{"id":"https://openalex.org/W7134168205","doi":"https://doi.org/10.1002/cpe.70626","title":"A Multi\u2010Level Acceleration Scheme for AI Model Training on ARM Architecture Processors","display_name":"A Multi\u2010Level Acceleration Scheme for AI Model Training on ARM Architecture Processors","publication_year":2026,"publication_date":"2026-02-25","ids":{"openalex":"https://openalex.org/W7134168205","doi":"https://doi.org/10.1002/cpe.70626"},"language":"en","primary_location":{"id":"doi:10.1002/cpe.70626","is_oa":false,"landing_page_url":"https://doi.org/10.1002/cpe.70626","pdf_url":null,"source":{"id":"https://openalex.org/S11065456","display_name":"Concurrency and Computation Practice and Experience","issn_l":"1532-0626","issn":["1532-0626","1532-0634"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320595","host_organization_name":"Wiley","host_organization_lineage":["https://openalex.org/P4310320595"],"host_organization_lineage_names":["Wiley"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Concurrency and Computation: Practice and Experience","raw_type":"journal-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/A5128373187","display_name":"Yang Bai","orcid":null},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang Bai","raw_affiliation_strings":["School of Cyberspace Science Harbin Institute of Technology  Harbin China"],"raw_orcid":"https://orcid.org/0009-0000-3422-8634","affiliations":[{"raw_affiliation_string":"School of Cyberspace Science Harbin Institute of Technology  Harbin China","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5008455219","display_name":"Mingdong Xie","orcid":null},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"MingDong Xie","raw_affiliation_strings":["School of Cyberspace Science Harbin Institute of Technology  Harbin China"],"raw_orcid":"https://orcid.org/0009-0007-4957-5672","affiliations":[{"raw_affiliation_string":"School of Cyberspace Science Harbin Institute of Technology  Harbin China","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060954730","display_name":"Meng Hao","orcid":"https://orcid.org/0000-0003-0043-4370"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Meng Hao","raw_affiliation_strings":["School of Cyberspace Science Harbin Institute of Technology  Harbin China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Cyberspace Science Harbin Institute of Technology  Harbin China","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128336983","display_name":"WeiZhe Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"WeiZhe Zhang","raw_affiliation_strings":["School of Cyberspace Science Harbin Institute of Technology  Harbin China"],"raw_orcid":"https://orcid.org/0000-0003-4783-876X","affiliations":[{"raw_affiliation_string":"School of Cyberspace Science Harbin Institute of Technology  Harbin China","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":null,"display_name":"HongWei Yang","orcid":null},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"HongWei Yang","raw_affiliation_strings":["School of Cyberspace Science Harbin Institute of Technology  Harbin China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Cyberspace Science Harbin Institute of Technology  Harbin China","institution_ids":["https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128337980","display_name":"Li He","orcid":null},"institutions":[{"id":"https://openalex.org/I180662265","display_name":"China Mobile (China)","ror":"https://ror.org/05gftfe97","country_code":"CN","type":"company","lineage":["https://openalex.org/I180662265"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li He","raw_affiliation_strings":["State Key Laboratory of Mobile Network and Mobile Multimedia Technology, ZTE  Shenzhen China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"State Key Laboratory of Mobile Network and Mobile Multimedia Technology, ZTE  Shenzhen China","institution_ids":["https://openalex.org/I180662265"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057125041","display_name":"Wei Duan","orcid":"https://orcid.org/0000-0003-1678-9520"},"institutions":[{"id":"https://openalex.org/I180662265","display_name":"China Mobile (China)","ror":"https://ror.org/05gftfe97","country_code":"CN","type":"company","lineage":["https://openalex.org/I180662265"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wei Duan","raw_affiliation_strings":["State Key Laboratory of Mobile Network and Mobile Multimedia Technology, ZTE  Shenzhen China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"State Key Laboratory of Mobile Network and Mobile Multimedia Technology, ZTE  Shenzhen China","institution_ids":["https://openalex.org/I180662265"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5128363152","display_name":"NingCheng Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I180662265","display_name":"China Mobile (China)","ror":"https://ror.org/05gftfe97","country_code":"CN","type":"company","lineage":["https://openalex.org/I180662265"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"NingCheng Wang","raw_affiliation_strings":["State Key Laboratory of Mobile Network and Mobile Multimedia Technology, ZTE  Shenzhen China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"State Key Laboratory of Mobile Network and Mobile Multimedia Technology, ZTE  Shenzhen China","institution_ids":["https://openalex.org/I180662265"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5060954730"],"corresponding_institution_ids":["https://openalex.org/I204983213"],"apc_list":{"value":4740,"currency":"USD","value_usd":4740},"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.28218752,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"38","issue":"5","first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.41110000014305115,"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":0.41110000014305115,"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/T14347","display_name":"Big Data and Digital Economy","score":0.04100000113248825,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.039500001817941666,"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/speedup","display_name":"Speedup","score":0.9153000116348267},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.5810999870300293},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5684999823570251},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.4959999918937683},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.47999998927116394},{"id":"https://openalex.org/keywords/simd","display_name":"SIMD","score":0.44780001044273376},{"id":"https://openalex.org/keywords/edge-device","display_name":"Edge device","score":0.4350000023841858},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.4348999857902527}],"concepts":[{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.9153000116348267},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.888700008392334},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.5810999870300293},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5684999823570251},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.5029000043869019},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.4959999918937683},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4832000136375427},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.47999998927116394},{"id":"https://openalex.org/C150552126","wikidata":"https://www.wikidata.org/wiki/Q339387","display_name":"SIMD","level":2,"score":0.44780001044273376},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.4350000023841858},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.4348999857902527},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.41850000619888306},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.39320001006126404},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.37689998745918274},{"id":"https://openalex.org/C26771161","wikidata":"https://www.wikidata.org/wiki/Q16980","display_name":"ARM architecture","level":2,"score":0.3718000054359436},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.3443000018596649},{"id":"https://openalex.org/C2778456923","wikidata":"https://www.wikidata.org/wiki/Q5337692","display_name":"Edge computing","level":3,"score":0.3375999927520752},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3199000060558319},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.31119999289512634},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.2718999981880188},{"id":"https://openalex.org/C160633673","wikidata":"https://www.wikidata.org/wiki/Q355198","display_name":"Pixel","level":2,"score":0.2572999894618988}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1002/cpe.70626","is_oa":false,"landing_page_url":"https://doi.org/10.1002/cpe.70626","pdf_url":null,"source":{"id":"https://openalex.org/S11065456","display_name":"Concurrency and Computation Practice and Experience","issn_l":"1532-0626","issn":["1532-0626","1532-0634"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320595","host_organization_name":"Wiley","host_organization_lineage":["https://openalex.org/P4310320595"],"host_organization_lineage_names":["Wiley"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Concurrency and Computation: Practice and Experience","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G8725492322","display_name":null,"funder_award_id":"62202123","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":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":14,"referenced_works":["https://openalex.org/W2112796928","https://openalex.org/W2604319603","https://openalex.org/W2883780447","https://openalex.org/W2963122961","https://openalex.org/W2967733054","https://openalex.org/W2982083293","https://openalex.org/W3040269689","https://openalex.org/W3158216587","https://openalex.org/W3177243942","https://openalex.org/W3210601829","https://openalex.org/W4301418577","https://openalex.org/W4379158112","https://openalex.org/W4386918877","https://openalex.org/W4399304422"],"related_works":[],"abstract_inverted_index":{"ABSTRACT":[0],"With":[1],"the":[2,64,70,93,135,155,204,225,229,241],"widespread":[3],"application":[4],"of":[5,78,228],"reinforcement":[6,190],"learning":[7,10,30,191],"and":[8,54,74,112,127,138,159,164,171,208,212,219,235,244],"deep":[9,29],"on":[11,17,39,63,83,125,182],"edge":[12,252],"devices,":[13],"training":[14,37,45,59,72,89,117,180],"neural":[15],"networks":[16],"ARM":[18,40,84,104],"architecture":[19],"processors":[20],"has":[21],"become":[22],"an":[23,114,142,246],"urgent":[24],"demand.":[25],"However,":[26],"existing":[27],"mainstream":[28,139],"frameworks":[31],"are":[32],"not":[33],"sufficiently":[34],"optimized":[35],"for":[36,192,250],"workloads":[38],"CPUs,":[41],"resulting":[42],"in":[43,179],"low":[44],"efficiency.":[46],"To":[47],"address":[48],"this":[49,51],"problem,":[50],"paper":[52],"proposes":[53],"implements":[55],"a":[56,183,198],"multi\u2010level":[57],"AI":[58],"acceleration":[60],"scheme":[61,87],"based":[62],"open\u2010source":[65],"C++":[66],"library":[67],"mlpack,":[68],"targeting":[69],"slow":[71],"speed":[73],"high":[75],"resource":[76],"consumption":[77],"Convolutional":[79],"Neural":[80],"Networks":[81],"(CNNs)":[82],"platforms.":[85],"The":[86],"accelerates":[88],"by":[90],"deeply":[91],"optimizing":[92],"im2col":[94],"algorithm":[95],"to":[96,108,151,161,240],"convert":[97],"convolutions":[98],"into":[99],"efficient":[100],"matrix":[101],"multiplications,":[102],"utilizing":[103],"NEON":[105],"SIMD":[106],"instructions":[107],"optimize":[109],"linear":[110],"operators,":[111],"integrating":[113],"FP64/FP16":[115],"mixed\u2010precision":[116,236],"strategy":[118],"with":[119,232],"dynamic":[120],"loss":[121],"scaling.":[122],"Experimental":[123],"results":[124],"LeNet\u20105":[126],"VGG11\u2010style":[128],"CNNs":[129],"show":[130],"substantial":[131],"performance":[132],"gains":[133],"over":[134,154,169,203,216],"original":[136,156],"mlpack":[137,157],"frameworks.":[140],"On":[141],"NVIDIA":[143],"Jetson":[144],"AGX":[145],"Orin,":[146],"our":[147,195],"implementation":[148],"achieves":[149],"up":[150,160],"7.3":[152],"speedup":[153,202,243],"baseline":[158,207],"11.3":[162],"\u00d7":[163,166,200,211,214],"5.69":[165],"end\u2010to\u2010end":[167,201],"speedups":[168],"PyTorch":[170,205,218],"TensorFlow,":[172],"respectively,":[173],"while":[174],"still":[175],"delivering":[176],"multi\u2010fold":[177],"reductions":[178],"time":[181],"low\u2010resource":[184],"Raspberry":[185],"Pi":[186],"platform.":[187],"In":[188],"DQN\u2010based":[189],"Atari":[193],"Breakout,":[194],"solution":[196],"attains":[197],"4.98":[199],"single\u2010threaded":[206],"maintains":[209],"2.37":[210],"4.23":[213],"advantages":[215],"4\u2010threaded":[217],"TensorFlow":[220],"implementations.":[221],"Ablation":[222],"studies":[223],"confirm":[224],"complementary":[226],"nature":[227],"proposed":[230],"optimizations,":[231],"convolutional,":[233],"linear,":[234],"components":[237],"jointly":[238],"contributing":[239],"overall":[242],"enabling":[245],"attractive":[247],"performance\u2013accuracy":[248],"trade\u2010off":[249],"ARM\u2010based":[251],"computing.":[253]},"counts_by_year":[],"updated_date":"2026-07-21T08:15:58.654021","created_date":"2026-03-09T00:00:00"}