{"id":"https://openalex.org/W4408355220","doi":"https://doi.org/10.1109/icassp49660.2025.10887737","title":"Evaluation of Deep Audio Representations for Hearables","display_name":"Evaluation of Deep Audio Representations for Hearables","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408355220","doi":"https://doi.org/10.1109/icassp49660.2025.10887737"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10887737","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10887737","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":"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/A5062704748","display_name":"Fabian Gr\u00f6ger","orcid":"https://orcid.org/0000-0002-9699-688X"},"institutions":[{"id":"https://openalex.org/I81007117","display_name":"Lucerne University of Applied Sciences and Arts","ror":"https://ror.org/04nd0xd48","country_code":"CH","type":"education","lineage":["https://openalex.org/I81007117"]}],"countries":["CH"],"is_corresponding":true,"raw_author_name":"Fabian Gr\u00f6ger","raw_affiliation_strings":["Lucerne University of Applied Sciences and Arts,Rotkreuz,Switzerland"],"affiliations":[{"raw_affiliation_string":"Lucerne University of Applied Sciences and Arts,Rotkreuz,Switzerland","institution_ids":["https://openalex.org/I81007117"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051473693","display_name":"Pascal Baumann","orcid":"https://orcid.org/0000-0002-9371-0807"},"institutions":[{"id":"https://openalex.org/I81007117","display_name":"Lucerne University of Applied Sciences and Arts","ror":"https://ror.org/04nd0xd48","country_code":"CH","type":"education","lineage":["https://openalex.org/I81007117"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Pascal Baumann","raw_affiliation_strings":["Lucerne University of Applied Sciences and Arts,Rotkreuz,Switzerland"],"affiliations":[{"raw_affiliation_string":"Lucerne University of Applied Sciences and Arts,Rotkreuz,Switzerland","institution_ids":["https://openalex.org/I81007117"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005775483","display_name":"Ludovic Amruthalingam","orcid":"https://orcid.org/0000-0001-5980-5469"},"institutions":[{"id":"https://openalex.org/I81007117","display_name":"Lucerne University of Applied Sciences and Arts","ror":"https://ror.org/04nd0xd48","country_code":"CH","type":"education","lineage":["https://openalex.org/I81007117"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Ludovic Amruthalingam","raw_affiliation_strings":["Lucerne University of Applied Sciences and Arts,Rotkreuz,Switzerland"],"affiliations":[{"raw_affiliation_string":"Lucerne University of Applied Sciences and Arts,Rotkreuz,Switzerland","institution_ids":["https://openalex.org/I81007117"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025996497","display_name":"Laurent Simon","orcid":"https://orcid.org/0000-0002-8439-8081"},"institutions":[{"id":"https://openalex.org/I4210114179","display_name":"Sonova (Switzerland)","ror":"https://ror.org/027rr8795","country_code":"CH","type":"company","lineage":["https://openalex.org/I4210114179"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Laurent Simon","raw_affiliation_strings":["Sonova AG,Stäfa,Switzerland"],"affiliations":[{"raw_affiliation_string":"Sonova AG,Stäfa,Switzerland","institution_ids":["https://openalex.org/I4210114179"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5049999256","display_name":"Ruksana Giurda","orcid":null},"institutions":[{"id":"https://openalex.org/I4210114179","display_name":"Sonova (Switzerland)","ror":"https://ror.org/027rr8795","country_code":"CH","type":"company","lineage":["https://openalex.org/I4210114179"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Ruksana Giurda","raw_affiliation_strings":["Sonova AG,Stäfa,Switzerland"],"affiliations":[{"raw_affiliation_string":"Sonova AG,Stäfa,Switzerland","institution_ids":["https://openalex.org/I4210114179"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5005954103","display_name":"Simone Lionetti","orcid":"https://orcid.org/0000-0001-7305-8957"},"institutions":[{"id":"https://openalex.org/I81007117","display_name":"Lucerne University of Applied Sciences and Arts","ror":"https://ror.org/04nd0xd48","country_code":"CH","type":"education","lineage":["https://openalex.org/I81007117"]}],"countries":["CH"],"is_corresponding":false,"raw_author_name":"Simone Lionetti","raw_affiliation_strings":["Lucerne University of Applied Sciences and Arts,Rotkreuz,Switzerland"],"affiliations":[{"raw_affiliation_string":"Lucerne University of Applied Sciences and Arts,Rotkreuz,Switzerland","institution_ids":["https://openalex.org/I81007117"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5062704748"],"corresponding_institution_ids":["https://openalex.org/I81007117"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.05799354,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"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/T10860","display_name":"Speech and Audio Processing","score":0.9192000031471252,"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":0.9192000031471252,"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":0.7413243055343628},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.44943299889564514},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3544130027294159}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7413243055343628},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.44943299889564514},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3544130027294159}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49660.2025.10887737","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10887737","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"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":37,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W2052666245","https://openalex.org/W2075229564","https://openalex.org/W2129354862","https://openalex.org/W2163415800","https://openalex.org/W2560753069","https://openalex.org/W2566935005","https://openalex.org/W2593116425","https://openalex.org/W2780646907","https://openalex.org/W2789662602","https://openalex.org/W2893969570","https://openalex.org/W2973049979","https://openalex.org/W3005580638","https://openalex.org/W3006926732","https://openalex.org/W3010162163","https://openalex.org/W3119308075","https://openalex.org/W3148821605","https://openalex.org/W3159481202","https://openalex.org/W3166682767","https://openalex.org/W3189296823","https://openalex.org/W3197580070","https://openalex.org/W3198694222","https://openalex.org/W3203613879","https://openalex.org/W3209059054","https://openalex.org/W3209984917","https://openalex.org/W3210163254","https://openalex.org/W4225713393","https://openalex.org/W4293370787","https://openalex.org/W4311137818","https://openalex.org/W4321002349","https://openalex.org/W4362602198","https://openalex.org/W4401597538","https://openalex.org/W6755207826","https://openalex.org/W6780218876","https://openalex.org/W6809947431","https://openalex.org/W6848208918","https://openalex.org/W6858100026"],"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":{"Effectively":[0],"steering":[1],"hearable":[2,157],"devices":[3],"requires":[4],"understanding":[5],"the":[6,10,13,24,27,46,53,96,106,121,131,152],"acoustic":[7,61,87,103],"environment":[8,153],"around":[9],"user.":[11],"In":[12],"computational":[14],"analysis":[15],"of":[16,26,42,55,85,105,112,133,147],"sound":[17],"scenes,":[18],"foundation":[19,56],"models":[20,57,134],"have":[21],"emerged":[22],"as":[23],"state":[25],"art":[28],"to":[29,51,143],"produce":[30],"high-performance,":[31],"robust,":[32],"multi-purpose":[33],"audio":[34,69,107,115,138],"representations.":[35],"We":[36],"introduce":[37],"and":[38,49,101,162],"release":[39],"Deep":[40],"Evaluation":[41],"Audio":[43],"Representations":[44],"(DEAR),":[45],"first":[47],"dataset":[48,66,161],"benchmark":[50,90],"evaluate":[52],"efficacy":[54],"in":[58],"capturing":[59],"essential":[60],"properties":[62,104,154],"for":[63,156],"hearables.":[64],"The":[65,159],"includes":[67],"1,158":[68],"tracks,":[70],"each":[71],"30":[72],"seconds":[73],"long,":[74],"created":[75],"by":[76],"spatially":[77],"mixing":[78],"proprietary":[79],"monologues":[80],"with":[81],"commercial,":[82],"high-quality":[83],"recordings":[84],"everyday":[86],"scenes.":[88,108],"Our":[89],"encompasses":[91],"eight":[92],"tasks":[93],"that":[94,120],"assess":[95],"general":[97],"context,":[98],"speech":[99],"sources,":[100],"technical":[102],"Through":[109],"our":[110],"evaluation":[111],"four":[113],"general-purpose":[114],"representation":[116],"models,":[117],"we":[118],"demonstrate":[119],"BEATs":[122],"model":[123],"significantly":[124],"surpasses":[125],"its":[126],"counterparts.":[127],"This":[128],"superiority":[129],"underscores":[130],"advantage":[132],"trained":[135],"on":[136],"diverse":[137],"collections,":[139],"confirming":[140],"their":[141],"applicability":[142],"a":[144],"wide":[145],"array":[146],"auditory":[148],"tasks,":[149],"including":[150],"encoding":[151],"necessary":[155],"steering.":[158],"DEAR":[160],"associated":[163],"code":[164],"are":[165],"available":[166],"at":[167],"https://dear-dataset.github.io.":[168]},"counts_by_year":[],"updated_date":"2025-12-28T23:10:05.387466","created_date":"2025-10-10T00:00:00"}