{"id":"https://openalex.org/W7162105034","doi":"https://doi.org/10.1109/access.2026.3696226","title":"Hardware-Friendly Fully Quantized Mamba-2 Model and Its FPGA-Based Accelerator","display_name":"Hardware-Friendly Fully Quantized Mamba-2 Model and Its FPGA-Based Accelerator","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7162105034","doi":"https://doi.org/10.1109/access.2026.3696226"},"language":"en","primary_location":{"id":"doi:10.1109/access.2026.3696226","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3696226","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"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 Access","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1109/access.2026.3696226","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5004321250","display_name":"Hiroshi Fuketa","orcid":"https://orcid.org/0000-0003-0171-6679"},"institutions":[{"id":"https://openalex.org/I73613424","display_name":"National Institute of Advanced Industrial Science and Technology","ror":"https://ror.org/01703db54","country_code":"JP","type":"government","lineage":["https://openalex.org/I73613424"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Hiroshi Fuketa","raw_affiliation_strings":["National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan"],"raw_orcid":"https://orcid.org/0000-0003-0171-6679","affiliations":[{"raw_affiliation_string":"National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan","institution_ids":["https://openalex.org/I73613424"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5049602728","display_name":"Masakazu Hioki","orcid":"https://orcid.org/0000-0002-5641-9252"},"institutions":[{"id":"https://openalex.org/I73613424","display_name":"National Institute of Advanced Industrial Science and Technology","ror":"https://ror.org/01703db54","country_code":"JP","type":"government","lineage":["https://openalex.org/I73613424"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Masakazu Hioki","raw_affiliation_strings":["National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan"],"raw_orcid":"https://orcid.org/0000-0002-5641-9252","affiliations":[{"raw_affiliation_string":"National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan","institution_ids":["https://openalex.org/I73613424"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I73613424"],"apc_list":{"value":1850,"currency":"USD","value_usd":1850},"apc_paid":{"value":1850,"currency":"USD","value_usd":1850},"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.6641329,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"14","issue":null,"first_page":"78920","last_page":"78930"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11697","display_name":"Numerical Methods and Algorithms","score":0.14800000190734863,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/T11697","display_name":"Numerical Methods and Algorithms","score":0.14800000190734863,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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":0.1363999992609024,"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/T10951","display_name":"Cryptographic Implementations and Security","score":0.056299999356269836,"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/acceleration","display_name":"Acceleration","score":0.3564000129699707},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.3005000054836273},{"id":"https://openalex.org/keywords/signal-processing","display_name":"Signal processing","score":0.2526000142097473},{"id":"https://openalex.org/keywords/particle-accelerator","display_name":"Particle accelerator","score":0.24819999933242798},{"id":"https://openalex.org/keywords/probability-density-function","display_name":"Probability density function","score":0.22920000553131104}],"concepts":[{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.4927000105381012},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4584999978542328},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.3564000129699707},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.3005000054836273},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.2526000142097473},{"id":"https://openalex.org/C189166818","wikidata":"https://www.wikidata.org/wiki/Q130825","display_name":"Particle accelerator","level":3,"score":0.24819999933242798},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.24420000612735748},{"id":"https://openalex.org/C197055811","wikidata":"https://www.wikidata.org/wiki/Q207522","display_name":"Probability density function","level":2,"score":0.22920000553131104},{"id":"https://openalex.org/C3079626","wikidata":"https://www.wikidata.org/wiki/Q234881","display_name":"Quantum electrodynamics","level":1,"score":0.21649999916553497},{"id":"https://openalex.org/C2986605239","wikidata":"https://www.wikidata.org/wiki/Q925667","display_name":"Numerical models","level":3,"score":0.2152000069618225}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/access.2026.3696226","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3696226","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"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 Access","raw_type":"journal-article"},{"id":"pmh:oai:doaj.org/article:3df5db79adde4038999afa0fc3de8a05","is_oa":true,"landing_page_url":"https://doaj.org/article/3df5db79adde4038999afa0fc3de8a05","pdf_url":null,"source":{"id":"https://openalex.org/S4306401280","display_name":"DOAJ (DOAJ: Directory of Open Access Journals)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"IEEE Access, Vol 14, Pp 78920-78930 (2026)","raw_type":"article"}],"best_oa_location":{"id":"doi:10.1109/access.2026.3696226","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3696226","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"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 Access","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G531153412","display_name":null,"funder_award_id":"JP24K02918","funder_id":"https://openalex.org/F4320334764","funder_display_name":"Japan Society for the Promotion of Science"}],"funders":[{"id":"https://openalex.org/F4320334764","display_name":"Japan Society for the Promotion of Science","ror":"https://ror.org/00hhkn466"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"In":[0,40],"this":[1],"paper,":[2],"we":[3,114],"propose":[4,115],"a":[5,22,28,140,146,181,190],"fully":[6,123],"quantized":[7,49,124],"Mamba-2":[8,125],"model":[9,15,105],"to":[10,31,50,120,203],"enable":[11],"efficient":[12],"large":[13],"language":[14],"(LLM)":[16],"processing":[17],"for":[18,37,61],"edge":[19,38],"AI.":[20,39],"As":[21],"promising":[23],"next\u2011generation":[24],"architecture,":[25],"Mamba":[26,72],"offers":[27],"scalable":[29],"alternative":[30],"Transformer-based":[32],"models":[33],"with":[34,84,197],"better":[35],"suitability":[36],"the":[41,44,59,65,70,108,121,154,163,166,177,185,204],"proposed":[42,109,122,128,178],"model,":[43,73],"weights":[45],"and":[46,52,79,91,99,136,145,159,172,189],"activations":[47],"are":[48,82],"ternary":[51],"8-bit":[53],"integer":[54],"(INT8)":[55],"precision,":[56],"respectively,":[57],"eliminating":[58],"need":[60],"floating-point":[62],"operations.":[63],"Additionally,":[64],"non-linear":[66],"functions":[67],"included":[68],"in":[69,184,193],"original":[71],"such":[74],"as":[75],"SiLU,":[76],"softplus,":[77],"exponential,":[78],"normalization":[80],"functions,":[81,90],"replaced":[83],"hardware-friendly":[85],"alternatives,":[86],"including":[87],"ReLU,":[88],"power-of-two":[89],"table-lookup-based":[92],"normalization.":[93],"These":[94],"substitutions":[95],"substantially":[96],"reduce":[97],"computational":[98],"hardware":[100],"complexity":[101],"while":[102],"largely":[103],"preserving":[104],"accuracy":[106],"through":[107],"two-stage":[110],"quantization-aware":[111],"training.":[112],"Furthermore,":[113],"an":[116,133],"FPGA-based":[117,199],"accelerator":[118,129,179],"tailored":[119],"model.":[126],"The":[127,151],"is":[130],"implemented":[131],"on":[132],"Ultra96-V2":[134],"board":[135],"consists":[137],"primarily":[138],"of":[139,156,168],"LUT-based":[141],"vector-matrix":[142],"multiplication":[143,149],"unit":[144],"DSP-based":[147],"vector-vector":[148],"unit.":[150],"former":[152],"computes":[153,165],"products":[155,167],"INT8-precision":[157,169],"vectors":[158],"ternary-precision":[160],"matrices,":[161],"whereas":[162],"latter":[164],"vectors.":[170],"Implementation":[171],"measurement":[173],"results":[174],"show":[175],"that":[176],"achieves":[180],"17%":[182],"reduction":[183],"FPGA":[186],"resource":[187],"cost":[188],"3.6-fold":[191],"improvement":[192],"energy":[194],"efficiency":[195],"compared":[196],"conventional":[198],"LLM":[200],"accelerators,":[201],"owing":[202],"hardware\u2011friendly":[205],"design.":[206]},"counts_by_year":[],"updated_date":"2026-05-30T06:14:24.967023","created_date":"2026-05-23T00:00:00"}