{"id":"https://openalex.org/W3178511319","doi":"https://doi.org/10.21437/interspeech.2021-712","title":"Amortized Neural Networks for Low-Latency Speech Recognition","display_name":"Amortized Neural Networks for Low-Latency Speech Recognition","publication_year":2021,"publication_date":"2021-08-27","ids":{"openalex":"https://openalex.org/W3178511319","doi":"https://doi.org/10.21437/interspeech.2021-712","mag":"3178511319"},"language":"en","primary_location":{"id":"doi:10.21437/interspeech.2021-712","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2021-712","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Interspeech 2021","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2108.01553","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5082890917","display_name":"Jonathan J. Macoskey","orcid":"https://orcid.org/0000-0002-2776-230X"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jonathan Macoskey","raw_affiliation_strings":["Amazon (United States), Seattle, United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Amazon (United States), Seattle, United States","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023708721","display_name":"Grant P. Strimel","orcid":null},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Grant P. Strimel","raw_affiliation_strings":["Amazon (United States), Seattle, United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Amazon (United States), Seattle, United States","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017559398","display_name":"Jinru Su","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinru Su","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5110230355","display_name":"Ariya Rastrow","orcid":null},"institutions":[{"id":"https://openalex.org/I4210089985","display_name":"Amazon (Germany)","ror":"https://ror.org/00b9ktm87","country_code":"DE","type":"company","lineage":["https://openalex.org/I1311688040","https://openalex.org/I4210089985"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Ariya Rastrow","raw_affiliation_strings":["Amazon"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Amazon","institution_ids":["https://openalex.org/I4210089985"]}]}],"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":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9997000098228455,"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":0.9997000098228455,"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/T10320","display_name":"Neural Networks and Applications","score":0.9980999827384949,"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":0.9937000274658203,"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":0.7976561188697815},{"id":"https://openalex.org/keywords/recurrent-neural-network","display_name":"Recurrent neural network","score":0.6779060363769531},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.6411347389221191},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5884575247764587},{"id":"https://openalex.org/keywords/frame","display_name":"Frame (networking)","score":0.5168948173522949},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5044807195663452},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.5026402473449707},{"id":"https://openalex.org/keywords/network-architecture","display_name":"Network architecture","score":0.45086508989334106},{"id":"https://openalex.org/keywords/low-latency","display_name":"Low latency (capital markets)","score":0.4239698648452759},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.3381307125091553},{"id":"https://openalex.org/keywords/real-time-computing","display_name":"Real-time computing","score":0.3238251805305481},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2533232569694519},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.1632135808467865}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7976561188697815},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.6779060363769531},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.6411347389221191},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5884575247764587},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.5168948173522949},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5044807195663452},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.5026402473449707},{"id":"https://openalex.org/C193415008","wikidata":"https://www.wikidata.org/wiki/Q639681","display_name":"Network architecture","level":2,"score":0.45086508989334106},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.4239698648452759},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3381307125091553},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.3238251805305481},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2533232569694519},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.1632135808467865},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.0},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.21437/interspeech.2021-712","is_oa":false,"landing_page_url":"https://doi.org/10.21437/interspeech.2021-712","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Interspeech 2021","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2108.01553","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2108.01553","pdf_url":"https://arxiv.org/pdf/2108.01553","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":"","raw_type":"text"},{"id":"mag:3178511319","is_oa":true,"landing_page_url":"https://arxiv.org/pdf/2108.01553.pdf","pdf_url":null,"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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null},{"id":"doi:10.48550/arxiv.2108.01553","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2108.01553","pdf_url":null,"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":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2108.01553","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2108.01553","pdf_url":"https://arxiv.org/pdf/2108.01553","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":"","raw_type":"text"},"sustainable_development_goals":[{"score":0.6000000238418579,"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3178511319.pdf","grobid_xml":"https://content.openalex.org/works/W3178511319.grobid-xml"},"referenced_works_count":43,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W1523985187","https://openalex.org/W1689711448","https://openalex.org/W1987570579","https://openalex.org/W2005708641","https://openalex.org/W2058283171","https://openalex.org/W2061708033","https://openalex.org/W2064675550","https://openalex.org/W2143612262","https://openalex.org/W2158114082","https://openalex.org/W2160547390","https://openalex.org/W2179423374","https://openalex.org/W2312434537","https://openalex.org/W2325237720","https://openalex.org/W2396144419","https://openalex.org/W2497335382","https://openalex.org/W2547875792","https://openalex.org/W2556046966","https://openalex.org/W2754132561","https://openalex.org/W2764043458","https://openalex.org/W2886352393","https://openalex.org/W2915589364","https://openalex.org/W2952533036","https://openalex.org/W2962760690","https://openalex.org/W2962956675","https://openalex.org/W2963174729","https://openalex.org/W2963414781","https://openalex.org/W2964062240","https://openalex.org/W2964121744","https://openalex.org/W2971840980","https://openalex.org/W2975044525","https://openalex.org/W2976556660","https://openalex.org/W3006631416","https://openalex.org/W3014981215","https://openalex.org/W3015194534","https://openalex.org/W3037982576","https://openalex.org/W3094667432","https://openalex.org/W3095311338","https://openalex.org/W3096552437","https://openalex.org/W3097522836","https://openalex.org/W3151287998","https://openalex.org/W3162309234","https://openalex.org/W3162649911"],"related_works":["https://openalex.org/W3196783077","https://openalex.org/W2918006316","https://openalex.org/W3147053158","https://openalex.org/W2795328587","https://openalex.org/W2990266649","https://openalex.org/W2968095833","https://openalex.org/W3180266648","https://openalex.org/W3136839710","https://openalex.org/W2921965200","https://openalex.org/W3206624279","https://openalex.org/W3008827736","https://openalex.org/W2765537880","https://openalex.org/W3140099337","https://openalex.org/W2096426948","https://openalex.org/W3042267743","https://openalex.org/W3211659500","https://openalex.org/W3213528868","https://openalex.org/W2893550440","https://openalex.org/W3159727696","https://openalex.org/W3013407415"],"abstract_inverted_index":{"We":[0,19,104],"introduce":[1],"Amortized":[2],"Neural":[3,25],"Networks":[4],"(AmNets),":[5],"a":[6,56,134],"compute":[7,31,66,75,102],"cost-":[8],"and":[9,33,68,85,113,127],"latency-aware":[10],"network":[11,48,98],"architecture":[12,45,118],"particularly":[13],"well-suited":[14],"for":[15,35,76],"sequence":[16],"modeling":[17],"tasks.":[18],"apply":[20],"AmNets":[21,43],"to":[22,29,49,125,129],"the":[23,47,86],"Recurrent":[24],"Network":[26],"Transducer":[27],"(RNN-T)":[28],"reduce":[30,120],"cost":[32,67,122],"latency":[34,128],"an":[36,96],"automatic":[37],"speech":[38],"recognition":[39],"(ASR)":[40],"task.":[41],"The":[42],"RNN-T":[44],"enables":[46],"dynamically":[50],"switch":[51],"between":[52],"encoder":[53],"branches":[54],"on":[55,110],"frame-by-frame":[57],"basis.":[58],"Branches":[59],"are":[60],"constructed":[61],"with":[62],"variable":[63,74],"levels":[64],"of":[65],"model":[69],"capacity.":[70],"Here,":[71],"we":[72],"achieve":[73],"two":[77],"well-known":[78],"candidate":[79],"techniques:":[80],"one":[81],"using":[82,88,107],"sparse":[83],"pruning":[84],"other":[87],"matrix":[89],"factorization.":[90],"Frame-by-frame":[91],"switching":[92],"is":[93],"determined":[94],"by":[95,123],"arbitrator":[97],"that":[99,115],"requires":[100],"negligible":[101],"overhead.":[103],"present":[105],"results":[106],"both":[108],"architectures":[109],"LibriSpeech":[111],"data":[112],"show":[114],"our":[116],"proposed":[117],"can":[119],"inference":[121],"up":[124],"45\\%":[126],"nearly":[130],"real-time":[131],"without":[132],"incurring":[133],"loss":[135],"in":[136],"accuracy.":[137]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}