{"id":"https://openalex.org/W4408354505","doi":"https://doi.org/10.1109/icassp49660.2025.10888920","title":"Effective and Efficient Mixed Precision Quantization of Speech Foundation Models","display_name":"Effective and Efficient Mixed Precision Quantization of Speech Foundation Models","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408354505","doi":"https://doi.org/10.1109/icassp49660.2025.10888920"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10888920","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10888920","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/A5063907750","display_name":"Haoning Xu","orcid":null},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Haoning Xu","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101675941","display_name":"Zhaoqing Li","orcid":"https://orcid.org/0000-0001-8649-4934"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Zhaoqing Li","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5075023049","display_name":"Zengrui Jin","orcid":"https://orcid.org/0000-0002-2637-7880"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Zengrui Jin","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000600118","display_name":"Huimeng Wang","orcid":"https://orcid.org/0000-0002-4280-503X"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Huimeng Wang","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113020902","display_name":"Youjun Chen","orcid":"https://orcid.org/0000-0002-6484-2325"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Youjun Chen","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042515126","display_name":"Guinan Li","orcid":"https://orcid.org/0000-0002-2206-0237"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Guinan Li","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003895235","display_name":"Mengzhe Geng","orcid":"https://orcid.org/0000-0002-7886-439X"},"institutions":[{"id":"https://openalex.org/I4210159778","display_name":"National Research Council Canada","ror":"https://ror.org/04mte1k06","country_code":"CA","type":"government","lineage":["https://openalex.org/I4210159778"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Mengzhe Geng","raw_affiliation_strings":["National Research Council Canada,Canada"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Research Council Canada,Canada","institution_ids":["https://openalex.org/I4210159778"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026334377","display_name":"Shujie Hu","orcid":"https://orcid.org/0000-0002-8475-4912"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Shujie Hu","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5081269555","display_name":"Jiajun Deng","orcid":"https://orcid.org/0000-0002-6834-0322"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Jiajun Deng","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5037109470","display_name":"Xunying Liu","orcid":"https://orcid.org/0000-0001-6725-1160"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xunying Liu","raw_affiliation_strings":["The Chinese University of Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,China","institution_ids":["https://openalex.org/I177725633"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":2,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9697999954223633,"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":0.9697999954223633,"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":0.9394000172615051,"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/T10901","display_name":"Advanced Data Compression Techniques","score":0.9290000200271606,"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/computer-science","display_name":"Computer science","score":0.6567035913467407},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.6276721954345703},{"id":"https://openalex.org/keywords/foundation","display_name":"Foundation (evidence)","score":0.6028002500534058},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5280997157096863},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3376131057739258},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.16132205724716187}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6567035913467407},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.6276721954345703},{"id":"https://openalex.org/C2780966255","wikidata":"https://www.wikidata.org/wiki/Q5474306","display_name":"Foundation (evidence)","level":2,"score":0.6028002500534058},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5280997157096863},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3376131057739258},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.16132205724716187},{"id":"https://openalex.org/C95457728","wikidata":"https://www.wikidata.org/wiki/Q309","display_name":"History","level":0,"score":0.0},{"id":"https://openalex.org/C166957645","wikidata":"https://www.wikidata.org/wiki/Q23498","display_name":"Archaeology","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/icassp49660.2025.10888920","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10888920","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:cisti-icist.nrc-cnrc.ca:cistinparc:56faf904-e182-46a7-b427-7135c3d72b95","is_oa":false,"landing_page_url":"https://nrc-publications.canada.ca/eng/view/object/?id=56faf904-e182-46a7-b427-7135c3d72b95","pdf_url":null,"source":{"id":"https://openalex.org/S7407055245","display_name":"NPARC","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":"acceptedVersion","is_accepted":true,"is_published":false,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":43,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W1932968309","https://openalex.org/W3034309359","https://openalex.org/W3034940165","https://openalex.org/W3162309234","https://openalex.org/W3189437251","https://openalex.org/W3197148831","https://openalex.org/W3203140070","https://openalex.org/W3209059054","https://openalex.org/W3209984917","https://openalex.org/W3217016993","https://openalex.org/W4206375145","https://openalex.org/W4221138270","https://openalex.org/W4224821750","https://openalex.org/W4224918069","https://openalex.org/W4297841520","https://openalex.org/W4297841557","https://openalex.org/W4297841786","https://openalex.org/W4319862415","https://openalex.org/W4319862444","https://openalex.org/W4319862633","https://openalex.org/W4375869065","https://openalex.org/W4385822445","https://openalex.org/W4385822766","https://openalex.org/W4385822878","https://openalex.org/W4385823026","https://openalex.org/W4385823182","https://openalex.org/W4391306611","https://openalex.org/W4392903331","https://openalex.org/W4392903837","https://openalex.org/W4392904500","https://openalex.org/W4392909098","https://openalex.org/W4402111990","https://openalex.org/W6730091202","https://openalex.org/W6752515464","https://openalex.org/W6780218876","https://openalex.org/W6788001715","https://openalex.org/W6793959113","https://openalex.org/W6796551075","https://openalex.org/W6796815506","https://openalex.org/W6838633097","https://openalex.org/W6850634149","https://openalex.org/W6857740148"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2381393187","https://openalex.org/W2332779545","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W2358060160","https://openalex.org/W2035483685"],"abstract_inverted_index":{"This":[0],"paper":[1],"presents":[2],"a":[3,129],"novel":[4],"mixed-precision":[5,15,42,63,114,124],"quantization":[6,73],"approach":[7],"for":[8],"speech":[9],"foundation":[10],"models":[11,38,44,101],"that":[12,66],"tightly":[13],"integrates":[14],"learning":[16,69],"and":[17,36,55,61,70,76,99,108],"quantized":[18,43,64,125],"model":[19,25,71,127],"parameter":[20],"estimation":[21],"into":[22],"one":[23],"single":[24],"compression":[26,48,95,131],"stage.":[27],"Experiments":[28],"conducted":[29],"on":[30],"LibriSpeech":[31],"dataset":[32],"with":[33],"fine-tuned":[34],"wav2vec2.0-base":[35,98],"HuBERT-large":[37,100,126],"suggest":[39],"the":[40,46,58,89,112,136],"resulting":[41],"increased":[45],"lossless":[47,130],"ratio":[49,132],"by":[50,104],"factors":[51],"up":[52,105],"to":[53,106],"1.7x":[54],"1.9x":[56],"over":[57,88,111,135],"respective":[59],"uniform-precision":[60],"two-stage":[62,113],"baselines":[65],"perform":[67],"precision":[68],"parameters":[72],"in":[74],"separate":[75],"disjointed":[77],"stages,":[78],"while":[79,116],"incurring":[80],"no":[81],"statistically":[82],"word":[83],"error":[84],"rate":[85],"(WER)":[86],"increase":[87],"32-bit":[90,137],"full-precision":[91,138],"models.":[92],"The":[93,121],"system":[94],"time":[96],"of":[97,133],"is":[102],"reduced":[103],"1.9":[107],"1.5":[109],"times":[110],"baselines,":[115],"both":[117],"produce":[118],"lower":[119],"WERs.":[120],"best-performing":[122],"3.5-bit":[123],"produces":[128],"8.6x":[134],"system.":[139]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}