{"id":"https://openalex.org/W3189437251","doi":"https://doi.org/10.24963/ijcai.2021/472","title":"Automatic Mixed-Precision Quantization Search of BERT","display_name":"Automatic Mixed-Precision Quantization Search of BERT","publication_year":2021,"publication_date":"2021-08-01","ids":{"openalex":"https://openalex.org/W3189437251","doi":"https://doi.org/10.24963/ijcai.2021/472","mag":"3189437251"},"language":"en","primary_location":{"id":"doi:10.24963/ijcai.2021/472","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2021/472","pdf_url":"https://www.ijcai.org/proceedings/2021/0472.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://www.ijcai.org/proceedings/2021/0472.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5109628225","display_name":"Changsheng Zhao","orcid":null},"institutions":[{"id":"https://openalex.org/I4210101778","display_name":"Samsung (United States)","ror":"https://ror.org/01bfbvm65","country_code":"US","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210101778"]},{"id":"https://openalex.org/I4210133173","display_name":"Research!America (United States)","ror":"https://ror.org/044pgyv50","country_code":"US","type":"company","lineage":["https://openalex.org/I4210133173"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Changsheng Zhao","raw_affiliation_strings":["Samsung Research America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Samsung Research America","institution_ids":["https://openalex.org/I4210101778","https://openalex.org/I4210133173"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112108996","display_name":"Hua Ting","orcid":null},"institutions":[{"id":"https://openalex.org/I4210101778","display_name":"Samsung (United States)","ror":"https://ror.org/01bfbvm65","country_code":"US","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210101778"]},{"id":"https://openalex.org/I4210133173","display_name":"Research!America (United States)","ror":"https://ror.org/044pgyv50","country_code":"US","type":"company","lineage":["https://openalex.org/I4210133173"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ting Hua","raw_affiliation_strings":["Samsung Research America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Samsung Research America","institution_ids":["https://openalex.org/I4210101778","https://openalex.org/I4210133173"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058974318","display_name":"Yilin Shen","orcid":"https://orcid.org/0000-0002-1955-1529"},"institutions":[{"id":"https://openalex.org/I4210101778","display_name":"Samsung (United States)","ror":"https://ror.org/01bfbvm65","country_code":"US","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210101778"]},{"id":"https://openalex.org/I4210133173","display_name":"Research!America (United States)","ror":"https://ror.org/044pgyv50","country_code":"US","type":"company","lineage":["https://openalex.org/I4210133173"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yilin Shen","raw_affiliation_strings":["Samsung Research America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Samsung Research America","institution_ids":["https://openalex.org/I4210101778","https://openalex.org/I4210133173"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006753786","display_name":"Qian Lou","orcid":"https://orcid.org/0000-0001-5462-2567"},"institutions":[{"id":"https://openalex.org/I4210101778","display_name":"Samsung (United States)","ror":"https://ror.org/01bfbvm65","country_code":"US","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210101778"]},{"id":"https://openalex.org/I4210133173","display_name":"Research!America (United States)","ror":"https://ror.org/044pgyv50","country_code":"US","type":"company","lineage":["https://openalex.org/I4210133173"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Qian Lou","raw_affiliation_strings":["Samsung Research America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Samsung Research America","institution_ids":["https://openalex.org/I4210101778","https://openalex.org/I4210133173"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5044205212","display_name":"Hongxia Jin","orcid":"https://orcid.org/0009-0000-0222-4217"},"institutions":[{"id":"https://openalex.org/I4210101778","display_name":"Samsung (United States)","ror":"https://ror.org/01bfbvm65","country_code":"US","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210101778"]},{"id":"https://openalex.org/I4210133173","display_name":"Research!America (United States)","ror":"https://ror.org/044pgyv50","country_code":"US","type":"company","lineage":["https://openalex.org/I4210133173"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hongxia Jin","raw_affiliation_strings":["Samsung Research America"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Samsung Research America","institution_ids":["https://openalex.org/I4210101778","https://openalex.org/I4210133173"]}]}],"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":16,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"3427","last_page":"3433"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","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/T10028","display_name":"Topic Modeling","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/T10181","display_name":"Natural Language Processing Techniques","score":0.9984999895095825,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9973999857902527,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/quantization","display_name":"Quantization (signal processing)","score":0.8052293062210083},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7948065996170044},{"id":"https://openalex.org/keywords/compression-ratio","display_name":"Compression ratio","score":0.6142523288726807},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.6010951399803162},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.46491485834121704},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3973212242126465},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.332447350025177}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.8052293062210083},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7948065996170044},{"id":"https://openalex.org/C25797200","wikidata":"https://www.wikidata.org/wiki/Q828137","display_name":"Compression ratio","level":3,"score":0.6142523288726807},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6010951399803162},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.46491485834121704},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3973212242126465},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.332447350025177},{"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/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0},{"id":"https://openalex.org/C171146098","wikidata":"https://www.wikidata.org/wiki/Q124192","display_name":"Automotive engineering","level":1,"score":0.0},{"id":"https://openalex.org/C511840579","wikidata":"https://www.wikidata.org/wiki/Q12757","display_name":"Internal combustion engine","level":2,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.24963/ijcai.2021/472","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2021/472","pdf_url":"https://www.ijcai.org/proceedings/2021/0472.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2112.14938","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2112.14938","pdf_url":"https://arxiv.org/pdf/2112.14938","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,"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":null,"raw_type":"text"}],"best_oa_location":{"id":"doi:10.24963/ijcai.2021/472","is_oa":true,"landing_page_url":"https://doi.org/10.24963/ijcai.2021/472","pdf_url":"https://www.ijcai.org/proceedings/2021/0472.pdf","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3189437251.pdf","grobid_xml":"https://content.openalex.org/works/W3189437251.grobid-xml"},"referenced_works_count":26,"referenced_works":["https://openalex.org/W1821462560","https://openalex.org/W2242818861","https://openalex.org/W2300242332","https://openalex.org/W2524428287","https://openalex.org/W2544912866","https://openalex.org/W2553303224","https://openalex.org/W2786771851","https://openalex.org/W2805003733","https://openalex.org/W2896457183","https://openalex.org/W2903260438","https://openalex.org/W2951104886","https://openalex.org/W2952088488","https://openalex.org/W2962847160","https://openalex.org/W2963273111","https://openalex.org/W2963828549","https://openalex.org/W2963854351","https://openalex.org/W2964029277","https://openalex.org/W2970120757","https://openalex.org/W2978017171","https://openalex.org/W2998183051","https://openalex.org/W3034309359","https://openalex.org/W3038012435","https://openalex.org/W3105966348","https://openalex.org/W4288256350","https://openalex.org/W4297813615","https://openalex.org/W4300687381"],"related_works":["https://openalex.org/W2051487156","https://openalex.org/W2073681303","https://openalex.org/W2053286651","https://openalex.org/W2181743346","https://openalex.org/W2187401768","https://openalex.org/W2181413294","https://openalex.org/W2989452537","https://openalex.org/W3026162553","https://openalex.org/W2356664282","https://openalex.org/W2052122378"],"abstract_inverted_index":{"Pre-trained":[0],"language":[1,13,86],"models":[2,18,52],"such":[3,188],"as":[4,189],"BERT":[5,109,150],"have":[6],"shown":[7],"remarkable":[8],"effectiveness":[9],"in":[10,47,133],"various":[11],"natural":[12,85],"processing":[14,87],"tasks.":[15],"However,":[16,50],"these":[17],"usually":[19,91],"contain":[20],"millions":[21],"of":[22,145,175],"parameters,":[23],"which":[24],"prevent":[25],"them":[26],"from":[27,59],"the":[28,44,71,139,162,173,177],"practical":[29],"deployment":[30],"on":[31,81,95,149],"resource-constrained":[32],"devices.":[33],"Knowledge":[34],"distillation,":[35],"Weight":[36],"pruning,":[37],"and":[38,89,114,129,137],"Quantization":[39],"are":[40,75],"known":[41],"to":[42,126],"be":[43],"main":[45],"directions":[46],"model":[48,168,180],"compression.":[49],"compact":[51],"obtained":[53],"through":[54],"knowledge":[55],"distillation":[56],"may":[57],"suffer":[58],"significant":[60],"accuracy":[61],"drop":[62],"even":[63],"for":[64,84,108,131],"a":[65,77],"relatively":[66],"small":[67],"compression":[68],"ratio.":[69],"On":[70],"other":[72],"hand,":[73],"there":[74],"only":[76],"few":[78],"attempts":[79],"based":[80],"quantization":[82,105,113],"designed":[83,107],"tasks,":[88],"they":[90],"require":[92],"manual":[93],"setting":[94],"hyper-parameters.":[96],"In":[97],"this":[98],"paper,":[99],"we":[100],"proposed":[101,119,156],"an":[102],"automatic":[103],"mixed-precision":[104],"framework":[106],"that":[110,154],"can":[111],"conduct":[112],"pruning":[115,141],"simultaneously.":[116],"Specifically,":[117],"our":[118,155,183],"method":[120,157],"leverages":[121],"Differentiable":[122],"Neural":[123],"Architecture":[124],"Search":[125],"assign":[127],"scale":[128],"precision":[130],"parameters":[132],"each":[134],"sub-group":[135],"automatically,":[136],"at":[138],"same":[140,163],"out":[142],"redundant":[143],"groups":[144],"parameters.":[146],"Extensive":[147],"evaluations":[148],"downstream":[151],"tasks":[152],"reveal":[153],"beats":[158],"baselines":[159],"by":[160,181],"providing":[161],"performance":[164],"with":[165,185],"much":[166],"smaller":[167],"size.":[169],"We":[170],"also":[171],"show":[172],"possibility":[174],"obtaining":[176],"extremely":[178],"light-weight":[179],"combining":[182],"solution":[184],"orthogonal":[186],"methods":[187],"DistilBERT.":[190]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":3},{"year":2022,"cited_by_count":3}],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2025-10-10T00:00:00"}