{"id":"https://openalex.org/W4372349720","doi":"https://doi.org/10.1109/icassp49357.2023.10096570","title":"Training Large-Vocabulary Neural Language Models by Private Federated Learning for Resource-Constrained Devices","display_name":"Training Large-Vocabulary Neural Language Models by Private Federated Learning for Resource-Constrained Devices","publication_year":2023,"publication_date":"2023-05-05","ids":{"openalex":"https://openalex.org/W4372349720","doi":"https://doi.org/10.1109/icassp49357.2023.10096570"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49357.2023.10096570","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10096570","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":"article","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/A5086280874","display_name":"Mingbin Xu","orcid":null},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Mingbin Xu","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048446895","display_name":"Congzheng Song","orcid":"https://orcid.org/0000-0002-5076-7299"},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Congzheng Song","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024018042","display_name":"Ye Tian","orcid":"https://orcid.org/0000-0002-3428-1889"},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Ye Tian","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101667681","display_name":"Neha Agrawal","orcid":"https://orcid.org/0000-0002-6904-2060"},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Neha Agrawal","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005144024","display_name":"Filip Granqvist","orcid":null},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Filip Granqvist","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5045282611","display_name":"Rogier van Dalen","orcid":null},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Rogier van Dalen","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100320831","display_name":"Xiao Zhang","orcid":"https://orcid.org/0000-0001-5050-2621"},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Xiao Zhang","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013339281","display_name":"Arturo Argueta","orcid":null},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Arturo Argueta","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054574376","display_name":"Shiyi Han","orcid":null},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Shiyi Han","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050980161","display_name":"Yaqiao Deng","orcid":null},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Yaqiao Deng","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101842894","display_name":"Leo Liu","orcid":"https://orcid.org/0009-0008-1720-8533"},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Leo Liu","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065974400","display_name":"Anmol Walia","orcid":null},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Anmol Walia","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5085670252","display_name":"Alex Jin","orcid":null},"institutions":[{"id":"https://openalex.org/I4210107260","display_name":"Apple (United Kingdom)","ror":"https://ror.org/01vpeym60","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210107260"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Alex Jin","raw_affiliation_strings":["Apple"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Apple","institution_ids":["https://openalex.org/I4210107260"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I4210107260"],"apc_list":null,"apc_paid":null,"fwci":2.2583,"has_fulltext":false,"cited_by_count":14,"citation_normalized_percentile":{"value":0.90217891,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":98},"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/T10764","display_name":"Privacy-Preserving Technologies in Data","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/T10764","display_name":"Privacy-Preserving Technologies in Data","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/T11704","display_name":"Mobile Crowdsensing and Crowdsourcing","score":0.9506000280380249,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9492999911308289,"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":0.8652421236038208},{"id":"https://openalex.org/keywords/differential-privacy","display_name":"Differential privacy","score":0.8558623194694519},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.6982265710830688},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.5479061007499695},{"id":"https://openalex.org/keywords/edge-device","display_name":"Edge device","score":0.5291260480880737},{"id":"https://openalex.org/keywords/payload","display_name":"Payload (computing)","score":0.5289782285690308},{"id":"https://openalex.org/keywords/vocabulary","display_name":"Vocabulary","score":0.5032040476799011},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.49724701046943665},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.48809874057769775},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4408625364303589},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.42254117131233215},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.38213446736335754},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.33440878987312317},{"id":"https://openalex.org/keywords/data-mining","display_name":"Data mining","score":0.29113101959228516},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.22355660796165466},{"id":"https://openalex.org/keywords/cloud-computing","display_name":"Cloud computing","score":0.1769120991230011}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8652421236038208},{"id":"https://openalex.org/C23130292","wikidata":"https://www.wikidata.org/wiki/Q5275358","display_name":"Differential privacy","level":2,"score":0.8558623194694519},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6982265710830688},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.5479061007499695},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.5291260480880737},{"id":"https://openalex.org/C134066672","wikidata":"https://www.wikidata.org/wiki/Q1424639","display_name":"Payload (computing)","level":3,"score":0.5289782285690308},{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.5032040476799011},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.49724701046943665},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.48809874057769775},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4408625364303589},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.42254117131233215},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.38213446736335754},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.33440878987312317},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.29113101959228516},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.22355660796165466},{"id":"https://openalex.org/C79974875","wikidata":"https://www.wikidata.org/wiki/Q483639","display_name":"Cloud computing","level":2,"score":0.1769120991230011},{"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/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C120665830","wikidata":"https://www.wikidata.org/wiki/Q14620","display_name":"Optics","level":1,"score":0.0},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"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/C158379750","wikidata":"https://www.wikidata.org/wiki/Q214111","display_name":"Network packet","level":2,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49357.2023.10096570","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49357.2023.10096570","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":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education","score":0.44999998807907104}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":44,"referenced_works":["https://openalex.org/W1873763122","https://openalex.org/W1992926795","https://openalex.org/W2046932483","https://openalex.org/W2064675550","https://openalex.org/W2120861206","https://openalex.org/W2152790380","https://openalex.org/W2251640092","https://openalex.org/W2405883473","https://openalex.org/W2473418344","https://openalex.org/W2784621220","https://openalex.org/W2896457183","https://openalex.org/W2900120080","https://openalex.org/W2946930197","https://openalex.org/W2963347649","https://openalex.org/W2963456518","https://openalex.org/W2970408908","https://openalex.org/W2972570881","https://openalex.org/W3015480556","https://openalex.org/W3168867926","https://openalex.org/W3188505388","https://openalex.org/W3196902142","https://openalex.org/W3206066344","https://openalex.org/W3207429447","https://openalex.org/W4287553002","https://openalex.org/W4287663285","https://openalex.org/W4288333953","https://openalex.org/W4299283926","https://openalex.org/W4318619660","https://openalex.org/W4385245566","https://openalex.org/W6639246211","https://openalex.org/W6678040779","https://openalex.org/W6682948231","https://openalex.org/W6728757088","https://openalex.org/W6739901393","https://openalex.org/W6747855403","https://openalex.org/W6755207826","https://openalex.org/W6763393573","https://openalex.org/W6764838729","https://openalex.org/W6767676916","https://openalex.org/W6773976177","https://openalex.org/W6787335730","https://openalex.org/W6796581206","https://openalex.org/W6802279528","https://openalex.org/W6802709103"],"related_works":["https://openalex.org/W3038283795","https://openalex.org/W2973622361","https://openalex.org/W3176282186","https://openalex.org/W4387489555","https://openalex.org/W3185576471","https://openalex.org/W4288024917","https://openalex.org/W4293053895","https://openalex.org/W2983364019","https://openalex.org/W2998183476","https://openalex.org/W3215372595"],"abstract_inverted_index":{"Federated":[0],"Learning":[1],"(FL)":[2],"is":[3,38],"a":[4,27,41,82],"technique":[5,84],"to":[6,25,39,62,85,107,134],"train":[7,40,135],"models":[8,114,138],"on":[9,34,48,115],"distributed":[10],"edge":[11],"devices":[12,50,125],"with":[13,23,123],"local":[14],"data":[15,33],"samples.":[16],"Differential":[17],"Privacy":[18],"(DP)":[19],"can":[20],"be":[21],"applied":[22],"FL":[24,55],"provide":[26],"formal":[28],"privacy":[29,53,65],"guarantee":[30,63],"for":[31],"sensitive":[32],"device.":[35],"Our":[36],"goal":[37],"large":[42,113],"neural":[43],"network":[44],"language":[45,137],"model":[46,69],"(NNLM)":[47],"compute-constrained":[49,116],"while":[51,139],"preserving":[52,140],"using":[54],"and":[56,102,122,142],"DP.":[57],"However,":[58],"the":[59,68,87,109],"noise":[60],"required":[61],"differential":[64],"increases":[66],"as":[67],"size":[70],"grows,":[71],"which":[72],"often":[73],"prevents":[74],"convergence.":[75],"We":[76,118],"propose":[77],"Partial":[78],"Embedding":[79],"Updates":[80],"(PEU),":[81],"novel":[83],"reduce":[86,108],"impact":[88],"of":[89,112,129],"DP-noise":[90],"by":[91],"decreasing":[92],"payload":[93],"size.":[94],"Furthermore,":[95],"we":[96],"adopt":[97],"Low":[98],"Rank":[99],"Adaptation":[100],"(LoRA)":[101],"Noise":[103],"Contrastive":[104],"Estimation":[105],"(NCE)":[106],"memory":[110],"demands":[111],"devices.":[117],"demonstrate":[119],"in":[120],"simulation":[121],"real":[124],"that":[126],"this":[127],"combination":[128],"techniques":[130],"makes":[131],"it":[132],"possible":[133],"large-vocabulary":[136],"accuracy":[141],"privacy.":[143]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":5},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":5}],"updated_date":"2026-06-26T08:34:08.712188","created_date":"2025-10-10T00:00:00"}