","country_code":"FR","type":"education","lineage":["https://openalex.org/I39804081"]},{"id":"https://openalex.org/I4389425508","display_name":"Sciences et Technologies de la Musique et du Son","ror":"https://ror.org/025xvn046","country_code":"FR","type":"facility","lineage":["https://openalex.org/I1294671590","https://openalex.org/I2802818602","https://openalex.org/I35345632","https://openalex.org/I39804081","https://openalex.org/I4210102700","https://openalex.org/I4389425508"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"David Genova","raw_affiliation_strings":["IRCAM - Sorbonne University - CNRS,UMR 9912 STMS,Paris,France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IRCAM - Sorbonne University - CNRS,UMR 9912 STMS,Paris,France","institution_ids":["https://openalex.org/I1294671590","https://openalex.org/I35345632","https://openalex.org/I39804081","https://openalex.org/I4389425508"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085170922","display_name":"Philippe Esling","orcid":"https://orcid.org/0000-0002-1655-7909"},"institutions":[{"id":"https://openalex.org/I1294671590","display_name":"Centre National de la Recherche Scientifique","ror":"https://ror.org/02feahw73","country_code":"FR","type":"government","lineage":["https://openalex.org/I1294671590"]},{"id":"https://openalex.org/I35345632","display_name":"Institut de Recherche et Coordination Acoustique Musique","ror":"https://ror.org/0121jnt59","country_code":"FR","type":"education","lineage":["https://openalex.org/I35345632"]},{"id":"https://openalex.org/I39804081","display_name":"Sorbonne Universit\u00e9","ror":"https://ror.org/02en5vm52","country_code":"FR","type":"education","lineage":["https://openalex.org/I39804081"]},{"id":"https://openalex.org/I4389425508","display_name":"Sciences et Technologies de la Musique et du Son","ror":"https://ror.org/025xvn046","country_code":"FR","type":"facility","lineage":["https://openalex.org/I1294671590","https://openalex.org/I2802818602","https://openalex.org/I35345632","https://openalex.org/I39804081","https://openalex.org/I4210102700","https://openalex.org/I4389425508"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"Philippe Esling","raw_affiliation_strings":["IRCAM - Sorbonne University - CNRS,UMR 9912 STMS,Paris,France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IRCAM - Sorbonne University - CNRS,UMR 9912 STMS,Paris,France","institution_ids":["https://openalex.org/I1294671590","https://openalex.org/I35345632","https://openalex.org/I39804081","https://openalex.org/I4389425508"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5116335834","display_name":"Tom Hurlin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tom Hurlin","raw_affiliation_strings":["Squarp Instruments,Paris,France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Squarp Instruments,Paris,France","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.5078,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.75321956,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11309","display_name":"Music and Audio Processing","score":0.994700014591217,"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/T11309","display_name":"Music and Audio Processing","score":0.994700014591217,"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":0.9696000218391418,"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":0.9624999761581421,"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/computer-science","display_name":"Computer science","score":0.773516058921814},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.6222130060195923},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.3228607475757599}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.773516058921814},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.6222130060195923},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3228607475757599},{"id":"https://openalex.org/C94625758","wikidata":"https://www.wikidata.org/wiki/Q7163","display_name":"Politics","level":2,"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/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/icassp49660.2025.10887836","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10887836","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},{"id":"pmh:oai:HAL:hal-05499231v1","is_oa":false,"landing_page_url":"https://hal.science/hal-05499231","pdf_url":null,"source":{"id":"https://openalex.org/S4306402512","display_name":"HAL (Le Centre pour la Communication Scientifique Directe)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1294671590","host_organization_name":"Centre National de la Recherche Scientifique","host_organization_lineage":["https://openalex.org/I1294671590"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"International Conference on Acoustics, Speech and Signal Processing, Apr 2025, Hyderabad, India. ⟨10.1109/ICASSP49660.2025.10887836⟩","raw_type":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":33,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W1597121597","https://openalex.org/W2038484192","https://openalex.org/W2052666245","https://openalex.org/W2133824856","https://openalex.org/W2156150815","https://openalex.org/W2972584841","https://openalex.org/W3015591594","https://openalex.org/W3094550259","https://openalex.org/W3097777922","https://openalex.org/W3162331882","https://openalex.org/W3162391496","https://openalex.org/W3209059054","https://openalex.org/W4205689591","https://openalex.org/W4226442948","https://openalex.org/W4372266552","https://openalex.org/W4379251869","https://openalex.org/W4392904213","https://openalex.org/W6677103964","https://openalex.org/W6678969435","https://openalex.org/W6736723571","https://openalex.org/W6750665317","https://openalex.org/W6755166560","https://openalex.org/W6759579507","https://openalex.org/W6771400683","https://openalex.org/W6772453130","https://openalex.org/W6780218876","https://openalex.org/W6781288810","https://openalex.org/W6781581198","https://openalex.org/W6791353385","https://openalex.org/W6810673746","https://openalex.org/W6849105126","https://openalex.org/W6852985890"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Recently,":[0],"research":[1],"on":[2,17,64,148,167,208,216,225,236],"audio":[3,31,105,211,227],"foundation":[4,127,176,212],"models":[5,87],"has":[6],"witnessed":[7],"notable":[8],"advances,":[9],"as":[10,231,233],"illustrated":[11],"by":[12],"the":[13,51,59,79,82,136,145,158,172,175,189,198,246],"ever":[14],"improving":[15],"results":[16,247],"complex":[18],"downstream":[19,150],"tasks.":[20],"Subsequently,":[21],"those":[22],"pretrained":[23,140],"networks":[24,63],"have":[25,35],"quickly":[26],"been":[27],"used":[28],"for":[29,71,84,244],"various":[30],"applications.":[32,73],"These":[33],"improvements":[34],"however":[36],"resulted":[37],"in":[38,43],"a":[39,97,115,139,149,154,163,168,217],"considerable":[40],"increase":[41],"both":[42,237],"size":[44],"and":[45,67,221,240,248],"complexity":[46],"of":[47,61,81,104,138,174],"these":[48,86],"models.":[49,128],"Along":[50],"environmental":[52],"concerns":[53],"this":[54,57,75,108,112],"issue":[55,113],"raises,":[56],"prevents":[58],"deployment":[60],"such":[62],"consumer-level":[65],"devices,":[66],"precludes":[68],"their":[69],"use":[70],"real-time":[72],"Moreover,":[74],"appears":[76],"contradictory":[77],"with":[78,114],"specificity":[80],"tasks":[83],"which":[85,90],"are":[88,91,178,251],"used,":[89],"often":[92],"simpler":[93],"compared":[94],"to":[95,120,157],"extracting":[96],"rich,":[98],"multi-purpose":[99],"representation":[100,141,228],"from":[101,125,197],"any":[102],"type":[103],"data.":[106],"In":[107],"paper,":[109],"we":[110,130,152],"address":[111],"simple,":[116],"yet":[117],"effective":[118],"method":[119,207],"extract":[121],"lightweight":[122],"specialist":[123],"subnetworks":[124],"large":[126],"Specifically,":[129],"introduce":[131],"learnable":[132],"binary":[133],"masks":[134],"in-between":[135],"layers":[137],"model.":[142],"When":[143],"training":[144,185],"end-to-end":[146],"model":[147,177],"task,":[151],"add":[153],"sparsity-inducing":[155],"loss":[156],"overall":[159],"objective,":[160],"hence":[161],"learning":[162],"compact":[164],"subnetwork":[165],"specialized":[166],"single":[169],"task.":[170],"Importantly,":[171],"weights":[173],"kept":[179],"frozen,":[180],"resulting":[181],"into":[182],"low":[183],"additional":[184],"costs.":[186],"Once":[187],"trained,":[188],"masked":[190],"computational":[191],"units":[192],"can":[193],"then":[194],"be":[195],"removed":[196],"network,":[199],"implying":[200],"significant":[201],"performance":[202],"gains.":[203],"We":[204],"assess":[205],"our":[206],"three":[209],"widespread":[210],"models,":[213],"each":[214],"based":[215],"different":[218],"backbone":[219],"architecture,":[220],"illustrate":[222],"its":[223,234],"effectiveness":[224],"common":[226],"evaluation":[229],"tasks,":[230],"well":[232],"versatility":[235],"speech,":[238],"music,":[239],"general":[241],"audio.":[242],"Code":[243],"reproducing":[245],"supporting":[249],"webpage":[250],"available":[252],"at":[253],"https://github.com/gnvIRCAM/Audio-representation-trimming.":[254]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}