{"id":"https://openalex.org/W4416925291","doi":"https://doi.org/10.1109/wimob66857.2025.11257519","title":"Stability-Driven Quantization-Aware Training for Low-Bit Models","display_name":"Stability-Driven Quantization-Aware Training for Low-Bit Models","publication_year":2025,"publication_date":"2025-10-20","ids":{"openalex":"https://openalex.org/W4416925291","doi":"https://doi.org/10.1109/wimob66857.2025.11257519"},"language":null,"primary_location":{"id":"doi:10.1109/wimob66857.2025.11257519","is_oa":false,"landing_page_url":"https://doi.org/10.1109/wimob66857.2025.11257519","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 21th International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob)","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/A5116975251","display_name":"Du Tran-Ngoc","orcid":null},"institutions":[{"id":"https://openalex.org/I4210153650","display_name":"Viet Duc Hospital","ror":"https://ror.org/049ymah45","country_code":"VN","type":"healthcare","lineage":["https://openalex.org/I4210153650"]}],"countries":["VN"],"is_corresponding":true,"raw_author_name":"Du Tran-Ngoc","raw_affiliation_strings":["Viettel Semiconductor Center,Hanoi,Vietnam"],"affiliations":[{"raw_affiliation_string":"Viettel Semiconductor Center,Hanoi,Vietnam","institution_ids":["https://openalex.org/I4210153650"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5116975250","display_name":"Quang Le-Hoang-Minh","orcid":null},"institutions":[{"id":"https://openalex.org/I4210153650","display_name":"Viet Duc Hospital","ror":"https://ror.org/049ymah45","country_code":"VN","type":"healthcare","lineage":["https://openalex.org/I4210153650"]}],"countries":["VN"],"is_corresponding":false,"raw_author_name":"Quang Le-Hoang-Minh","raw_affiliation_strings":["Viettel Semiconductor Center,Hanoi,Vietnam"],"affiliations":[{"raw_affiliation_string":"Viettel Semiconductor Center,Hanoi,Vietnam","institution_ids":["https://openalex.org/I4210153650"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120627827","display_name":"Thang Nguyen-Minh","orcid":null},"institutions":[{"id":"https://openalex.org/I4210153650","display_name":"Viet Duc Hospital","ror":"https://ror.org/049ymah45","country_code":"VN","type":"healthcare","lineage":["https://openalex.org/I4210153650"]}],"countries":["VN"],"is_corresponding":false,"raw_author_name":"Thang Nguyen-Minh","raw_affiliation_strings":["Viettel Semiconductor Center,Hanoi,Vietnam"],"affiliations":[{"raw_affiliation_string":"Viettel Semiconductor Center,Hanoi,Vietnam","institution_ids":["https://openalex.org/I4210153650"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5120536717","display_name":"Trung Dong","orcid":null},"institutions":[{"id":"https://openalex.org/I4210153650","display_name":"Viet Duc Hospital","ror":"https://ror.org/049ymah45","country_code":"VN","type":"healthcare","lineage":["https://openalex.org/I4210153650"]}],"countries":["VN"],"is_corresponding":false,"raw_author_name":"Trung Dong","raw_affiliation_strings":["Viettel Semiconductor Center,Hanoi,Vietnam"],"affiliations":[{"raw_affiliation_string":"Viettel Semiconductor Center,Hanoi,Vietnam","institution_ids":["https://openalex.org/I4210153650"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5116975251"],"corresponding_institution_ids":["https://openalex.org/I4210153650"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.41504942,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9204999804496765,"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":0.9204999804496765,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.014000000432133675,"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":0.006399999838322401,"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/quantization","display_name":"Quantization (signal processing)","score":0.7498000264167786},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5741000175476074},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.5471000075340271},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4187999963760376},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.41499999165534973},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.36340001225471497},{"id":"https://openalex.org/keywords/object-detection","display_name":"Object detection","score":0.35830000042915344}],"concepts":[{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7498000264167786},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6890000104904175},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5741000175476074},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.5471000075340271},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5194000005722046},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4643999934196472},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4187999963760376},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.41499999165534973},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.36340001225471497},{"id":"https://openalex.org/C2776151529","wikidata":"https://www.wikidata.org/wiki/Q3045304","display_name":"Object detection","level":3,"score":0.35830000042915344},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.34200000762939453},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.33309999108314514},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.31929999589920044},{"id":"https://openalex.org/C29265498","wikidata":"https://www.wikidata.org/wiki/Q7047719","display_name":"Noise measurement","level":3,"score":0.30880001187324524},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2948000133037567},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.29159998893737793},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.289900004863739},{"id":"https://openalex.org/C34736171","wikidata":"https://www.wikidata.org/wiki/Q918333","display_name":"Preprocessor","level":2,"score":0.27390000224113464}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/wimob66857.2025.11257519","is_oa":false,"landing_page_url":"https://doi.org/10.1109/wimob66857.2025.11257519","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 21th International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":16,"referenced_works":["https://openalex.org/W2194775991","https://openalex.org/W2752782242","https://openalex.org/W2883780447","https://openalex.org/W2963163009","https://openalex.org/W2981751377","https://openalex.org/W2982083293","https://openalex.org/W2999803881","https://openalex.org/W3004061291","https://openalex.org/W3035183452","https://openalex.org/W3113272803","https://openalex.org/W3176946833","https://openalex.org/W3177245350","https://openalex.org/W4390638513","https://openalex.org/W4394625638","https://openalex.org/W4403770406","https://openalex.org/W4404521479"],"related_works":[],"abstract_inverted_index":{"Quantization-Aware":[0],"Training":[1],"(QAT)":[2],"has":[3],"gained":[4],"significant":[5],"attention":[6],"thanks":[7],"to":[8,22,32,43,48],"reduced":[9],"memory":[10],"usage":[11],"and":[12,52,89,132],"accelerated":[13],"inference,":[14],"which":[15],"are":[16],"achieved":[17],"by":[18,148],"transforming":[19],"fullprecision":[20],"models":[21],"low-bit":[23,73],"integer":[24],"formats":[25],"while":[26,144],"preserving":[27],"accuracy.":[28],"However,":[29],"applying":[30],"QAT":[31,110],"ultra-low":[33],"precision":[34],"(e.g.,":[35],"4":[36],"bits":[37],"or":[38],"3":[39],"bits)":[40],"can":[41],"lead":[42],"severe":[44],"accuracy":[45,70,118],"degradation":[46,71],"due":[47],"the":[49,66,77,124,128],"improper":[50],"weights":[51],"quantization":[53,96],"parameters":[54],"during":[55],"training.":[56],"In":[57,112],"this":[58],"paper,":[59],"we":[60,79],"present":[61],"a":[62],"comprehensive":[63],"analysis":[64],"of":[65,69,123],"root":[67],"causes":[68],"in":[72],"QAT.":[74],"Based":[75],"on":[76,127,138],"analysis,":[78],"propose":[80],"three":[81],"novel":[82],"techniques":[83],"Gradient":[84],"Scaling-Aware":[85],"Distance,":[86,88],"EMA-Aware":[87],"Adaptive":[90],"Fine-Tuning":[91],"Weight":[92],"that":[93,105],"effectively":[94],"mitigate":[95],"noise":[97],"without":[98],"adding":[99],"inference":[100],"overhead.":[101],"Experimental":[102],"results":[103],"demonstrate":[104],"our":[106,115],"method":[107,116],"outperforms":[108],"existing":[109],"approaches.":[111],"4-bit":[113],"quantization,":[114],"achieves":[117],"drop":[119],"within":[120],"0.7":[121],"%":[122],"full-precision":[125],"model":[126,146],"ImageNet-1k":[129],"classification":[130],"task":[131],"<tex":[133],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[134],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">$\\mathbf{1.":[135],"8":[136],"\\%}$</tex>":[137],"MS":[139],"COCO":[140],"object":[141],"detection":[142],"task,":[143],"reducing":[145],"size":[147],"87.5":[149],"%.":[150]},"counts_by_year":[],"updated_date":"2026-03-07T16:01:11.037858","created_date":"2025-12-02T00:00:00"}