{"id":"https://openalex.org/W4406092772","doi":"https://doi.org/10.1145/3706628.3708877","title":"TreeLUT: An Efficient Alternative to Deep Neural Networks for Inference Acceleration Using Gradient Boosted Decision Trees","display_name":"TreeLUT: An Efficient Alternative to Deep Neural Networks for Inference Acceleration Using Gradient Boosted Decision Trees","publication_year":2025,"publication_date":"2025-02-26","ids":{"openalex":"https://openalex.org/W4406092772","doi":"https://doi.org/10.1145/3706628.3708877"},"language":"en","primary_location":{"id":"doi:10.1145/3706628.3708877","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3706628.3708877","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 ACM/SIGDA International Symposium on Field Programmable Gate Arrays","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2501.01511","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5065614453","display_name":"Alireza Khataei","orcid":"https://orcid.org/0000-0002-9146-5684"},"institutions":[{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Alireza Khataei","raw_affiliation_strings":["Department of Electrical and Computer Engineering, University of Minnesota, MN, MN, USA","Department of Electrical and Computer Engineering University of Minnesota Minneapolis, MN, USA"],"raw_orcid":"https://orcid.org/0000-0002-9146-5684","affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering, University of Minnesota, MN, MN, USA","institution_ids":["https://openalex.org/I130238516"]},{"raw_affiliation_string":"Department of Electrical and Computer Engineering University of Minnesota Minneapolis, MN, USA","institution_ids":["https://openalex.org/I130238516"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5013257804","display_name":"Kia Bazargan","orcid":"https://orcid.org/0000-0003-3624-7366"},"institutions":[{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kia Bazargan","raw_affiliation_strings":["Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, USA"],"raw_orcid":"https://orcid.org/0000-0003-3624-7366","affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, USA","institution_ids":["https://openalex.org/I130238516"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5065614453"],"corresponding_institution_ids":["https://openalex.org/I130238516"],"apc_list":null,"apc_paid":null,"fwci":29.0378,"has_fulltext":true,"cited_by_count":21,"citation_normalized_percentile":{"value":0.99788751,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":99,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"14","last_page":"24"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11512","display_name":"Anomaly Detection Techniques and Applications","score":0.9976000189781189,"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/T11512","display_name":"Anomaly Detection Techniques and Applications","score":0.9976000189781189,"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/T10320","display_name":"Neural Networks and Applications","score":0.9884999990463257,"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/T12205","display_name":"Time Series Analysis and Forecasting","score":0.9851999878883362,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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":0.6903022527694702},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6588508486747742},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6507819890975952},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.6169434785842896},{"id":"https://openalex.org/keywords/decision-tree","display_name":"Decision tree","score":0.5397243499755859},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5356549620628357},{"id":"https://openalex.org/keywords/deep-neural-networks","display_name":"Deep neural networks","score":0.5226154923439026},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.4412483274936676},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.10011857748031616}],"concepts":[{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.6903022527694702},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6588508486747742},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6507819890975952},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6169434785842896},{"id":"https://openalex.org/C84525736","wikidata":"https://www.wikidata.org/wiki/Q831366","display_name":"Decision tree","level":2,"score":0.5397243499755859},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5356549620628357},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.5226154923439026},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4412483274936676},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.10011857748031616},{"id":"https://openalex.org/C74650414","wikidata":"https://www.wikidata.org/wiki/Q11397","display_name":"Classical mechanics","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3706628.3708877","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3706628.3708877","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 ACM/SIGDA International Symposium on Field Programmable Gate Arrays","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2501.01511","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.01511","pdf_url":"https://arxiv.org/pdf/2501.01511","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2501.01511","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2501.01511","pdf_url":"https://arxiv.org/pdf/2501.01511","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3854537970","display_name":null,"funder_award_id":"PFI-TT 2016390","funder_id":"https://openalex.org/F4320323817","funder_display_name":"Universitas Brawijaya"},{"id":"https://openalex.org/G6725241808","display_name":"PFI-TT: Harnessing the power of uncompressed number representation for modern computations","funder_award_id":"2016390","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320309636","display_name":"University of Minnesota","ror":"https://ror.org/03grvy078"},{"id":"https://openalex.org/F4320323817","display_name":"Universitas Brawijaya","ror":"https://ror.org/01wk3d929"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4406092772.pdf","grobid_xml":"https://content.openalex.org/works/W4406092772.grobid-xml"},"referenced_works_count":18,"referenced_works":["https://openalex.org/W1678356000","https://openalex.org/W2198739892","https://openalex.org/W2295598076","https://openalex.org/W2408172093","https://openalex.org/W2565125333","https://openalex.org/W2787398007","https://openalex.org/W2907163239","https://openalex.org/W2951537853","https://openalex.org/W3093982621","https://openalex.org/W3126460454","https://openalex.org/W3183144345","https://openalex.org/W4281689464","https://openalex.org/W4312692239","https://openalex.org/W4391455313","https://openalex.org/W4401935097","https://openalex.org/W4402196955","https://openalex.org/W4403278560","https://openalex.org/W4403278591"],"related_works":["https://openalex.org/W2565094479","https://openalex.org/W2390829436","https://openalex.org/W1989791859","https://openalex.org/W602859758","https://openalex.org/W1971289376","https://openalex.org/W2379101322","https://openalex.org/W2055243143","https://openalex.org/W1992553864","https://openalex.org/W3042419602","https://openalex.org/W2966649771"],"abstract_inverted_index":{"Accelerating":[0],"machine":[1],"learning":[2],"inference":[3],"has":[4],"been":[5],"an":[6,157,164,254],"active":[7],"research":[8],"area":[9,203],"in":[10,28,95,185,273],"recent":[11,279],"years.":[12],"In":[13,138],"this":[14,60],"context,":[15],"field-programmable":[16],"gate":[17],"arrays":[18],"(FPGAs)":[19],"have":[20,57],"demonstrated":[21],"compelling":[22],"performance":[23],"by":[24],"providing":[25],"massive":[26,44],"parallelism":[27],"deep":[29],"neural":[30],"networks":[31,34],"(DNNs).":[32],"Neural":[33],"(NNs)":[35],"are":[36,89,129],"computationally":[37],"intensive":[38],"during":[39,147],"inference,":[40,148],"as":[41,156,222],"they":[42],"require":[43],"amounts":[45],"of":[46,68,74,99,112,122,192,256,275],"multiplication":[47],"and":[48,72,101,151,170,204,218,229,241],"addition,":[49],"which":[50,118],"makes":[51],"their":[52],"implementations":[53],"costly.":[54],"Numerous":[55],"studies":[56],"recently":[58],"addressed":[59],"challenge":[61],"to":[62,93,132,140,200,247,267],"some":[63],"extent":[64],"using":[65,163,194],"a":[66,90,96,115,135],"combination":[67],"sparsity":[69],"induction,":[70],"quantization,":[71],"transformation":[73],"neurons":[75],"or":[76,180],"sub-networks":[77],"into":[78],"lookup":[79],"tables":[80],"(LUTs)":[81],"on":[82,182,259],"FPGAs.":[83],"Gradient":[84],"boosted":[85],"decision":[86,116,124,141],"trees":[87],"(GBDTs)":[88],"high-accuracy":[91],"alternative":[92],"DNNs":[94],"wide":[97],"range":[98],"regression":[100],"classification":[102,196],"tasks,":[103],"particularly":[104],"for":[105,160],"tabular":[106],"datasets.":[107],"The":[108],"basic":[109],"building":[110],"block":[111],"GBDTs":[113,143,162],"is":[114],"tree,":[117],"resembles":[119],"the":[120,190,260],"structure":[121,136],"binary":[123],"diagrams.":[125],"FPGA":[126],"design":[127],"flows":[128],"heavily":[130],"optimized":[131],"implement":[133],"such":[134,221],"efficiently.":[137],"addition":[139],"trees,":[142],"perform":[144],"simple":[145],"operations":[146],"including":[149],"comparison":[150],"addition.":[152],"We":[153,188],"present":[154],"TreeLUT":[155,193,235],"open-source":[158],"tool":[159],"implementing":[161],"efficient":[165],"quantization":[166],"scheme,":[167],"hardware":[168,238,271],"architecture,":[169],"pipelining":[171],"strategy.":[172],"It":[173],"primarily":[174],"utilizes":[175],"LUTs":[176],"with":[177,215],"no":[178],"BRAMs":[179],"DSPs":[181],"FPGAs,":[183],"resulting":[184],"high":[186],"efficiency.":[187],"show":[189,233],"effectiveness":[191],"multiple":[195],"datasets,":[197],"commonly":[198],"used":[199],"evaluate":[201],"ultra-low":[202],"latency":[205],"architectures.":[206],"Using":[207],"these":[208],"benchmarks,":[209],"we":[210],"compare":[211],"our":[212],"implementation":[213],"results":[214,232],"existing":[216],"DNN":[217],"GBDT":[219],"methods,":[220],"DWN,":[223],"PolyLUT-Add,":[224],"NeuraLUT,":[225],"LogicNets,":[226],"FINN,":[227],"hls4ml,":[228],"others.":[230],"Our":[231],"that":[234],"significantly":[236],"improves":[237],"utilization,":[239],"latency,":[240],"throughput":[242],"at":[243],"competitive":[244],"accuracy":[245,255],"compared":[246],"previous":[248],"works.":[249],"For":[250],"instance,":[251],"it":[252],"achieves":[253],"around":[257,265],"97%":[258],"MNIST":[261],"dataset":[262],"while":[263],"delivering":[264],"4":[266],"101":[268],"times":[269],"lower":[270],"cost":[272],"terms":[274],"area-delay":[276],"product":[277],"than":[278],"LUT-based":[280],"NNs.":[281]},"counts_by_year":[{"year":2026,"cited_by_count":12},{"year":2025,"cited_by_count":9}],"updated_date":"2026-09-01T08:31:24.195828","created_date":"2025-10-10T00:00:00"}