{"id":"https://openalex.org/W4387799863","doi":"https://doi.org/10.48550/arxiv.2310.11230","title":"Zipformer: A faster and better encoder for automatic speech recognition","display_name":"Zipformer: A faster and better encoder for automatic speech recognition","publication_year":2023,"publication_date":"2023-10-17","ids":{"openalex":"https://openalex.org/W4387799863","doi":"https://doi.org/10.48550/arxiv.2310.11230"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2310.11230","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.11230","pdf_url":"https://arxiv.org/pdf/2310.11230","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2310.11230","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5066643565","display_name":"Zengwei Yao","orcid":"https://orcid.org/0000-0002-2331-2387","observed_orcids":["https://orcid.org/0000-0002-2331-2387"]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yao, Zengwei","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102707172","display_name":"Liyong Guo","orcid":"https://orcid.org/0009-0002-2465-3914","observed_orcids":["https://orcid.org/0009-0002-2465-3914"]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guo, Liyong","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100435428","display_name":"Xiaoyu Yang","orcid":"https://orcid.org/0000-0002-3570-8979","observed_orcids":["https://orcid.org/0000-0002-3570-8979"]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Xiaoyu","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068307574","display_name":"Wei Kang","orcid":"https://orcid.org/0000-0003-0193-2654","observed_orcids":["https://orcid.org/0000-0003-0193-2654"]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kang, Wei","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004231067","display_name":"Fangjun Kuang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kuang, Fangjun","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100743978","display_name":"Yifan Yang","orcid":"https://orcid.org/0000-0003-3313-5473","observed_orcids":["https://orcid.org/0000-0003-3313-5473"]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Yifan","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5075023049","display_name":"Zengrui Jin","orcid":"https://orcid.org/0000-0002-2637-7880","observed_orcids":["https://orcid.org/0000-0002-2637-7880"]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jin, Zengrui","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Lin, Long","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lin, Long","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5084286453","display_name":"Daniel Povey","orcid":"https://orcid.org/0000-0002-0611-3634","observed_orcids":["https://orcid.org/0000-0002-0611-3634"]},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Povey, Daniel","raw_affiliation_strings":["Xiaomi Corp. , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaomi Corp. , Beijing , China","institution_ids":["https://openalex.org/I862669128"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I862669128"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":52,"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":1.0,"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":1.0,"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":9.999999747378752e-05,"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/T10028","display_name":"Topic Modeling","score":0.0,"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/conformers","display_name":"conformational isomers","score":0.7649999856948853},{"id":"https://openalex.org/keywords/speech-encoder","display_name":"speech encoder","score":0.5950000286102295},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"speech recognition","score":0.843999981880188},{"id":"https://openalex.org/keywords/memory-efficiency","display_name":"memory efficiency","score":0.7770000100135803},{"id":"https://openalex.org/keywords/librispeech","display_name":"LibriSpeech","score":0.6439999938011169},{"id":"https://openalex.org/keywords/aishell-1","display_name":"AISHELL-1","score":0.9459999799728394}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8009679317474365},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.7662205100059509},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6485761404037476},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5041447877883911},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4494483470916748},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3894374668598175},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.32824552059173584},{"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/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2310.11230","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.11230","pdf_url":"https://arxiv.org/pdf/2310.11230","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2310.11230","doi":"https://doi.org/10.48550/arxiv.2310.11230","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2310.11230","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2310.11230","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.11230","pdf_url":"https://arxiv.org/pdf/2310.11230","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4387799863.pdf","grobid_xml":"https://content.openalex.org/works/W4387799863.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4417003861","https://openalex.org/W7133208400","https://openalex.org/W4378771706","https://openalex.org/W7213533256","https://openalex.org/W4393407971","https://openalex.org/W4297841818","https://openalex.org/W4292215849","https://openalex.org/W4362598042","https://openalex.org/W4412886879","https://openalex.org/W4391021542"],"abstract_inverted_index":{"The":[0],"Conformer":[1],"has":[2],"become":[3],"the":[4,109,118,122,128,149],"most":[5],"popular":[6],"encoder":[7,48],"model":[8],"for":[9,71],"automatic":[10],"speech":[11],"recognition":[12],"(ASR).":[13],"It":[14,131],"adds":[15],"convolution":[16],"modules":[17],"to":[18,21,83,116],"a":[19,33,46,74,102],"transformer":[20],"learn":[22],"both":[23],"local":[24],"and":[25,37,93,124,135,145],"global":[26],"dependencies.":[27],"In":[28],"this":[29],"work":[30,95],"we":[31,67],"describe":[32],"faster,":[34],"more":[35,63],"memory-efficient,":[36],"better-performing":[38],"transformer,":[39],"called":[40,79,105],"Zipformer.":[41],"Modeling":[42],"changes":[43],"include:":[44],"1)":[45],"U-Net-like":[47],"structure":[49,61],"where":[50],"middle":[51],"stacks":[52],"operate":[53],"at":[54,165],"lower":[55],"frame":[56],"rates;":[57],"2)":[58],"reorganized":[59],"block":[60],"with":[62],"modules,":[64],"within":[65],"which":[66,107],"re-use":[68],"attention":[69],"weights":[70],"efficiency;":[72],"3)":[73],"modified":[75],"form":[76],"of":[77,151],"LayerNorm":[78],"BiasNorm":[80],"allows":[81],"us":[82],"retain":[84],"some":[85],"length":[86],"information;":[87],"4)":[88],"new":[89,103],"activation":[90],"functions":[91],"SwooshR":[92],"SwooshL":[94],"better":[96,136],"than":[97,138],"Swish.":[98],"We":[99],"also":[100,125],"propose":[101],"optimizer,":[104],"ScaledAdam,":[106],"scales":[108],"update":[110],"by":[111],"each":[112],"tensor's":[113],"current":[114],"scale":[115],"keep":[117],"relative":[119],"change":[120],"about":[121],"same,":[123],"explictly":[126],"learns":[127],"parameter":[129],"scale.":[130],"achieves":[132],"faster":[133],"convergence":[134],"performance":[137],"Adam.":[139],"Extensive":[140],"experiments":[141],"on":[142],"LibriSpeech,":[143],"Aishell-1,":[144],"WenetSpeech":[146],"datasets":[147],"demonstrate":[148],"effectiveness":[150],"our":[152],"proposed":[153],"Zipformer":[154],"over":[155],"other":[156],"state-of-the-art":[157],"ASR":[158],"models.":[159],"Our":[160],"code":[161],"is":[162],"publicly":[163],"available":[164],"https://github.com/k2-fsa/icefall.":[166]},"counts_by_year":[{"year":2025,"cited_by_count":31},{"year":2024,"cited_by_count":21}],"updated_date":"2026-10-16T09:59:22.235173","created_date":"2023-10-21T00:00:00"}