{"id":"https://openalex.org/W4372260579","doi":"https://doi.org/10.1109/icassp49357.2023.10095255","title":"Dataset Balancing Can Hurt Model Performance","display_name":"Dataset Balancing Can Hurt Model Performance","publication_year":2023,"publication_date":"2023-05-05","ids":{"openalex":"https://openalex.org/W4372260579","doi":"https://doi.org/10.1109/icassp49357.2023.10095255"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49357.2023.10095255","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10095255","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2307.00079","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5110435728","display_name":"Robert C. Moore","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"R. Channing Moore","raw_affiliation_strings":["Google Research,Mountain View,CA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google Research,Mountain View,CA,USA","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5055458377","display_name":"Daniel P. W. Ellis","orcid":"https://orcid.org/0000-0002-5076-4311"},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Daniel P. W. Ellis","raw_affiliation_strings":["Google Research,Mountain View,CA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google Research,Mountain View,CA,USA","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101918091","display_name":"Eduardo Fonseca","orcid":"https://orcid.org/0009-0002-7098-9419"},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Eduardo Fonseca","raw_affiliation_strings":["Google Research,Mountain View,CA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google Research,Mountain View,CA,USA","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014428891","display_name":"Shawn Hershey","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Shawn Hershey","raw_affiliation_strings":["Google Research,Mountain View,CA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google Research,Mountain View,CA,USA","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103622427","display_name":"Aren Jansen","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Aren Jansen","raw_affiliation_strings":["Google Research,Mountain View,CA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google Research,Mountain View,CA,USA","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5059222327","display_name":"Manoj Plakal","orcid":"https://orcid.org/0000-0001-6759-3873"},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Manoj Plakal","raw_affiliation_strings":["Google Research,Mountain View,CA,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google Research,Mountain View,CA,USA","institution_ids":["https://openalex.org/I1291425158"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I1291425158"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":11,"citation_normalized_percentile":null,"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/T11309","display_name":"Music and Audio Processing","score":0.9998999834060669,"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":0.9998999834060669,"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/T10860","display_name":"Speech and Audio Processing","score":0.9940000176429749,"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":0.989300012588501,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7945400476455688},{"id":"https://openalex.org/keywords/class","display_name":"Class (philosophy)","score":0.5739344954490662},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.553045392036438},{"id":"https://openalex.org/keywords/prior-probability","display_name":"Prior probability","score":0.522328794002533},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.49342480301856995},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.4599624574184418},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4541482627391815},{"id":"https://openalex.org/keywords/load-balancing","display_name":"Load balancing (electrical power)","score":0.4471375346183777},{"id":"https://openalex.org/keywords/range","display_name":"Range (aeronautics)","score":0.43398359417915344},{"id":"https://openalex.org/keywords/performance-improvement","display_name":"Performance improvement","score":0.4101082682609558},{"id":"https://openalex.org/keywords/data-mining","display_name":"Data mining","score":0.38858768343925476},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.09009996056556702},{"id":"https://openalex.org/keywords/operations-management","display_name":"Operations management","score":0.07459554076194763},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.07235556840896606}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7945400476455688},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.5739344954490662},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.553045392036438},{"id":"https://openalex.org/C177769412","wikidata":"https://www.wikidata.org/wiki/Q278090","display_name":"Prior probability","level":3,"score":0.522328794002533},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.49342480301856995},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.4599624574184418},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4541482627391815},{"id":"https://openalex.org/C138959212","wikidata":"https://www.wikidata.org/wiki/Q1806783","display_name":"Load balancing (electrical power)","level":3,"score":0.4471375346183777},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.43398359417915344},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.4101082682609558},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.38858768343925476},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.09009996056556702},{"id":"https://openalex.org/C21547014","wikidata":"https://www.wikidata.org/wiki/Q1423657","display_name":"Operations management","level":1,"score":0.07459554076194763},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.07235556840896606},{"id":"https://openalex.org/C146978453","wikidata":"https://www.wikidata.org/wiki/Q3798668","display_name":"Aerospace engineering","level":1,"score":0.0},{"id":"https://openalex.org/C187691185","wikidata":"https://www.wikidata.org/wiki/Q2020720","display_name":"Grid","level":2,"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/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/icassp49357.2023.10095255","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10095255","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2307.00079","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2307.00079","pdf_url":"https://arxiv.org/pdf/2307.00079","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2307.00079","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2307.00079","pdf_url":"https://arxiv.org/pdf/2307.00079","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","score":0.6000000238418579,"display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4372260579.pdf","grobid_xml":"https://content.openalex.org/works/W4372260579.grobid-xml"},"referenced_works_count":24,"referenced_works":["https://openalex.org/W587635713","https://openalex.org/W1522301498","https://openalex.org/W1821462560","https://openalex.org/W2117539524","https://openalex.org/W2118978333","https://openalex.org/W2339172597","https://openalex.org/W2526050071","https://openalex.org/W2593116425","https://openalex.org/W2765407302","https://openalex.org/W2896791666","https://openalex.org/W2936774411","https://openalex.org/W3094502228","https://openalex.org/W3094550259","https://openalex.org/W3189075232","https://openalex.org/W3196974791","https://openalex.org/W3205743929","https://openalex.org/W4287672208","https://openalex.org/W4385245566","https://openalex.org/W6631190155","https://openalex.org/W6638523607","https://openalex.org/W6739901393","https://openalex.org/W6745136726","https://openalex.org/W6754940743","https://openalex.org/W6784333009"],"related_works":["https://openalex.org/W2580650124","https://openalex.org/W4386190339","https://openalex.org/W2968424575","https://openalex.org/W3142333283","https://openalex.org/W3122088529","https://openalex.org/W3041320102","https://openalex.org/W2111669074","https://openalex.org/W2085259108","https://openalex.org/W3123087812","https://openalex.org/W2063076820"],"abstract_inverted_index":{"Machine":[0],"learning":[1],"from":[2],"training":[3],"data":[4,101],"with":[5],"a":[6,33,53,172],"skewed":[7],"distribution":[8],"of":[9,26,37,120,158],"examples":[10],"per":[11],"class":[12,146],"can":[13],"lead":[14],"to":[15,69,83,149,168],"models":[16],"that":[17,60,91,124,142],"favor":[18],"performance":[19,27,46,61,71,85,95,105,147],"on":[20,28,47,62,72,86,96,106,131,171],"common":[21,73,150],"classes":[22,64],"at":[23],"the":[24,70,97,113,118,132],"expense":[25],"rare":[29,63,145],"ones.":[30,74],"AudioSet":[31,48,99],"has":[32],"very":[34],"wide":[35],"range":[36],"priors":[38],"over":[39,56],"its":[40,125],"527":[41],"sound":[42],"event":[43],"classes.":[44,151],"Classification":[45],"is":[49,65],"usually":[50],"evaluated":[51],"by":[52],"simple":[54],"average":[55],"per-class":[57],"metrics,":[58],"meaning":[59],"equal":[66],"in":[67],"importance":[68],"Several":[75],"recent":[76],"papers":[77],"have":[78],"used":[79],"dataset":[80],"balancing":[81,93,143],"techniques":[82],"improve":[84],"AudioSet.":[87],"We":[88,135,152],"find,":[89],"however,":[90],"while":[92],"improves":[94,144],"public":[98,173],"evaluation":[100,109,133,174],"it":[102],"simultaneously":[103],"hurts":[104],"an":[107],"unpublished":[108],"set":[110],"collected":[111],"under":[112],"same":[114],"conditions.":[115],"By":[116],"varying":[117],"degree":[119],"balancing,":[121,159],"we":[122],"show":[123],"benefits":[126],"are":[127],"fragile":[128],"and":[129],"depend":[130],"set.":[134,175],"also":[136],"do":[137],"not":[138],"find":[139],"evidence":[140],"indicating":[141],"relative":[148],"therefore":[153],"caution":[154],"against":[155,163],"blind":[156],"application":[157],"as":[160,162],"well":[161],"paying":[164],"too":[165],"much":[166],"attention":[167],"small":[169],"improvements":[170]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":4}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}