{"id":"https://openalex.org/W7027953124","doi":null,"title":"Efficient Look-Up Table Implementation of Neural Networks with Binary Activation Through Layerwise Pruning","display_name":"Efficient Look-Up Table Implementation of Neural Networks with Binary Activation Through Layerwise Pruning","alternate_titles":[],"publication_year":2023,"publication_date":"2023-04-01","ids":{"openalex":"https://openalex.org/W7027953124"},"language":"fr","primary_location":{"id":"pmh:oai:publications.polymtl.ca:53415","doi":null,"is_oa":true,"landing_page_url":"https://publications.polymtl.ca/53415/","pdf_url":"https://publications.polymtl.ca/53415/1/2023_AmiraliEbrahimi.pdf","source":{"id":"https://openalex.org/S4306401013","display_name":"PolyPublie (\u00c9cole Polytechnique de Montr\u00e9al)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I45683168","host_organization_name":"Polytechnique Montr\u00e9al","host_organization_lineage":["https://openalex.org/I45683168"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Ebrahimi, A. (2023). Efficient Look-Up Table Implementation of Neural Networks with Binary Activation Through Layerwise Pruning [Master's thesis, Polytechnique Montr\u00e9al].","raw_type":"m\u00e9moire de ma\u00eetrise"},"parent_works":[],"lineage":["https://openalex.org/W7027953124"],"child_works_count":0,"type":"dissertation","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://publications.polymtl.ca/53415/1/2023_AmiraliEbrahimi.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Ebrahimi, Amirali","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I45683168","display_name":"Polytechnique Montr\u00e9al","ror":"https://ror.org/05f8d4e86","country_code":"CA","type":"education","lineage":["https://openalex.org/I45683168"]}],"countries":["CA"],"is_corresponding":true,"contributor_roles":[],"raw_author_name":"Ebrahimi, Amirali","raw_affiliation_strings":["D\u00e9partement de g\u00e9nie \u00e9lectrique, POLYTECHNIQUE MONTR\u00c9AL"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"D\u00e9partement de g\u00e9nie \u00e9lectrique, POLYTECHNIQUE MONTR\u00c9AL","institution_ids":["https://openalex.org/I45683168"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I45683168"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_pulled":false,"is_paratext":false,"is_xpac":true,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.8521000146865845,"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":0.8521000146865845,"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/T10904","display_name":"Embedded Systems Design Techniques","score":0.03620000183582306,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10320","display_name":"Neural Networks and Applications","score":0.025299999862909317,"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/neural-network-hardware-acceleration","display_name":"neural network hardware acceleration","score":0.6129999756813049},{"id":"https://openalex.org/keywords/lookup-tables","display_name":"look-up table","score":0.625},{"id":"https://openalex.org/keywords/fpga","display_name":"FPGA","score":0.6869999766349792},{"id":"https://openalex.org/keywords/binary-neural-networks","display_name":"binary neural networks","score":0.7590000033378601},{"id":"https://openalex.org/keywords/model-compression","display_name":"model compression","score":0.7319999933242798},{"id":"https://openalex.org/keywords/quantization","display_name":"quantization","score":0.7070000171661377},{"id":"https://openalex.org/keywords/energy-efficiency","display_name":"energy efficiency","score":0.5820000171661377},{"id":"https://openalex.org/keywords/mnist","display_name":"MNIST","score":0.6230000257492065},{"id":"https://openalex.org/keywords/embedded-devices","display_name":"embedded devices","score":0.4909999966621399},{"id":"https://openalex.org/keywords/mobile-devices","display_name":"mobile devices","score":0.6200000047683716},{"id":"https://openalex.org/keywords/loss-of-precision","display_name":"precision loss","score":0.32499998807907104},{"id":"https://openalex.org/keywords/throughput","display_name":"throughput","score":0.38600000739097595},{"id":"https://openalex.org/keywords/memory-reduction","display_name":"memory reduction","score":0.4169999957084656},{"id":"https://openalex.org/keywords/hardware-latency","display_name":"hardware latency","score":0.38199999928474426}],"concepts":[{"id":"https://openalex.org/C106131492","wikidata":"https://www.wikidata.org/wiki/Q3072260","display_name":"Filter (signal processing)","level":2,"score":0.2612000107765198},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.2500999867916107},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.2223999947309494},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.2013999968767166},{"id":"https://openalex.org/C144251240","wikidata":"https://www.wikidata.org/wiki/Q7068229","display_name":"Nucleofection","level":4,"score":0.1941000074148178},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.186599999666214},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.1808999925851822},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.1754000037908554},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.17520000040531158},{"id":"https://openalex.org/C2777158700","wikidata":"https://www.wikidata.org/wiki/Q1419356","display_name":"Hyporeflexia","level":3,"score":0.17419999837875366}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:publications.polymtl.ca:53415","doi":null,"is_oa":true,"landing_page_url":"https://publications.polymtl.ca/53415/","pdf_url":"https://publications.polymtl.ca/53415/1/2023_AmiraliEbrahimi.pdf","source":{"id":"https://openalex.org/S4306401013","display_name":"PolyPublie (\u00c9cole Polytechnique de Montr\u00e9al)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I45683168","host_organization_name":"Polytechnique Montr\u00e9al","host_organization_lineage":["https://openalex.org/I45683168"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Ebrahimi, A. (2023). Efficient Look-Up Table Implementation of Neural Networks with Binary Activation Through Layerwise Pruning [Master's thesis, Polytechnique Montr\u00e9al].","raw_type":"m\u00e9moire de ma\u00eetrise"}],"best_oa_location":{"id":"pmh:oai:publications.polymtl.ca:53415","doi":null,"is_oa":true,"landing_page_url":"https://publications.polymtl.ca/53415/","pdf_url":"https://publications.polymtl.ca/53415/1/2023_AmiraliEbrahimi.pdf","source":{"id":"https://openalex.org/S4306401013","display_name":"PolyPublie (\u00c9cole Polytechnique de Montr\u00e9al)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I45683168","host_organization_name":"Polytechnique Montr\u00e9al","host_organization_lineage":["https://openalex.org/I45683168"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Ebrahimi, A. (2023). Efficient Look-Up Table Implementation of Neural Networks with Binary Activation Through Layerwise Pruning [Master's thesis, Polytechnique Montr\u00e9al].","raw_type":"m\u00e9moire de ma\u00eetrise"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7027953124.pdf","grobid_xml":"https://content.openalex.org/works/W7027953124.grobid-xml"},"referenced_works_count":56,"referenced_works":["https://openalex.org/W1686810756","https://openalex.org/W1910657905","https://openalex.org/W1913356549","https://openalex.org/W1982366717","https://openalex.org/W2108598243","https://openalex.org/W2112796928","https://openalex.org/W2117539524","https://openalex.org/W2163605009","https://openalex.org/W2167222293","https://openalex.org/W2194775991","https://openalex.org/W2242818861","https://openalex.org/W2267635276","https://openalex.org/W2296509296","https://openalex.org/W2300242332","https://openalex.org/W2395611524","https://openalex.org/W2412782625","https://openalex.org/W2513491952","https://openalex.org/W2515385951","https://openalex.org/W2565125333","https://openalex.org/W2604700561","https://openalex.org/W2617825408","https://openalex.org/W2786288715","https://openalex.org/W2792643794","https://openalex.org/W2795765840","https://openalex.org/W2798084934","https://openalex.org/W2809624076","https://openalex.org/W2887447938","https://openalex.org/W2904282303","https://openalex.org/W2927678548","https://openalex.org/W2951537853","https://openalex.org/W2952797486","https://openalex.org/W2963163009","https://openalex.org/W2963374099","https://openalex.org/W2964195914","https://openalex.org/W2970971581","https://openalex.org/W2996432061","https://openalex.org/W2997106510","https://openalex.org/W3004379541","https://openalex.org/W3006266031","https://openalex.org/W3007633413","https://openalex.org/W3008515144","https://openalex.org/W3010695715","https://openalex.org/W3016036342","https://openalex.org/W3034480170","https://openalex.org/W3104151879","https://openalex.org/W3116219244","https://openalex.org/W3117511472","https://openalex.org/W3143008962","https://openalex.org/W3166559827","https://openalex.org/W4200506235","https://openalex.org/W4231109964","https://openalex.org/W4286906772","https://openalex.org/W4287901333","https://openalex.org/W4289744896","https://openalex.org/W4311007713","https://openalex.org/W4394642372"],"related_works":["https://openalex.org/W7030051489","https://openalex.org/W4385624928","https://openalex.org/W2924130671","https://openalex.org/W3083246355","https://openalex.org/W7135169239","https://openalex.org/W2947996606","https://openalex.org/W3012650805","https://openalex.org/W4392871178","https://openalex.org/W2555421572","https://openalex.org/W3101144093"],"abstract_inverted_index":{"R\u00c9SUM\u00c9:":[0],"R\u00c9SUM\u00c9":[1],"Les":[2,280,336],"r\u00e9seaux":[3,71,351],"de":[4,17,26,30,32,44,47,65,70,72,79,81,113,126,172,178,221,236,238,247,257,263,287,309,326,332,334,350,352],"neurones":[5,73,114,353],"artificiels":[6],"ont":[7],"gagn\u00e9":[8],"en":[9,15,21,24,38,68,100,122,261,297,347],"popularit\u00e9":[10],"dans":[11],"divers":[12],"domaines.":[13],"Cependant,":[14],"raison":[16,43],"leurs":[18],"besoins":[19],"\u00e9lev\u00e9s":[20],"m\u00e9moire":[22],"et":[23,49,56,87,217,291,319],"ressources":[25,371],"traitement,":[27],"la":[28,45,66,85,98,133,179,184,206,245,292,345],"s\u00e9lection":[29],"plates-formes":[31],"traitement":[33,48],"appropri\u00e9es":[34],"pour":[35,92,196,213,266,366],"leur":[36,201],"mise":[37,67,99,346],"\u0153uvre":[39,69,348],"est":[40,312,364],"cruciale.":[41],"En":[42,135],"puissance":[46],"du":[50,186,259,278],"stockage":[51],"limit\u00e9s":[52],"des":[53,58,116,120,124,141,229,275,355,359,370],"appareils":[54],"mobiles":[55],"embarqu\u00e9s,":[57],"mesures":[59],"sp\u00e9ciales":[60],"doivent":[61],"\u00eatre":[62,90,211,272],"prises":[63],"lors":[64],"sur":[74,132,183,205,244,307,358],"ces":[75],"appareils.":[76],"Des":[77],"techniques":[78,426],"compression":[80,425],"mod\u00e8le":[82,216],"telles":[83],"que":[84,154,283,340],"quantification":[86],"l'\u00e9lagage":[88],"peuvent":[89],"utilis\u00e9es":[91,195],"r\u00e9duire":[93],"les":[94,198,226],"co\u00fbts":[95],"associ\u00e9s":[96],"\u00e0":[97,119,140,274,303],"\u0153uvre.":[101],"Cette":[102],"recherche":[103],"pr\u00e9sente":[104],"un":[105,111,129,253,299],"nouvel":[106],"algorithme":[107],"qui":[108,175,208,269,363],"peut":[109,209,270,343],"associer":[110],"r\u00e9seau":[112,148,300,325],"avec":[115,128,328,354,369],"activations":[117,356,456,660],"binaires":[118,190,357],"LUT,":[121,159,207,268,288,585],"particulier":[123],"LUT":[125,276,489],"FPGA,":[127,361],"impact":[130,464],"minimal":[131,463],"pr\u00e9cision.":[134,223],"int\u00e9grant":[136],"le":[137,147,155,162,215,239,285,289,324],"concept":[138,471],"d'\u00e9lagage":[139,174],"consid\u00e9rations":[142],"architecturales,":[143],"l'algorithme":[144,237],"propos\u00e9":[145],"\u00e9lague":[146],"couch\u00e9":[149,151],"par":[150,161,200],"jusqu'\u00e0":[152],"ce":[153,362],"nombre":[156,286],"souhait\u00e9":[157],"d'entr\u00e9es":[158],"d\u00e9fini":[160],"concepteur,":[163],"soit":[164],"atteint.":[165],"Nous":[166],"introduisons":[167],"\u00e9galement":[168],"une":[169,219,234,330],"nouvelle":[170],"m\u00e9trique":[171],"d\u00e9faut":[173],"consid\u00e8re":[176],"l'effet":[177],"suppression":[180],"d'une":[181],"connexion":[182],"pr\u00e9cision":[185,333],"mod\u00e8le.":[187],"Nos":[188],"couches":[189,199,305],"co-\u00e9qales":[191],"propos\u00e9es":[192],"sont":[193,295],"ensuite":[194,210,271],"remplacer":[197],"version":[202],"\u00e9lagu\u00e9e":[203],"bas\u00e9e":[204],"utilis\u00e9e":[212],"recycler":[214],"r\u00e9cup\u00e9rer":[218],"partie":[220],"sa":[222],"Pour":[224],"remplir":[225],"entr\u00e9es":[227],"manquantes":[228],"LUTs":[230,594],"g\u00e9n\u00e9r\u00e9es,":[231],"nous":[232,251],"utilisons":[233],"variante":[235],"plus":[240,314,317],"proches":[241],"voisins":[242],"base":[243],"distance":[246],"Hamming.":[248],"De":[249],"plus,":[250],"fournissons":[252],"script":[254,575],"Python":[255,574],"capable":[256],"g\u00e9n\u00e9rer":[258],"code":[260,582],"langage":[262],"description":[264],"mat\u00e9rielle":[265],"chaque":[267],"associ\u00e9":[273],"physiques":[277],"FPGA.":[279,597],"r\u00e9sultats":[281,337],"montrent":[282],"lorsque":[284],"d\u00e9bit":[290],"consommation":[293],"d'\u00e9nergie":[294],"pris":[296],"compte,":[298],"enti\u00e8rement":[301],"connect\u00e9":[302],"quatre":[304],"\u00e9lagu\u00e9":[306],"l'ensemble":[308],"donn\u00e9es":[310],"MNIST":[311,621],"7.95\u00d7":[313,624],"petit,":[315],"1.66\u00d7":[316,626],"rapide":[318],"5.66\u00d7":[320,629],"moins":[321],"\u00e9nergivore.":[322],"Que":[323],"base,":[327],"seulement":[329],"baisse":[331],"0,47%.":[335],"empiriques":[338],"sugg\u00e8rent":[339],"l'approche":[341],"propos\u00e9e":[342],"permettre":[344],"efficace":[349],"dispositifs":[360],"crucial":[365,410,666],"diverses":[367],"applications":[368,669],"limit\u00e9es.":[372],"ABSTRACT:":[373],"ABSTRACT":[374],"Artificial":[375],"Neural":[376],"Networks":[377],"have":[378,418],"gained":[379],"significant":[380],"popularity":[381],"across":[382],"different":[383],"domains.":[384],"To":[385,548],"implement":[386],"these":[387],"networks":[388,657],"successfully,":[389],"it":[390],"is":[391,408,495,623,665],"essential":[392],"to":[393,435,457,525,538],"choose":[394],"suitable":[395],"processing":[396,405,420],"platforms":[397],"that":[398,448,505,576,601,646],"can":[399,432,449,534,577,587,650],"meet":[400],"their":[401,529],"high":[402],"memory":[403],"and":[404,413,422,430,542,608,628],"demands.":[406],"This":[407,442],"especially":[409],"for":[411,583,667],"mobile":[412],"embedded":[414],"devices,":[415,663],"which":[416,533,586,664],"often":[417],"limited":[419],"power":[421,609,631],"storage.":[423],"Model":[424],"such":[427],"as":[428],"quantization":[429],"pruning":[431,470,502],"be":[433,536,589],"employed":[434],"reduce":[436],"the":[437,469,475,479,485,493,507,540,554,562,567,592,596,603,620,634,647,652],"costs":[438],"associated":[439],"with":[440,454,462,472,528,636,658,670],"implementation.":[441],"research":[443],"presents":[444],"a":[445,451,500,511,559,573,613,638],"new":[446],"algorithm":[447,477,564],"map":[450],"neural":[452,656],"network":[453,480,618],"binary":[455,659],"LUTs,":[458,461,556,606],"specifically":[459],"FPGA":[460,662],"on":[465,513,566,595,619,661],"accuracy.":[466,515,547],"By":[467],"integrating":[468],"architectural":[473],"considerations,":[474],"proposed":[476,517,648],"prunes":[478],"layer":[481,483],"by":[482,492],"until":[484],"desired":[486],"number":[487,604],"of":[488,509,545,553,561,605,655],"inputs,":[490],"defined":[491],"designer,":[494],"reached.":[496],"We":[497],"also":[498],"introduce":[499],"novel":[501],"fault":[503],"metric":[504],"considers":[506],"effect":[508],"removing":[510],"connection":[512],"model":[514,541],"Our":[516],"Binary":[518],"Layer":[519],"Coequal":[520],"layers":[521,527],"are":[522,611],"then":[523,535,588],"used":[524,537],"replace":[526],"LUT-based":[530],"pruned":[531,614],"version,":[532],"retrain":[539],"recover":[543],"some":[544],"its":[546],"fill":[549],"in":[550],"missing":[551],"entries":[552],"generated":[555],"we":[557,571],"use":[558],"variant":[560],"K-Nearest-neighbours":[563],"based":[565],"Hamming":[568],"distance.":[569],"Furthermore,":[570],"provide":[572],"generate":[578],"Hardware":[579],"Description":[580],"Language":[581],"each":[584],"mapped":[590],"onto":[591],"physical":[593],"The":[598,642],"results":[599,644],"show":[600],"when":[602],"throughput,":[607],"consumption":[610],"considered,":[612],"four-layer":[615],"fully":[616],"connected":[617],"dataset":[622],"smaller,":[625],"faster,":[627],"more":[630],"efficient":[632,653],"than":[633],"baseline,":[635],"only":[637],"0.47%":[639],"accuracy":[640],"drop.":[641],"empirical":[643],"suggest":[645],"approach":[649],"enable":[651],"implementation":[654],"various":[668],"constrained":[671],"resources.":[672]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W7027953124","counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-11-20T13:11:24.684918","created_date":"2025-10-10T00:00:00"}
