{"id":"https://openalex.org/W4213433078","doi":"https://doi.org/10.1109/tpds.2022.3153682","title":"Accelerating Bayesian Neural Networks via Algorithmic and Hardware Optimizations","display_name":"Accelerating Bayesian Neural Networks via Algorithmic and Hardware Optimizations","publication_year":2022,"publication_date":"2022-02-23","ids":{"openalex":"https://openalex.org/W4213433078","doi":"https://doi.org/10.1109/tpds.2022.3153682"},"language":"en","primary_location":{"id":"doi:10.1109/tpds.2022.3153682","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2022.3153682","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Parallel and Distributed Systems","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://discovery.ucl.ac.uk/10150056/1/tpds21_bayescnn_hf5_final.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5057043409","display_name":"Hongxiang Fan","orcid":"https://orcid.org/0000-0003-2387-5611"},"institutions":[{"id":"https://openalex.org/I47508984","display_name":"Imperial College London","ror":"https://ror.org/041kmwe10","country_code":"GB","type":"education","lineage":["https://openalex.org/I47508984"]}],"countries":["GB"],"is_corresponding":true,"raw_author_name":"Hongxiang Fan","raw_affiliation_strings":["Department of Computing, Imperial College London, London, U.K"],"affiliations":[{"raw_affiliation_string":"Department of Computing, Imperial College London, London, U.K","institution_ids":["https://openalex.org/I47508984"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5059088556","display_name":"Martin Ferianc","orcid":"https://orcid.org/0000-0002-4031-6398"},"institutions":[{"id":"https://openalex.org/I45129253","display_name":"University College London","ror":"https://ror.org/02jx3x895","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I45129253"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Martin Ferianc","raw_affiliation_strings":["Department of Electronic and Electrical Engineering, University College London, London, U.K"],"affiliations":[{"raw_affiliation_string":"Department of Electronic and Electrical Engineering, University College London, London, U.K","institution_ids":["https://openalex.org/I45129253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054475218","display_name":"Zhiqiang Que","orcid":"https://orcid.org/0000-0002-9263-6529"},"institutions":[{"id":"https://openalex.org/I47508984","display_name":"Imperial College London","ror":"https://ror.org/041kmwe10","country_code":"GB","type":"education","lineage":["https://openalex.org/I47508984"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Zhiqiang Que","raw_affiliation_strings":["Department of Computing, Imperial College London, London, U.K"],"affiliations":[{"raw_affiliation_string":"Department of Computing, Imperial College London, London, U.K","institution_ids":["https://openalex.org/I47508984"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103060695","display_name":"Xinyu Niu","orcid":"https://orcid.org/0000-0003-0202-9408"},"institutions":[{"id":"https://openalex.org/I4210149914","display_name":"MSIGHT Technologies (China)","ror":"https://ror.org/03rjyvk34","country_code":"CN","type":"company","lineage":["https://openalex.org/I4210149914"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xinyu Niu","raw_affiliation_strings":["Corerain Technologies Ltd., Shenzhen, Guangdong, China"],"affiliations":[{"raw_affiliation_string":"Corerain Technologies Ltd., Shenzhen, Guangdong, China","institution_ids":["https://openalex.org/I4210149914"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5044634366","display_name":"Miguel R. D. Rodrigues","orcid":"https://orcid.org/0000-0002-8908-847X"},"institutions":[{"id":"https://openalex.org/I45129253","display_name":"University College London","ror":"https://ror.org/02jx3x895","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I45129253"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Miguel Rodrigues","raw_affiliation_strings":["Department of Electronic and Electrical Engineering, University College London, London, U.K"],"affiliations":[{"raw_affiliation_string":"Department of Electronic and Electrical Engineering, University College London, London, U.K","institution_ids":["https://openalex.org/I45129253"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5057940557","display_name":"Wayne Luk","orcid":"https://orcid.org/0000-0002-6750-927X"},"institutions":[{"id":"https://openalex.org/I47508984","display_name":"Imperial College London","ror":"https://ror.org/041kmwe10","country_code":"GB","type":"education","lineage":["https://openalex.org/I47508984"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Wayne Luk","raw_affiliation_strings":["Department of Computing, Imperial College London, London, U.K"],"affiliations":[{"raw_affiliation_string":"Department of Computing, Imperial College London, London, U.K","institution_ids":["https://openalex.org/I47508984"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5057043409"],"corresponding_institution_ids":["https://openalex.org/I47508984"],"apc_list":null,"apc_paid":null,"fwci":1.2495,"has_fulltext":true,"cited_by_count":10,"citation_normalized_percentile":{"value":0.82432847,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":99},"biblio":{"volume":"33","issue":"12","first_page":"3387","last_page":"3399"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9993000030517578,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9993000030517578,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9976000189781189,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.994700014591217,"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":0.8716657161712646},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.8343673944473267},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.6657168865203857},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.5471852421760559},{"id":"https://openalex.org/keywords/convolutional-neural-network","display_name":"Convolutional neural network","score":0.5136772990226746},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.46902695298194885},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.4552146792411804},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.45008131861686707},{"id":"https://openalex.org/keywords/monte-carlo-method","display_name":"Monte Carlo method","score":0.4331973195075989},{"id":"https://openalex.org/keywords/hardware-architecture","display_name":"Hardware architecture","score":0.42489734292030334},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3837761878967285},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.3197508454322815},{"id":"https://openalex.org/keywords/computer-hardware","display_name":"Computer hardware","score":0.31627416610717773},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2604570686817169},{"id":"https://openalex.org/keywords/software","display_name":"Software","score":0.14850234985351562}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8716657161712646},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.8343673944473267},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.6657168865203857},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.5471852421760559},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.5136772990226746},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.46902695298194885},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.4552146792411804},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.45008131861686707},{"id":"https://openalex.org/C19499675","wikidata":"https://www.wikidata.org/wiki/Q232207","display_name":"Monte Carlo method","level":2,"score":0.4331973195075989},{"id":"https://openalex.org/C65232700","wikidata":"https://www.wikidata.org/wiki/Q5656403","display_name":"Hardware architecture","level":3,"score":0.42489734292030334},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3837761878967285},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3197508454322815},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.31627416610717773},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2604570686817169},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.14850234985351562},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","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},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.0},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/tpds.2022.3153682","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpds.2022.3153682","pdf_url":null,"source":{"id":"https://openalex.org/S97130795","display_name":"IEEE Transactions on Parallel and Distributed Systems","issn_l":"1045-9219","issn":["1045-9219","1558-2183","2161-9883"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Parallel and Distributed Systems","raw_type":"journal-article"},{"id":"pmh:oai:eprints.ucl.ac.uk.OAI2:10150056","is_oa":true,"landing_page_url":"https://discovery.ucl.ac.uk/id/eprint/10150056/","pdf_url":"https://discovery.ucl.ac.uk/10150056/1/tpds21_bayescnn_hf5_final.pdf","source":{"id":"https://openalex.org/S4306400024","display_name":"UCL Discovery (University College London)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I45129253","host_organization_name":"University College London","host_organization_lineage":["https://openalex.org/I45129253"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"   IEEE Transactions on Parallel and Distributed Systems , 33  (12)   pp. 3387-3399.   (2022)     (In press).  ","raw_type":"Article"},{"id":"pmh:oai:spiral.imperial.ac.uk:10044/1/94643","is_oa":false,"landing_page_url":"http://hdl.handle.net/10044/1/94643","pdf_url":null,"source":{"id":"https://openalex.org/S4306401396","display_name":"Spiral (Imperial College London)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I47508984","host_organization_name":"Imperial College London","host_organization_lineage":["https://openalex.org/I47508984"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"Journal Article"}],"best_oa_location":{"id":"pmh:oai:eprints.ucl.ac.uk.OAI2:10150056","is_oa":true,"landing_page_url":"https://discovery.ucl.ac.uk/id/eprint/10150056/","pdf_url":"https://discovery.ucl.ac.uk/10150056/1/tpds21_bayescnn_hf5_final.pdf","source":{"id":"https://openalex.org/S4306400024","display_name":"UCL Discovery (University College London)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I45129253","host_organization_name":"University College London","host_organization_lineage":["https://openalex.org/I45129253"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"   IEEE Transactions on Parallel and Distributed Systems , 33  (12)   pp. 3387-3399.   (2022)     (In press).  ","raw_type":"Article"},"sustainable_development_goals":[{"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7","score":0.8999999761581421}],"awards":[{"id":"https://openalex.org/G4587427570","display_name":null,"funder_award_id":"EP/S030069/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"},{"id":"https://openalex.org/G7493804148","display_name":null,"funder_award_id":"EP/N031768/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"},{"id":"https://openalex.org/G774180880","display_name":null,"funder_award_id":"EP/P010040/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"},{"id":"https://openalex.org/G8051675292","display_name":null,"funder_award_id":"EP/V028251/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"}],"funders":[{"id":"https://openalex.org/F4320334627","display_name":"Engineering and Physical Sciences Research Council","ror":"https://ror.org/0439y7842"}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4213433078.pdf"},"referenced_works_count":70,"referenced_works":["https://openalex.org/W582134693","https://openalex.org/W1494192115","https://openalex.org/W1921523184","https://openalex.org/W2095705004","https://openalex.org/W2119144962","https://openalex.org/W2167433878","https://openalex.org/W2194775991","https://openalex.org/W2335728318","https://openalex.org/W2626967530","https://openalex.org/W2794952988","https://openalex.org/W2928560789","https://openalex.org/W2951266961","https://openalex.org/W2953172087","https://openalex.org/W2962782102","https://openalex.org/W2963122961","https://openalex.org/W2963167203","https://openalex.org/W2963363373","https://openalex.org/W2963674932","https://openalex.org/W2964309882","https://openalex.org/W2979310060","https://openalex.org/W2979439447","https://openalex.org/W3000310304","https://openalex.org/W3016832937","https://openalex.org/W3035600444","https://openalex.org/W3035963492","https://openalex.org/W3043571714","https://openalex.org/W3047846843","https://openalex.org/W3049695724","https://openalex.org/W3082020764","https://openalex.org/W3083397657","https://openalex.org/W3098989436","https://openalex.org/W3102100346","https://openalex.org/W3103168911","https://openalex.org/W3105753905","https://openalex.org/W3106125969","https://openalex.org/W3117961507","https://openalex.org/W3118608800","https://openalex.org/W3129147932","https://openalex.org/W3130791942","https://openalex.org/W3164731060","https://openalex.org/W3166008066","https://openalex.org/W3187908937","https://openalex.org/W3211621070","https://openalex.org/W4236868170","https://openalex.org/W4251999575","https://openalex.org/W4287363917","https://openalex.org/W4295312788","https://openalex.org/W4300126339","https://openalex.org/W4301660609","https://openalex.org/W4308233353","https://openalex.org/W6617145748","https://openalex.org/W6638632666","https://openalex.org/W6640185926","https://openalex.org/W6674330103","https://openalex.org/W6675415620","https://openalex.org/W6677580257","https://openalex.org/W6684809622","https://openalex.org/W6703116779","https://openalex.org/W6729999211","https://openalex.org/W6739651123","https://openalex.org/W6748481559","https://openalex.org/W6764214684","https://openalex.org/W6765283766","https://openalex.org/W6766978945","https://openalex.org/W6772650050","https://openalex.org/W6782952729","https://openalex.org/W6787972765","https://openalex.org/W6790503700","https://openalex.org/W6790898931","https://openalex.org/W6803615721"],"related_works":["https://openalex.org/W2518118925","https://openalex.org/W2532502681","https://openalex.org/W3159273459","https://openalex.org/W4383503138","https://openalex.org/W2281932057","https://openalex.org/W2585970387","https://openalex.org/W3131592046","https://openalex.org/W2010977501","https://openalex.org/W2009779270","https://openalex.org/W3214999411"],"abstract_inverted_index":{"Bayesian":[0,232],"neural":[1],"networks":[2],"(BayesNNs)":[3],"have":[4],"demonstrated":[5,151,216],"their":[6,20,43,50],"advantages":[7],"in":[8,70,79],"various":[9],"safety-critical":[10],"applications,":[11],"such":[12],"as":[13],"autonomous":[14],"driving":[15],"or":[16,76],"healthcare,":[17],"due":[18],"to":[19,22,32,41,139,158,164,172,185,224,245,251],"ability":[21],"capture":[23],"and":[24,67,94,102,111,128,167,200,249],"represent":[25],"model":[26],"uncertainty.":[27],"However,":[28],"standard":[29,80],"BayesNNs":[30,87,220],"require":[31],"be":[33,106,156],"repeatedly":[34],"run":[35],"because":[36],"of":[37,86],"Monte":[38,91],"Carlo":[39,92],"sampling":[40,97],"quantify":[42],"uncertainty,":[44],"which":[45,104,188],"puts":[46],"a":[47,135,179,190],"burden":[48],"on":[49,211],"real-world":[51],"hardware":[52,112,177,181,192,198,209],"performance.":[53,149],"To":[54],"address":[55],"this":[56,59],"performance":[57,193],"issue,":[58],"article":[60],"systematically":[61],"exploits":[62],"the":[63,74,83,117,132,176,195,201,207,218,236],"extensive":[64],"structured":[65,77,84,153],"sparsity":[66,78,85,119,127,144,154],"redundant":[68],"computation":[69],"BayesNNs.":[71],"Different":[72],"from":[73],"unstructured":[75],"convolutional":[81],"NNs,":[82],"is":[88,137,183],"introduced":[89],"by":[90,162,170],"Dropout":[93],"its":[95],"associated":[96],"required":[98],"during":[99],"uncertainty":[100],"estimation":[101],"prediction,":[103],"can":[105,155,221],"exploited":[107,157],"through":[108],"both":[109],"algorithmic":[110,133,148],"optimizations.":[113],"We":[114,150],"first":[115],"classify":[116],"observed":[118],"patterns":[120],"into":[121],"three":[122,143],"categories:":[123],"channel":[124],"sparsity,":[125],"layer":[126],"sample":[129],"sparsity.":[130],"On":[131,175],"side,":[134,178],"framework":[136],"proposed":[138,184,208],"automatically":[140],"explore":[141],"these":[142],"categories":[145],"without":[146],"sacrificing":[147],"that":[152,217],"accelerate":[159,186],"CPU":[160],"designs":[161,169],"up":[163,171,223,244,250],"49":[165],"times,":[166],"GPU":[168,238],"40":[173],"times.":[174],"novel":[180],"architecture":[182],"BayesNNs,":[187],"achieves":[189],"high":[191],"using":[194],"runtime":[196],"adaptable":[197],"engines":[199],"intelligent":[202],"skipping":[203],"support.":[204],"Upon":[205],"implementing":[206],"design":[210,242],"an":[212],"FPGA,":[213],"our":[214,240],"experiments":[215],"algorithm-optimized":[219],"achieve":[222],"56":[225],"times":[226,247,253],"speedup":[227,248],"when":[228],"compared":[229],"with":[230,235],"unoptimized":[231],"nets.":[233],"Comparing":[234],"optimized":[237],"implementation,":[239],"FPGA":[241],"achieved":[243],"7.6":[246],"39.3":[252],"higher":[254],"energy":[255],"efficiency.":[256]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":3},{"year":2023,"cited_by_count":2}],"updated_date":"2026-04-05T17:49:38.594831","created_date":"2025-10-10T00:00:00"}