{"id":"https://openalex.org/W7137898003","doi":"https://doi.org/10.1609/aaai.v40i26.39345","title":"Double Rounding: Nearly Lossless Adaptive Bit Switching for QAT","display_name":"Double Rounding: Nearly Lossless Adaptive Bit Switching for QAT","publication_year":2026,"publication_date":"2026-03-14","ids":{"openalex":"https://openalex.org/W7137898003","doi":"https://doi.org/10.1609/aaai.v40i26.39345"},"language":null,"primary_location":{"id":"doi:10.1609/aaai.v40i26.39345","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i26.39345","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39345/43306","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39345/43306","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5003944732","display_name":"Haiduo Huang","orcid":"https://orcid.org/0009-0000-6726-9221"},"institutions":[{"id":"https://openalex.org/I87445476","display_name":"Xi'an Jiaotong University","ror":"https://ror.org/017zhmm22","country_code":"CN","type":"education","lineage":["https://openalex.org/I87445476"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Haiduo Huang","raw_affiliation_strings":["Xi'an Jiaotong University"],"affiliations":[{"raw_affiliation_string":"Xi'an Jiaotong University","institution_ids":["https://openalex.org/I87445476"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129669952","display_name":"Zhenhua Liu","orcid":null},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhenhua Liu","raw_affiliation_strings":["Peking University"],"affiliations":[{"raw_affiliation_string":"Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129730098","display_name":"Tian Xia","orcid":null},"institutions":[{"id":"https://openalex.org/I87445476","display_name":"Xi'an Jiaotong University","ror":"https://ror.org/017zhmm22","country_code":"CN","type":"education","lineage":["https://openalex.org/I87445476"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tian Xia","raw_affiliation_strings":["Xi'an Jiaotong University"],"affiliations":[{"raw_affiliation_string":"Xi'an Jiaotong University","institution_ids":["https://openalex.org/I87445476"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129676131","display_name":"Pengju Ren","orcid":null},"institutions":[{"id":"https://openalex.org/I87445476","display_name":"Xi'an Jiaotong University","ror":"https://ror.org/017zhmm22","country_code":"CN","type":"education","lineage":["https://openalex.org/I87445476"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Pengju Ren","raw_affiliation_strings":["Xi'an Jiaotong University"],"affiliations":[{"raw_affiliation_string":"Xi'an Jiaotong University","institution_ids":["https://openalex.org/I87445476"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5003944732"],"corresponding_institution_ids":["https://openalex.org/I87445476"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.13020833,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"40","issue":"26","first_page":"21921","last_page":"21929"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9151999950408936,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9151999950408936,"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"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.02199999988079071,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.007499999832361937,"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/rounding","display_name":"Rounding","score":0.8492000102996826},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.7806000113487244},{"id":"https://openalex.org/keywords/lossless-compression","display_name":"Lossless compression","score":0.671500027179718},{"id":"https://openalex.org/keywords/segmentation","display_name":"Segmentation","score":0.42579999566078186},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.39649999141693115},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.3919999897480011},{"id":"https://openalex.org/keywords/vector-quantization","display_name":"Vector quantization","score":0.3677999973297119},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.351500004529953}],"concepts":[{"id":"https://openalex.org/C136625980","wikidata":"https://www.wikidata.org/wiki/Q663208","display_name":"Rounding","level":2,"score":0.8492000102996826},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7806000113487244},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7019000053405762},{"id":"https://openalex.org/C81081738","wikidata":"https://www.wikidata.org/wiki/Q55542","display_name":"Lossless compression","level":3,"score":0.671500027179718},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5292999744415283},{"id":"https://openalex.org/C89600930","wikidata":"https://www.wikidata.org/wiki/Q1423946","display_name":"Segmentation","level":2,"score":0.42579999566078186},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.39649999141693115},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.3919999897480011},{"id":"https://openalex.org/C199833920","wikidata":"https://www.wikidata.org/wiki/Q612536","display_name":"Vector quantization","level":2,"score":0.3677999973297119},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.351500004529953},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.34470000863075256},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.33709999918937683},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.32190001010894775},{"id":"https://openalex.org/C761482","wikidata":"https://www.wikidata.org/wiki/Q118093","display_name":"Transmission (telecommunications)","level":2,"score":0.2953000068664551},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.29440000653266907},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.2872999906539917},{"id":"https://openalex.org/C2780056265","wikidata":"https://www.wikidata.org/wiki/Q106239881","display_name":"High dynamic range","level":3,"score":0.2851000130176544},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.28369998931884766},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.2680000066757202},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2660999894142151},{"id":"https://openalex.org/C87133666","wikidata":"https://www.wikidata.org/wiki/Q1161699","display_name":"Dynamic range","level":2,"score":0.2567000091075897},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.2538999915122986},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.2517000138759613}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1609/aaai.v40i26.39345","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i26.39345","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39345/43306","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1609/aaai.v40i26.39345","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i26.39345","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39345/43306","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1309785897","display_name":null,"funder_award_id":"52441602","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G1343088833","display_name":null,"funder_award_id":"2088102","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G2087396116","display_name":null,"funder_award_id":"China","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G3317480652","display_name":null,"funder_award_id":"Science","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G3833898155","display_name":null,"funder_award_id":"62088102","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G391238517","display_name":null,"funder_award_id":", and","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G5340218887","display_name":null,"funder_award_id":"62302381","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G5994120800","display_name":null,"funder_award_id":"Natural","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G6058138561","display_name":null,"funder_award_id":", No.","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8467678729","display_name":null,"funder_award_id":"6208810","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7137898003.pdf","grobid_xml":"https://content.openalex.org/works/W7137898003.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Model":[0],"quantization":[1,87,97,143],"is":[2,51],"widely":[3],"applied":[4],"for":[5,71,166],"compressing":[6],"and":[7,32,39,185,214,225],"accelerating":[8],"deep":[9],"neural":[10],"networks":[11],"(DNNs).":[12],"However,":[13],"conventional":[14],"Quantization-Aware":[15],"Training":[16],"(QAT)":[17],"focuses":[18],"on":[19,195,223,231],"training":[20,38,48,172,184],"DNNs":[21],"with":[22],"uniform":[23],"bit-width.":[24],"The":[25],"bit-width":[26],"settings":[27],"vary":[28],"across":[29],"different":[30,68,131],"hardware":[31],"transmission":[33],"demands,":[34],"which":[35,99],"induces":[36],"considerable":[37],"storage":[40,113],"costs.":[41],"Hence,":[42],"the":[43,94,102,116,171,196,218],"scheme":[44],"of":[45,121,142,220],"one-shot":[46,134,181,208],"joint":[47,135,209],"multiple":[49],"precisions":[50,132,168],"proposed":[52],"to":[53,65,84,106,139,169,180,206],"address":[54],"this":[55,90,150],"issue.":[56],"Previous":[57],"works":[58],"either":[59],"store":[60,75],"a":[61,76,127,187],"larger":[62],"FP32":[63],"model":[64,79],"switch":[66],"between":[67],"precision":[69,119,183],"models":[70],"higher":[72],"accuracy":[73,82],"or":[74],"smaller":[77],"INT8":[78],"but":[80],"compromise":[81],"due":[83,138],"using":[85,115],"shared":[86],"parameters.":[88],"In":[89],"paper,":[91],"we":[92,125,152,175],"introduce":[93],"Double":[95,178],"Rounding":[96,179],"method,":[98],"fully":[100],"utilizes":[101],"quantized":[103],"representation":[104],"range":[105],"accomplish":[107],"nearly":[108],"lossless":[109],"bit-switching":[110],"while":[111],"reducing":[112],"by":[114],"highest":[117],"integer":[118],"instead":[120],"full":[122],"precision.":[123],"Furthermore,":[124],"observe":[126],"competitive":[128],"interference":[129],"among":[130],"during":[133,145],"training,":[136],"primarily":[137],"inconsistent":[140],"gradients":[141],"scales":[144],"backward":[146],"propagation.":[147],"To":[148],"tackle":[149],"problem,":[151],"propose":[153],"an":[154],"Adaptive":[155],"Learning":[156],"Rate":[157],"Scaling":[158],"(AdaScale)":[159],"technique":[160],"that":[161,200],"dynamically":[162],"adapts":[163],"learning":[164],"rates":[165],"various":[167],"optimize":[170],"process.":[173],"Additionally,":[174],"extend":[176],"our":[177,201,221],"mixed":[182],"develop":[186],"Hessian-Aware":[188],"Stochastic":[189],"Bit-switching":[190],"(HessBit)":[191],"strategy.":[192],"Experimental":[193],"results":[194],"ImageNet-1K":[197],"classification":[198],"demonstrate":[199],"methods":[202],"have":[203],"enough":[204],"advantages":[205],"state-of-the-art":[207],"QAT":[210],"in":[211],"both":[212],"multi-precision":[213],"mixed-precision.":[215],"We":[216],"validate":[217],"feasibility":[219],"method":[222],"detection":[224],"segmentation":[226],"tasks,":[227],"as":[228,230],"well":[229],"LLMs":[232],"task.":[233]},"counts_by_year":[],"updated_date":"2026-04-03T22:45:19.894376","created_date":"2026-03-18T00:00:00"}