{"id":"https://openalex.org/W4404352974","doi":"https://doi.org/10.48550/arxiv.2411.01818","title":"Shrinking the Giant : Quasi-Weightless Transformers for Low Energy Inference","display_name":"Shrinking the Giant : Quasi-Weightless Transformers for Low Energy Inference","publication_year":2024,"publication_date":"2024-11-04","ids":{"openalex":"https://openalex.org/W4404352974","doi":"https://doi.org/10.48550/arxiv.2411.01818"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2411.01818","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.01818","pdf_url":"https://arxiv.org/pdf/2411.01818","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,"listed_in":[],"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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2411.01818","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5079937406","display_name":"Shashank Nag","orcid":"https://orcid.org/0009-0000-3619-6066","observed_orcids":["https://orcid.org/0009-0000-3619-6066"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nag, Shashank","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066894791","display_name":"Alan T. L. Bacellar","orcid":"https://orcid.org/0000-0002-3346-7665","observed_orcids":["https://orcid.org/0000-0002-3346-7665"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bacellar, Alan T. L.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5035725580","display_name":"Zachary Susskind","orcid":"https://orcid.org/0000-0002-7244-6285","observed_orcids":["https://orcid.org/0000-0002-7244-6285"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Susskind, Zachary","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010259580","display_name":"Anshul Jha","orcid":"https://orcid.org/0000-0002-5491-1884","observed_orcids":["https://orcid.org/0000-0002-5491-1884"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jha, Anshul","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114637645","display_name":"Logan Liberty","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liberty, Logan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Sivakumar, Aishwarya","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sivakumar, Aishwarya","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010151984","display_name":"Eugene B John","orcid":"https://orcid.org/0000-0001-9494-4894","observed_orcids":["https://orcid.org/0000-0001-9494-4894"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"John, Eugene B.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051377329","display_name":"Krishnan K. Kailas","orcid":"https://orcid.org/0000-0002-9935-9827","observed_orcids":["https://orcid.org/0000-0002-9935-9827"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kailas, Krishnan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058318838","display_name":"Priscila M. V. Lima","orcid":"https://orcid.org/0000-0002-8515-9904","observed_orcids":["https://orcid.org/0000-0002-8515-9904"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lima, Priscila M. V.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031510918","display_name":"Neeraja J. Yadwadkar","orcid":"https://orcid.org/0009-0007-7556-3069","observed_orcids":["https://orcid.org/0009-0007-7556-3069"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yadwadkar, Neeraja J.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033447808","display_name":"Felipe M. G. Fran\u00e7a","orcid":"https://orcid.org/0000-0002-8980-6208","observed_orcids":["https://orcid.org/0000-0002-8980-6208"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Franca, Felipe M. G.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5068885069","display_name":"Lizy K. John","orcid":"https://orcid.org/0000-0002-8747-5214","observed_orcids":["https://orcid.org/0000-0002-8747-5214"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"John, Lizy K.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9829000234603882,"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":0.9829000234603882,"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/T10363","display_name":"Low-power high-performance VLSI design","score":0.006300000008195639,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.0017999999690800905,"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/weightless-neural-networks","display_name":"weightless neural networks","score":0.8349999785423279},{"id":"https://openalex.org/keywords/energy-efficient-inference","display_name":"energy-efficient inference","score":0.875},{"id":"https://openalex.org/keywords/transformer-compression","display_name":"transformer compression","score":0.6029999852180481},{"id":"https://openalex.org/keywords/energy-efficiency","display_name":"energy efficiency","score":0.3930000066757202}],"concepts":[{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6743735074996948},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6230219006538391},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.3866485357284546},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.24383297562599182},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.20685040950775146},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.17503800988197327},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.07075533270835876}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2411.01818","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.01818","pdf_url":"https://arxiv.org/pdf/2411.01818","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,"listed_in":[],"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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2411.01818","doi":"https://doi.org/10.48550/arxiv.2411.01818","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2411.01818","pdf_url":null,"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,"listed_in":[],"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2411.01818","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.01818","pdf_url":"https://arxiv.org/pdf/2411.01818","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,"listed_in":[],"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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G6138234132","display_name":"Collaborative Research: SHF: Small: Quasi Weightless Neural Networks for Energy-Efficient Machine Learning on the Edge","funder_award_id":"2326894","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G653561462","display_name":null,"funder_award_id":"3148.001","funder_id":"https://openalex.org/F4320306087","funder_display_name":"Semiconductor Research Corporation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320306087","display_name":"Semiconductor Research Corporation","ror":"https://ror.org/047z4n946"},{"id":"https://openalex.org/F4320309480","display_name":"Nvidia","ror":"https://ror.org/03jdj4y14"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4404352974.pdf","grobid_xml":"https://content.openalex.org/works/W4404352974.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4383982669","https://openalex.org/W4404648795","https://openalex.org/W7154988726","https://openalex.org/W4286571858","https://openalex.org/W7131307777","https://openalex.org/W3117477522","https://openalex.org/W4319995060","https://openalex.org/W7171631283","https://openalex.org/W7215124850","https://openalex.org/W4200632953"],"abstract_inverted_index":{"Transformers":[0,119],"are":[1,57],"set":[2],"to":[3,14,42,115],"become":[4],"ubiquitous":[5],"with":[6,35,113,170],"applications":[7],"ranging":[8],"from":[9],"chatbots":[10],"and":[11,17,24,37,110,126,157],"educational":[12],"assistants":[13],"visual":[15],"recognition":[16],"remote":[18],"sensing.":[19],"However,":[20],"their":[21,65],"increasing":[22],"computational":[23,125],"memory":[25],"demands":[26],"is":[27,40],"resulting":[28],"in":[29,107,153],"growing":[30],"energy":[31,161],"consumption.":[32],"Building":[33],"models":[34,109],"fast":[36],"energy-efficient":[38,127],"inference":[39,66,128],"imperative":[41],"enable":[43],"a":[44,69,124,137,159],"variety":[45],"of":[46,100,140,149],"transformer-based":[47,131],"applications.":[48],"Look":[49],"Up":[50],"Table":[51],"(LUT)":[52],"based":[53],"Weightless":[54,118],"Neural":[55],"Networks":[56],"faster":[58],"than":[59],"the":[60,98,101,151,154,171],"conventional":[61],"neural":[62],"networks":[63,79],"as":[64],"only":[67],"involves":[68],"few":[70],"lookup":[71],"operations.":[72],"Recently,":[73],"an":[74,82],"approach":[75],"for":[76,96,123,130],"learning":[77],"LUT":[78],"directly":[80],"via":[81],"Extended":[83],"Finite":[84],"Difference":[85],"method":[86],"was":[87],"proposed.":[88],"We":[89,163],"build":[90],"on":[91,142,168],"this":[92],"idea,":[93],"extending":[94],"it":[95],"performing":[97],"functions":[99],"Multi":[102],"Layer":[103],"Perceptron":[104],"(MLP)":[105],"layers":[106],"transformer":[108],"integrating":[111],"them":[112],"transformers":[114],"propose":[116],"Quasi":[117],"(QuWeiT).":[120],"This":[121],"allows":[122],"solution":[129],"models.":[132],"On":[133],"I-ViT-T,":[134],"we":[135],"achieve":[136],"comparable":[138],"accuracy":[139],"95.64%":[141],"CIFAR-10":[143],"dataset":[144],"while":[145],"replacing":[146],"approximately":[147],"55%":[148],"all":[150],"multiplications":[152],"entire":[155],"model":[156],"achieving":[158],"2.2x":[160],"efficiency.":[162],"also":[164],"observe":[165],"similar":[166],"savings":[167],"experiments":[169],"nanoGPT":[172],"framework.":[173]},"counts_by_year":[],"updated_date":"2026-10-14T16:39:01.820549","created_date":"2024-11-14T00:00:00"}