{"id":"https://openalex.org/W4372259858","doi":"https://doi.org/10.1109/icassp49357.2023.10095006","title":"Sharing Low Rank Conformer Weights for Tiny Always-On Ambient Speech Recognition Models","display_name":"Sharing Low Rank Conformer Weights for Tiny Always-On Ambient Speech Recognition Models","publication_year":2023,"publication_date":"2023-05-05","ids":{"openalex":"https://openalex.org/W4372259858","doi":"https://doi.org/10.1109/icassp49357.2023.10095006"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49357.2023.10095006","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10095006","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":["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/A5000887058","display_name":"Steven M. Hernandez","orcid":"https://orcid.org/0000-0001-6386-5704"},"institutions":[{"id":"https://openalex.org/I184840846","display_name":"Virginia Commonwealth University","ror":"https://ror.org/02nkdxk79","country_code":"US","type":"education","lineage":["https://openalex.org/I184840846"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Steven M. Hernandez","raw_affiliation_strings":["Virginia Commonwealth University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Virginia Commonwealth University","institution_ids":["https://openalex.org/I184840846"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037644321","display_name":"Ding Zhao","orcid":"https://orcid.org/0000-0002-9400-8446"},"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":"Ding Zhao","raw_affiliation_strings":["Google"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058886181","display_name":"Shaojin Ding","orcid":"https://orcid.org/0000-0002-2108-3111"},"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":"Shaojin Ding","raw_affiliation_strings":["Google"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5091201956","display_name":"Antoine Bruguier","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":"Antoine Bruguier","raw_affiliation_strings":["Google"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032640894","display_name":"Rohit Prabhavalkar","orcid":"https://orcid.org/0000-0001-5331-6058"},"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":"Rohit Prabhavalkar","raw_affiliation_strings":["Google"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070513394","display_name":"Tara N. Sainath","orcid":"https://orcid.org/0000-0002-4126-6556"},"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":"Tara N. Sainath","raw_affiliation_strings":["Google"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101319167","display_name":"Yanzhang He","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":"Yanzhang He","raw_affiliation_strings":["Google"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5048236120","display_name":"Ian McGraw","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":"Ian McGraw","raw_affiliation_strings":["Google"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google","institution_ids":["https://openalex.org/I1291425158"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":9,"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":0.9998999834060669,"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":0.9998999834060669,"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/T11309","display_name":"Music and Audio Processing","score":0.9991000294685364,"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":0.9983000159263611,"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":0.7241244316101074},{"id":"https://openalex.org/keywords/rank","display_name":"Rank (graph theory)","score":0.5905447006225586},{"id":"https://openalex.org/keywords/wearable-computer","display_name":"Wearable computer","score":0.5245276093482971},{"id":"https://openalex.org/keywords/component","display_name":"Component (thermodynamics)","score":0.516319215297699},{"id":"https://openalex.org/keywords/reuse","display_name":"Reuse","score":0.515696108341217},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.47743162512779236},{"id":"https://openalex.org/keywords/conformational-isomerism","display_name":"Conformational isomerism","score":0.46055471897125244},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.33090752363204956},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.17166107892990112},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.13558632135391235}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7241244316101074},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.5905447006225586},{"id":"https://openalex.org/C150594956","wikidata":"https://www.wikidata.org/wiki/Q1334829","display_name":"Wearable computer","level":2,"score":0.5245276093482971},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.516319215297699},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.515696108341217},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.47743162512779236},{"id":"https://openalex.org/C18705241","wikidata":"https://www.wikidata.org/wiki/Q1128023","display_name":"Conformational isomerism","level":3,"score":0.46055471897125244},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.33090752363204956},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.17166107892990112},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.13558632135391235},{"id":"https://openalex.org/C32909587","wikidata":"https://www.wikidata.org/wiki/Q11369","display_name":"Molecule","level":2,"score":0.0},{"id":"https://openalex.org/C97355855","wikidata":"https://www.wikidata.org/wiki/Q11473","display_name":"Thermodynamics","level":1,"score":0.0},{"id":"https://openalex.org/C178790620","wikidata":"https://www.wikidata.org/wiki/Q11351","display_name":"Organic chemistry","level":1,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C18903297","wikidata":"https://www.wikidata.org/wiki/Q7150","display_name":"Ecology","level":1,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"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/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49357.2023.10095006","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10095006","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"}],"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/W2294543795","https://openalex.org/W2754084392","https://openalex.org/W2963000224","https://openalex.org/W2963273111","https://openalex.org/W2963928591","https://openalex.org/W2963977978","https://openalex.org/W2972824267","https://openalex.org/W2988271927","https://openalex.org/W2996428491","https://openalex.org/W2996834012","https://openalex.org/W3011347095","https://openalex.org/W3095173472","https://openalex.org/W3097777922","https://openalex.org/W3158216587","https://openalex.org/W3162309234","https://openalex.org/W3197900814","https://openalex.org/W4221138270","https://openalex.org/W4283065721","https://openalex.org/W4283809320","https://openalex.org/W4295262505","https://openalex.org/W4297032622","https://openalex.org/W4317796385","https://openalex.org/W4385245566","https://openalex.org/W6693397755","https://openalex.org/W6725543821","https://openalex.org/W6739901393","https://openalex.org/W6768021236","https://openalex.org/W6774875199","https://openalex.org/W6838147948","https://openalex.org/W6839186069","https://openalex.org/W6840030994","https://openalex.org/W6842996310"],"related_works":["https://openalex.org/W2023036309","https://openalex.org/W1972863456","https://openalex.org/W2089371831","https://openalex.org/W4385077663","https://openalex.org/W2052399476","https://openalex.org/W2077273959","https://openalex.org/W2810163160","https://openalex.org/W4220905048","https://openalex.org/W2037964119","https://openalex.org/W2365701604"],"abstract_inverted_index":{"Continued":[0],"improvements":[1],"in":[2],"machine":[3],"learning":[4],"techniques":[5],"offer":[6,27],"exciting":[7],"new":[8,29],"opportunities":[9],"through":[10],"the":[11,55,153,159,166],"use":[12],"of":[13,58,147,161,171,191],"larger":[14,17],"models":[15,62,66],"and":[16,38,132,174,182,193,198],"training":[18],"datasets.":[19],"However,":[20],"there":[21],"is":[22,46],"a":[23,84,169,183,189,202],"growing":[24],"need":[25],"to":[26,53,73,88,165],"these":[28],"capabilities":[30],"on-board":[31],"low-powered":[32],"devices":[33,96],"such":[34],"as":[35],"smart-phones,":[36],"wearables":[37],"other":[39],"embedded":[40],"environments":[41],"where":[42],"only":[43],"low":[44],"memory":[45],"available.":[47],"Towards":[48],"this,":[49],"we":[50,150,176,186],"consider":[51],"methods":[52],"reduce":[54],"model":[56,81,85,103,111,155],"size":[57],"Conformer-based":[59],"speech":[60,92],"recognition":[61,93],"which":[63],"typically":[64],"require":[65],"with":[67,97,179,201],"greater":[68],"than":[69],"100M":[70],"parameters":[71,76],"down":[72],"just":[74],"5M":[75,203],"while":[77,157],"minimizing":[78],"impact":[79],"on":[80,94],"quality.":[82],"Such":[83],"allows":[86],"us":[87],"achieve":[89,188],"always-on":[90],"ambient":[91],"edge":[95],"low-memory":[98],"neural":[99],"processors.":[100],"We":[101],"propose":[102],"weight":[104,180],"reuse":[105],"at":[106,144],"different":[107,145],"levels":[108,146],"within":[109],"our":[110,148],"architecture:":[112],"(i)":[113],"repeating":[114],"full":[115,154],"conformer":[116,122,130],"block":[117],"layers,":[118,125],"(ii)":[119],"sharing":[120,127,134,142,181],"specific":[121],"modules":[123],"across":[124],"(iii)":[126],"sub-components":[128],"per":[129],"module,":[131],"(iv)":[133],"decomposed":[135],"sub-component":[136],"weights":[137,143],"after":[138],"low-rank":[139,184],"decomposition.":[140],"By":[141],"model,":[149],"can":[151,187],"retain":[152],"in-memory":[156],"increasing":[158],"number":[160],"virtual":[162],"trans-formations":[163],"applied":[164],"input.":[167],"Through":[168],"series":[170],"ablation":[172],"studies":[173],"evaluations,":[175],"find":[177],"that":[178],"architecture,":[185],"WER":[190],"2.84":[192],"2.94":[194],"for":[195],"Librispeech":[196],"dev-clean":[197],"test-clean":[199],"respectively":[200],"parameter":[204],"model.":[205]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":3},{"year":2024,"cited_by_count":5}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}