{"id":"https://openalex.org/W3094071641","doi":"https://doi.org/10.1109/fpl50879.2020.00056","title":"An FPGA-Based Low-Latency Accelerator for Randomly Wired Neural Networks","display_name":"An FPGA-Based Low-Latency Accelerator for Randomly Wired Neural Networks","publication_year":2020,"publication_date":"2020-08-01","ids":{"openalex":"https://openalex.org/W3094071641","doi":"https://doi.org/10.1109/fpl50879.2020.00056","mag":"3094071641"},"language":"en","primary_location":{"id":"doi:10.1109/fpl50879.2020.00056","is_oa":false,"landing_page_url":"https://doi.org/10.1109/fpl50879.2020.00056","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 30th International Conference on Field-Programmable Logic and Applications (FPL)","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/A5001468686","display_name":"Ryosuke Kuramochi","orcid":null},"institutions":[{"id":"https://openalex.org/I114531698","display_name":"Tokyo Institute of Technology","ror":"https://ror.org/0112mx960","country_code":"JP","type":"education","lineage":["https://openalex.org/I114531698"]}],"countries":["JP"],"is_corresponding":true,"raw_author_name":"Ryosuke Kuramochi","raw_affiliation_strings":["Tokyo Institute of Technology, Japan"],"affiliations":[{"raw_affiliation_string":"Tokyo Institute of Technology, Japan","institution_ids":["https://openalex.org/I114531698"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5070734898","display_name":"Hiroki Nakahara","orcid":"https://orcid.org/0000-0002-5701-7466"},"institutions":[{"id":"https://openalex.org/I114531698","display_name":"Tokyo Institute of Technology","ror":"https://ror.org/0112mx960","country_code":"JP","type":"education","lineage":["https://openalex.org/I114531698"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Hiroki Nakahara","raw_affiliation_strings":["Tokyo Institute of Technology, Japan"],"affiliations":[{"raw_affiliation_string":"Tokyo Institute of Technology, Japan","institution_ids":["https://openalex.org/I114531698"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5001468686"],"corresponding_institution_ids":["https://openalex.org/I114531698"],"apc_list":null,"apc_paid":null,"fwci":0.977,"has_fulltext":false,"cited_by_count":14,"citation_normalized_percentile":{"value":0.7853902,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"298","last_page":"303"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":1.0,"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":1.0,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9983000159263611,"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/T12702","display_name":"Brain Tumor Detection and Classification","score":0.9980000257492065,"subfield":{"id":"https://openalex.org/subfields/2808","display_name":"Neurology"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8439279794692993},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.6795968413352966},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.6577919721603394},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.5788378715515137},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.5392739772796631},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.46365201473236084},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.4466726779937744},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.4281888008117676},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.42307692766189575},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.32304298877716064},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.3046068251132965}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8439279794692993},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.6795968413352966},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.6577919721603394},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5788378715515137},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.5392739772796631},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.46365201473236084},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.4466726779937744},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4281888008117676},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.42307692766189575},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.32304298877716064},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.3046068251132965},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/fpl50879.2020.00056","is_oa":false,"landing_page_url":"https://doi.org/10.1109/fpl50879.2020.00056","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 30th International Conference on Field-Programmable Logic and Applications (FPL)","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":33,"referenced_works":["https://openalex.org/W639708223","https://openalex.org/W1836465849","https://openalex.org/W2057392949","https://openalex.org/W2063747323","https://openalex.org/W2092333800","https://openalex.org/W2094756095","https://openalex.org/W2117539524","https://openalex.org/W2194775991","https://openalex.org/W2302255633","https://openalex.org/W2466675884","https://openalex.org/W2531409750","https://openalex.org/W2559085405","https://openalex.org/W2570343428","https://openalex.org/W2795915628","https://openalex.org/W2905110430","https://openalex.org/W2963122961","https://openalex.org/W2963881378","https://openalex.org/W2963911037","https://openalex.org/W2970971581","https://openalex.org/W2981985696","https://openalex.org/W2985161055","https://openalex.org/W2997828269","https://openalex.org/W3106250896","https://openalex.org/W3143219376","https://openalex.org/W4241669766","https://openalex.org/W4252973173","https://openalex.org/W4295312788","https://openalex.org/W6638444622","https://openalex.org/W6638667902","https://openalex.org/W6674479107","https://openalex.org/W6766978945","https://openalex.org/W6785652829","https://openalex.org/W6793127252"],"related_works":["https://openalex.org/W2111241003","https://openalex.org/W2355315220","https://openalex.org/W4200391368","https://openalex.org/W2210979487","https://openalex.org/W2316202402","https://openalex.org/W2074043759","https://openalex.org/W2082487009","https://openalex.org/W2373535795","https://openalex.org/W2518118925","https://openalex.org/W3159273459"],"abstract_inverted_index":{"Convolutional":[0],"neural":[1,74],"networks":[2,75],"(CNNs)":[3],"are":[4,31,80,131],"widely":[5],"used":[6],"for":[7,33,71,180,230],"image":[8,39,213,220],"tasks":[9,35],"in":[10,21,56,91,188,221],"both":[11],"embedded":[12],"systems":[13],"and":[14,28,128,199,249],"data":[15,23],"centers.":[16],"Particularly,":[17],"when":[18],"deploying":[19],"CNNs":[20],"a":[22,52,134,151,167,203,231,247,254],"center,":[24],"achieving":[25],"high":[26,59,63],"accuracy":[27],"low":[29],"latency":[30,97,241],"important":[32],"various":[34],"such":[36],"as":[37,112],"the":[38,96,106,125,140,162,173,181,211,240],"processing":[40],"of":[41],"streaming":[42],"videos.":[43],"However,":[44],"conventional":[45,232],"CNN":[46],"accelerators":[47],"often":[48],"use":[49,105],"architectures":[50],"with":[51,202,246,253],"deep":[53],"pipeline,":[54],"resulting":[55],"not":[57],"only":[58],"throughput":[60],"but":[61],"also":[62],"latency.":[64],"We":[65,104,143,190],"propose":[66],"an":[67,195,219],"FPGA-based":[68,204],"inference":[69,205],"accelerator":[70,206,237],"randomly":[72],"wired":[73],"(RWNNs),":[76],"whose":[77],"layer":[78,146,160],"structures":[79],"based":[81,171],"on":[82,172,194],"random":[83],"graph":[84,169],"models.":[85],"Because":[86],"RWNN":[87],"can":[88,94],"be":[89],"processed":[90],"parallel,":[92],"we":[93,157,216],"reduce":[95,239],"by":[98,165,242],"concurrently":[99],"using":[100,133,150],"multiple":[101,113],"computational":[102,129,148,182],"units.":[103],"HBM2":[107],"to":[108,117,137,147,161,184],"store":[109],"feature":[110,121,141],"maps,":[111],"computing":[114],"units":[115,130,149,183],"need":[116],"simultaneously":[118],"access":[119,185],"different":[120],"maps.":[122,142],"In":[123,155,210,234],"addition,":[124,156,235],"HBM":[126,163,186],"channels":[127,164,187],"connected":[132],"crossbar":[135],"switch":[136],"efficiently":[138],"transfer":[139],"allocate":[144,158],"each":[145,159],"simple":[152],"heuristic":[153],"algorithm.":[154],"coloring":[166],"conflict":[168],"built":[170],"allocated":[174],"schedule.":[175],"This":[176],"makes":[177],"it":[178,201],"possible":[179],"parallel.":[189],"implemented":[191],"our":[192,236],"architecture":[193],"Alveo":[196],"U50":[197],"FPGA":[198],"compared":[200,245,252],"that":[207,229],"targets":[208],"ResNet-50.":[209],"ImageNet":[212],"classification":[214],"task,":[215],"could":[217,238],"process":[218],"16.6":[222],"ms,":[223],"which":[224],"is":[225],"43%":[226],"lower":[227],"than":[228],"accelerator.":[233],"5.53":[243],"times":[244,251],"CPU":[248],"1.98":[250],"GPU.":[255]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":4},{"year":2022,"cited_by_count":3},{"year":2021,"cited_by_count":3}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}