{"id":"https://openalex.org/W4392904216","doi":"https://doi.org/10.1109/icassp48485.2024.10446321","title":"Residualtransformer: Residual Low-Rank Learning With Weight-Sharing For Transformer Layers","display_name":"Residualtransformer: Residual Low-Rank Learning With Weight-Sharing For Transformer Layers","publication_year":2024,"publication_date":"2024-03-18","ids":{"openalex":"https://openalex.org/W4392904216","doi":"https://doi.org/10.1109/icassp48485.2024.10446321"},"language":"en","primary_location":{"id":"doi:10.1109/icassp48485.2024.10446321","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp48485.2024.10446321","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2024 - 2024 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/A5100377920","display_name":"Yiming Wang","orcid":"https://orcid.org/0000-0001-7330-2235"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Yiming Wang","raw_affiliation_strings":["Microsoft Corporation,Redmond,WA,USA","Microsoft Corporation, Redmond, WA, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Corporation,Redmond,WA,USA","institution_ids":["https://openalex.org/I1290206253"]},{"raw_affiliation_string":"Microsoft Corporation, Redmond, WA, USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100365053","display_name":"Jinyu Li","orcid":"https://orcid.org/0000-0002-1089-9748"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jinyu Li","raw_affiliation_strings":["Microsoft Corporation,Redmond,WA,USA","Microsoft Corporation, Redmond, WA, USA"],"affiliations":[{"raw_affiliation_string":"Microsoft Corporation,Redmond,WA,USA","institution_ids":["https://openalex.org/I1290206253"]},{"raw_affiliation_string":"Microsoft Corporation, Redmond, WA, USA","institution_ids":["https://openalex.org/I1290206253"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5100377920"],"corresponding_institution_ids":["https://openalex.org/I1290206253"],"apc_list":null,"apc_paid":null,"fwci":1.7454,"has_fulltext":false,"cited_by_count":5,"citation_normalized_percentile":{"value":0.85960707,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"11161","last_page":"11165"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9998000264167786,"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":0.9998000264167786,"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":0.9965000152587891,"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":0.9962000250816345,"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/transformer","display_name":"Transformer","score":0.7571171522140503},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6556700468063354},{"id":"https://openalex.org/keywords/encoder","display_name":"Encoder","score":0.6178830862045288},{"id":"https://openalex.org/keywords/residual","display_name":"Residual","score":0.4991016387939453},{"id":"https://openalex.org/keywords/diagonal","display_name":"Diagonal","score":0.49436020851135254},{"id":"https://openalex.org/keywords/rank","display_name":"Rank (graph theory)","score":0.46726030111312866},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.35929742455482483},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.324106901884079},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.18057101964950562},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.10361611843109131},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.08206090331077576},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.07713550329208374}],"concepts":[{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.7571171522140503},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6556700468063354},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.6178830862045288},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.4991016387939453},{"id":"https://openalex.org/C130367717","wikidata":"https://www.wikidata.org/wiki/Q189791","display_name":"Diagonal","level":2,"score":0.49436020851135254},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.46726030111312866},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.35929742455482483},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.324106901884079},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.18057101964950562},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.10361611843109131},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.08206090331077576},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.07713550329208374},{"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/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"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/icassp48485.2024.10446321","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp48485.2024.10446321","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2024 - 2024 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":31,"referenced_works":["https://openalex.org/W2056738732","https://openalex.org/W2144499799","https://openalex.org/W2194775991","https://openalex.org/W2695252763","https://openalex.org/W2889129739","https://openalex.org/W2892009249","https://openalex.org/W2963250244","https://openalex.org/W2964110616","https://openalex.org/W3016010032","https://openalex.org/W3097777922","https://openalex.org/W3161873870","https://openalex.org/W3163237592","https://openalex.org/W3196509775","https://openalex.org/W3197726142","https://openalex.org/W4221138270","https://openalex.org/W4297639955","https://openalex.org/W4323066695","https://openalex.org/W4372259858","https://openalex.org/W4385570909","https://openalex.org/W6638523607","https://openalex.org/W6731672793","https://openalex.org/W6745148473","https://openalex.org/W6753640285","https://openalex.org/W6755207826","https://openalex.org/W6768021236","https://openalex.org/W6778883912","https://openalex.org/W6796581206","https://openalex.org/W6801862886","https://openalex.org/W6843363707","https://openalex.org/W6847363464","https://openalex.org/W6850218400"],"related_works":["https://openalex.org/W2135584473","https://openalex.org/W2560215812","https://openalex.org/W2949601986","https://openalex.org/W2964223869","https://openalex.org/W2792539140","https://openalex.org/W4390516098","https://openalex.org/W2788972299","https://openalex.org/W2498789492","https://openalex.org/W2521347458","https://openalex.org/W2169356733"],"abstract_inverted_index":{"Memory":[0],"constraint":[1],"of":[2,7,27,122,137,142],"always-on":[3],"devices":[4],"is":[5,39],"one":[6],"the":[8,36,75,138,153],"major":[9],"concerns":[10],"when":[11],"deploying":[12],"speech":[13,145,148],"processing":[14],"models":[15,21],"on":[16],"these":[17],"devices.":[18],"While":[19],"larger":[20],"trained":[22],"with":[23,101,162],"sufficiently":[24],"large":[25],"amount":[26,121],"data":[28],"generally":[29],"perform":[30],"better,":[31],"making":[32],"them":[33],"fit":[34],"in":[35,91],"device":[37],"memory":[38],"a":[40,62,92,97,107,119],"demanding":[41],"challenge.":[42],"In":[43,126],"this":[44],"paper,":[45],"we":[46,81,128],"aim":[47],"to":[48,111,133],"reduce":[49],"model":[50,54,123],"size":[51,124,156],"by":[52,71,160],"reparameterizing":[53],"weights":[55],"across":[56],"Transformer":[57,93,154],"encoder":[58,155],"layers":[59],"and":[60,66,74,105,147],"assuming":[61],"special":[63],"weight":[64,89,131],"composition":[65],"structure.":[67],"More":[68],"specifically,":[69],"inspired":[70],"ResNet":[72],"[1]":[73],"more":[76],"recent":[77],"LoRA":[78],"[2]":[79],"work,":[80],"propose":[82],"an":[83],"approach":[84],"named":[85],"ResidualTransformer,":[86],"where":[87],"each":[88],"matrix":[90],"layer":[94],"comprises":[95],"1)":[96],"shared":[98],"full-rank":[99],"component":[100,110],"its":[102],"adjacent":[103],"layers,":[104],"2)":[106],"unique":[108],"low-rank":[109,114,139],"itself.":[112],"The":[113],"matrices":[115,132],"only":[116],"account":[117],"for":[118],"small":[120],"increase.":[125],"addition,":[127],"add":[129],"diagonal":[130],"improve":[134],"modeling":[135],"capacity":[136],"matrices.":[140],"Experiments":[141],"our":[143],"10k-hour":[144],"recognition":[146],"translation":[149],"tasks":[150],"show":[151],"that":[152],"can":[157],"be":[158],"reduced":[159],"\u223c3\u00d7":[161],"very":[163],"slight":[164],"performance":[165],"degradation.":[166]},"counts_by_year":[{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":1}],"updated_date":"2026-03-27T14:29:43.386196","created_date":"2025-10-10T00:00:00"}