{"id":"https://openalex.org/W4415256962","doi":"https://doi.org/10.1109/les.2025.3600565","title":"Beyond BNNs: Design and Acceleration of Sub-Bit Neural Networks Using RISC-V Custom Functional Units","display_name":"Beyond BNNs: Design and Acceleration of Sub-Bit Neural Networks Using RISC-V Custom Functional Units","publication_year":2025,"publication_date":"2025-10-01","ids":{"openalex":"https://openalex.org/W4415256962","doi":"https://doi.org/10.1109/les.2025.3600565"},"language":null,"primary_location":{"id":"doi:10.1109/les.2025.3600565","is_oa":false,"landing_page_url":"https://doi.org/10.1109/les.2025.3600565","pdf_url":null,"source":{"id":"https://openalex.org/S22443479","display_name":"IEEE Embedded Systems Letters","issn_l":"1943-0663","issn":["1943-0663","1943-0671"],"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 Embedded Systems Letters","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/A5101507575","display_name":"Muhammad Sabih","orcid":"https://orcid.org/0000-0003-2066-646X"},"institutions":[{"id":"https://openalex.org/I181369854","display_name":"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg","ror":"https://ror.org/00f7hpc57","country_code":"DE","type":"education","lineage":["https://openalex.org/I181369854"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Muhammad Sabih","raw_affiliation_strings":["Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany"],"raw_orcid":"https://orcid.org/0000-0003-2066-646X","affiliations":[{"raw_affiliation_string":"Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany","institution_ids":["https://openalex.org/I181369854"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064320150","display_name":"Mohamed I. Abdo","orcid":null},"institutions":[{"id":"https://openalex.org/I181369854","display_name":"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg","ror":"https://ror.org/00f7hpc57","country_code":"DE","type":"education","lineage":["https://openalex.org/I181369854"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Mohamed Abdo","raw_affiliation_strings":["Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany"],"raw_orcid":"https://orcid.org/0009-0002-0701-2445","affiliations":[{"raw_affiliation_string":"Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany","institution_ids":["https://openalex.org/I181369854"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072039340","display_name":"Frank Hannig","orcid":"https://orcid.org/0000-0003-3663-6484"},"institutions":[{"id":"https://openalex.org/I181369854","display_name":"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg","ror":"https://ror.org/00f7hpc57","country_code":"DE","type":"education","lineage":["https://openalex.org/I181369854"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Frank Hannig","raw_affiliation_strings":["Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany"],"raw_orcid":"https://orcid.org/0000-0003-3663-6484","affiliations":[{"raw_affiliation_string":"Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany","institution_ids":["https://openalex.org/I181369854"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5076672029","display_name":"J\u00fcrgen Teich","orcid":"https://orcid.org/0000-0001-6285-5862"},"institutions":[{"id":"https://openalex.org/I181369854","display_name":"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg","ror":"https://ror.org/00f7hpc57","country_code":"DE","type":"education","lineage":["https://openalex.org/I181369854"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"J\u00fcrgen Teich","raw_affiliation_strings":["Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany"],"raw_orcid":"https://orcid.org/0000-0001-6285-5862","affiliations":[{"raw_affiliation_string":"Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany","institution_ids":["https://openalex.org/I181369854"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I181369854"],"apc_list":{"value":2645,"currency":"USD","value_usd":2645},"apc_paid":null,"fwci":1.1245,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.85183081,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":"17","issue":"5","first_page":"329","last_page":"332"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.9348000288009644,"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":0.9348000288009644,"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/codebook","display_name":"Codebook","score":0.7664999961853027},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.6105999946594238},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.59170001745224},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.5289999842643738},{"id":"https://openalex.org/keywords/hardware-acceleration","display_name":"Hardware acceleration","score":0.5167999863624573},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.5019000172615051},{"id":"https://openalex.org/keywords/binary-number","display_name":"Binary number","score":0.4814000129699707},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.45989999175071716}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8482000231742859},{"id":"https://openalex.org/C127759330","wikidata":"https://www.wikidata.org/wiki/Q637416","display_name":"Codebook","level":2,"score":0.7664999961853027},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.6105999946594238},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.59170001745224},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.5289999842643738},{"id":"https://openalex.org/C13164978","wikidata":"https://www.wikidata.org/wiki/Q600158","display_name":"Hardware acceleration","level":3,"score":0.5167999863624573},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.5019000172615051},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.4814000129699707},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.45989999175071716},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.4544999897480011},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.439300000667572},{"id":"https://openalex.org/C2908647359","wikidata":"https://www.wikidata.org/wiki/Q2625603","display_name":"Population","level":2,"score":0.42289999127388},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.40389999747276306},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.38100001215934753},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.3763999938964844},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.35830000042915344},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.34040001034736633},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.33399999141693115},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3001999855041504},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.2815999984741211},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.26249998807907104},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.2612000107765198},{"id":"https://openalex.org/C3017438360","wikidata":"https://www.wikidata.org/wiki/Q11188","display_name":"Global population","level":3,"score":0.25769999623298645},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2517000138759613}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/les.2025.3600565","is_oa":false,"landing_page_url":"https://doi.org/10.1109/les.2025.3600565","pdf_url":null,"source":{"id":"https://openalex.org/S22443479","display_name":"IEEE Embedded Systems Letters","issn_l":"1943-0663","issn":["1943-0663","1943-0671"],"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 Embedded Systems Letters","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G7348931839","display_name":"Optimierung neuronaler approximativer Beschleunigerarchitekturen f\u00fcr DNN Inferenz auf leichtgewichtigen FPGAs (NA^3Os)","funder_award_id":"524986327","funder_id":"https://openalex.org/F4320320879","funder_display_name":"Deutsche Forschungsgemeinschaft"}],"funders":[{"id":"https://openalex.org/F4320320879","display_name":"Deutsche Forschungsgemeinschaft","ror":"https://ror.org/018mejw64"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":7,"referenced_works":["https://openalex.org/W220440441","https://openalex.org/W2300242332","https://openalex.org/W3135547017","https://openalex.org/W3207102667","https://openalex.org/W4401568359","https://openalex.org/W4403582494","https://openalex.org/W4413278534"],"related_works":[],"abstract_inverted_index":{"Binary":[0],"neural":[1],"networks":[2],"(BNNs)":[3],"are":[4,20],"known":[5],"for":[6,16,133],"their":[7],"minimal":[8],"memory":[9],"requirements,":[10],"making":[11],"them":[12,33],"an":[13,95],"attractive":[14],"choice":[15],"resource-constrained":[17],"environments.":[18],"SBNN-nps":[19],"a":[21,56,63,81,91,109,131],"more":[22],"recent":[23],"advancement":[24],"that":[25,67],"extend":[26],"the":[27,121],"benefits":[28],"of":[29,111],"BNNs":[30,48,61,118],"by":[31,100],"compressing":[32,45],"even":[34,135],"further,":[35],"achieving":[36],"sub-bit":[37],"level":[38],"representations":[39],"to":[40,59,86,113,116,139],"maximize":[41],"efficiency.":[42],"However,":[43],"effectively":[44],"and":[46,102,129],"accelerating":[47],"presents":[49],"challenges.":[50],"In":[51],"this":[52],"letter,":[53],"we":[54],"propose":[55,76],"novel":[57],"approach":[58,107],"compress":[60],"using":[62],"fixed-length":[64],"compression":[65,137],"scheme":[66],"can":[68],"be":[69],"efficiently":[70],"decoded":[71],"at":[72],"runtime.":[73],"We":[74],"then":[75],"RISC-V":[77,122],"extensions,":[78],"implemented":[79],"as":[80],"custom":[82],"function":[83],"unit":[84],"(CFU),":[85],"decode":[87],"compressed":[88],"weights":[89],"via":[90],"codebook":[92],"stored":[93],"on":[94,120],"FPGA":[96],"on-board":[97],"memory,":[98],"followed":[99],"XOR":[101],"population":[103],"count":[104],"operations.":[105],"This":[106],"achieves":[108],"speedup":[110],"up":[112],"2\u00d7":[114],"compared":[115],"conventional":[117],"deployed":[119],"softcore,":[123],"with":[124],"Significantly":[125],"less":[126],"accuracy":[127],"degradation,":[128],"provides":[130],"foundation":[132],"exploring":[134],"higher":[136],"configurations":[138],"improve":[140],"performance":[141],"further.":[142]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-08-21T09:56:20.448147","created_date":"2025-10-17T00:00:00"}