{"id":"https://openalex.org/W4393145452","doi":"https://doi.org/10.1109/asp-dac58780.2024.10473822","title":"Extending Neural Processing Unit and Compiler for Advanced Binarized Neural Networks","display_name":"Extending Neural Processing Unit and Compiler for Advanced Binarized Neural Networks","publication_year":2024,"publication_date":"2024-01-22","ids":{"openalex":"https://openalex.org/W4393145452","doi":"https://doi.org/10.1109/asp-dac58780.2024.10473822"},"language":"en","primary_location":{"id":"doi:10.1109/asp-dac58780.2024.10473822","is_oa":false,"landing_page_url":"http://dx.doi.org/10.1109/asp-dac58780.2024.10473822","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 29th Asia and South Pacific Design Automation Conference (ASP-DAC)","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/A5101322788","display_name":"Minjoon Song","orcid":null},"institutions":[{"id":"https://openalex.org/I48566637","display_name":"Ulsan National Institute of Science and Technology","ror":"https://ror.org/017cjz748","country_code":"KR","type":"education","lineage":["https://openalex.org/I48566637"]}],"countries":["KR"],"is_corresponding":true,"raw_author_name":"Minjoon Song","raw_affiliation_strings":["Ulsan National Institute of Science and Technology (UNIST),AI Graduate School,Ulsan,Korea"],"affiliations":[{"raw_affiliation_string":"Ulsan National Institute of Science and Technology (UNIST),AI Graduate School,Ulsan,Korea","institution_ids":["https://openalex.org/I48566637"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003686675","display_name":"Faaiz Asim","orcid":null},"institutions":[{"id":"https://openalex.org/I48566637","display_name":"Ulsan National Institute of Science and Technology","ror":"https://ror.org/017cjz748","country_code":"KR","type":"education","lineage":["https://openalex.org/I48566637"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Faaiz Asim","raw_affiliation_strings":["Ulsan National Institute of Science and Technology (UNIST),Dept. of Electrical Engineering,Ulsan,Korea"],"affiliations":[{"raw_affiliation_string":"Ulsan National Institute of Science and Technology (UNIST),Dept. of Electrical Engineering,Ulsan,Korea","institution_ids":["https://openalex.org/I48566637"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100780257","display_name":"Jongeun Lee","orcid":"https://orcid.org/0000-0003-1523-2974"},"institutions":[{"id":"https://openalex.org/I48566637","display_name":"Ulsan National Institute of Science and Technology","ror":"https://ror.org/017cjz748","country_code":"KR","type":"education","lineage":["https://openalex.org/I48566637"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jongeun Lee","raw_affiliation_strings":["Ulsan National Institute of Science and Technology (UNIST),AI Graduate School,Ulsan,Korea"],"affiliations":[{"raw_affiliation_string":"Ulsan National Institute of Science and Technology (UNIST),AI Graduate School,Ulsan,Korea","institution_ids":["https://openalex.org/I48566637"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5101322788"],"corresponding_institution_ids":["https://openalex.org/I48566637"],"apc_list":null,"apc_paid":null,"fwci":0.3626,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.61874819,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":94},"biblio":{"volume":"29","issue":null,"first_page":"115","last_page":"120"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.8248999714851379,"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/T10320","display_name":"Neural Networks and Applications","score":0.8248999714851379,"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/computer-science","display_name":"Computer science","score":0.8093804717063904},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.7224436402320862},{"id":"https://openalex.org/keywords/compiler","display_name":"Compiler","score":0.6642193794250488},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.45109790563583374},{"id":"https://openalex.org/keywords/unit","display_name":"Unit (ring theory)","score":0.4107329845428467},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3722265958786011},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.3216460347175598},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.31291189789772034},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.05362847447395325}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8093804717063904},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.7224436402320862},{"id":"https://openalex.org/C169590947","wikidata":"https://www.wikidata.org/wiki/Q47506","display_name":"Compiler","level":2,"score":0.6642193794250488},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45109790563583374},{"id":"https://openalex.org/C122637931","wikidata":"https://www.wikidata.org/wiki/Q118084","display_name":"Unit (ring theory)","level":2,"score":0.4107329845428467},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3722265958786011},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3216460347175598},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.31291189789772034},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.05362847447395325},{"id":"https://openalex.org/C145420912","wikidata":"https://www.wikidata.org/wiki/Q853077","display_name":"Mathematics education","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/asp-dac58780.2024.10473822","is_oa":false,"landing_page_url":"http://dx.doi.org/10.1109/asp-dac58780.2024.10473822","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 29th Asia and South Pacific Design Automation Conference (ASP-DAC)","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":28,"referenced_works":["https://openalex.org/W2300242332","https://openalex.org/W2560662850","https://openalex.org/W2565125333","https://openalex.org/W2765235648","https://openalex.org/W2798341898","https://openalex.org/W2809624076","https://openalex.org/W2887447938","https://openalex.org/W2916954108","https://openalex.org/W2963114857","https://openalex.org/W2982165597","https://openalex.org/W2999093589","https://openalex.org/W3083679519","https://openalex.org/W3093035500","https://openalex.org/W3104151879","https://openalex.org/W3176945641","https://openalex.org/W3184782324","https://openalex.org/W3211180228","https://openalex.org/W4211153881","https://openalex.org/W4295262505","https://openalex.org/W4312121059","https://openalex.org/W4312600348","https://openalex.org/W6693397755","https://openalex.org/W6751109714","https://openalex.org/W6751349269","https://openalex.org/W6760069825","https://openalex.org/W6762586143","https://openalex.org/W6780119413","https://openalex.org/W6999268660"],"related_works":["https://openalex.org/W3096456556","https://openalex.org/W4240253816","https://openalex.org/W2169584677","https://openalex.org/W1713081424","https://openalex.org/W2132806979","https://openalex.org/W2127315869","https://openalex.org/W1523769955","https://openalex.org/W2245390655","https://openalex.org/W4379536100","https://openalex.org/W188282463"],"abstract_inverted_index":{"Binarized":[0],"neural":[1,14,61],"networks":[2],"(BNNs)":[3],"are":[4],"one":[5],"of":[6,56,91],"the":[7,35,124],"most":[8],"promising":[9],"approaches":[10],"to":[11,45,58,68],"deploy":[12],"deep":[13],"network":[15,47],"models":[16],"on":[17,26,105,136],"resource-constrained":[18],"devices.":[19],"However,":[20],"there":[21],"is":[22],"very":[23],"little":[24],"support":[25,69],"compilers":[27],"and":[28,42,65,85,97],"programmable":[29],"accelerators":[30],"for":[31,79,119],"BNNs":[32,37],"especially":[33],"with":[34,82,93,139],"modern":[36,70],"that":[38,110],"use":[39],"scale":[40,83],"factors":[41,84],"skip":[43,86],"connections":[44],"maximize":[46],"performance.":[48],"In":[49],"this":[50],"paper":[51],"we":[52],"present":[53],"a":[54,60,66],"set":[55],"methods":[57],"extend":[59],"processing":[62],"unit":[63],"(NPU)":[64],"compiler":[67,142],"BNNs.":[71],"Our":[72,101],"novel":[73],"ideas":[74],"include":[75],"(i)":[76],"batch-norm":[77],"folding":[78],"binarized":[80],"layers":[81,122],"connections,":[87],"(ii)":[88],"efficient":[89],"handling":[90],"convolutions":[92],"few":[94],"input":[95],"channels,":[96],"(iii)":[98],"bit-packing":[99],"pipelining.":[100],"evaluation":[102],"using":[103],"BiRealNet-18":[104,137],"an":[106],"FPGA":[107],"board":[108],"demonstrates":[109],"our":[111,128],"compiler-architecture":[112],"hybrid":[113],"approach":[114,129],"can":[115],"yield":[116],"significant":[117],"speedups":[118],"binary":[120],"convolution":[121],"over":[123],"baseline":[125],"NPU.":[126],"Also":[127],"gives":[130],"3.6~5.5":[131],"$\\times$":[132],"better":[133],"end-to-end":[134],"performance":[135],"compared":[138],"previous":[140],"BNN":[141],"approaches.":[143]},"counts_by_year":[{"year":2024,"cited_by_count":1}],"updated_date":"2025-12-19T19:40:27.379048","created_date":"2025-10-10T00:00:00"}