{"id":"https://openalex.org/W7130601595","doi":"https://doi.org/10.1109/tsusc.2026.3666456","title":"BEExformer: A Fast Inferencing Binarized Transformer With Early Exits","display_name":"BEExformer: A Fast Inferencing Binarized Transformer With Early Exits","publication_year":2026,"publication_date":"2026-02-19","ids":{"openalex":"https://openalex.org/W7130601595","doi":"https://doi.org/10.1109/tsusc.2026.3666456"},"language":null,"primary_location":{"id":"doi:10.1109/tsusc.2026.3666456","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tsusc.2026.3666456","pdf_url":null,"source":{"id":"https://openalex.org/S4210221417","display_name":"IEEE Transactions on Sustainable Computing","issn_l":"2377-3782","issn":["2377-3782","2377-3790"],"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 Sustainable Computing","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/A5071155612","display_name":"Wazib Ansar","orcid":"https://orcid.org/0000-0001-9191-1771"},"institutions":[{"id":"https://openalex.org/I106542073","display_name":"University of Calcutta","ror":"https://ror.org/01e7v7w47","country_code":"IN","type":"education","lineage":["https://openalex.org/I106542073"]}],"countries":["IN"],"is_corresponding":true,"raw_author_name":"Wazib Ansar","raw_affiliation_strings":["A. K. Choudhury School of IT, University of Calcutta, Kolkata, West Bengal, India"],"raw_orcid":"https://orcid.org/0000-0001-9191-1771","affiliations":[{"raw_affiliation_string":"A. K. Choudhury School of IT, University of Calcutta, Kolkata, West Bengal, India","institution_ids":["https://openalex.org/I106542073"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112442569","display_name":"Saptarsi Goswami","orcid":null},"institutions":[{"id":"https://openalex.org/I2800966917","display_name":"Bangur Institute of Neurosciences","ror":"https://ror.org/031yk2194","country_code":"IN","type":"healthcare","lineage":["https://openalex.org/I2800966917"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Saptarsi Goswami","raw_affiliation_strings":["Department of Computer Science, Bangabasi Morning College, Kolkata, West Bengal, India"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Bangabasi Morning College, Kolkata, West Bengal, India","institution_ids":["https://openalex.org/I2800966917"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5074619222","display_name":"A. Chakrabarti","orcid":null},"institutions":[{"id":"https://openalex.org/I106542073","display_name":"University of Calcutta","ror":"https://ror.org/01e7v7w47","country_code":"IN","type":"education","lineage":["https://openalex.org/I106542073"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Amlan Chakrabarti","raw_affiliation_strings":["A. K. Choudhury School of IT, University of Calcutta, Kolkata, West Bengal, India"],"raw_orcid":"https://orcid.org/0000-0003-4380-3172","affiliations":[{"raw_affiliation_string":"A. K. Choudhury School of IT, University of Calcutta, Kolkata, West Bengal, India","institution_ids":["https://openalex.org/I106542073"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5071155612"],"corresponding_institution_ids":["https://openalex.org/I106542073"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.34929528,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"11","issue":"2","first_page":"98","last_page":"110"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.3346000015735626,"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/T10028","display_name":"Topic Modeling","score":0.3346000015735626,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.0869000032544136,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.0731000006198883,"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/transformer","display_name":"Transformer","score":0.7253999710083008},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5967000126838684},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.53329998254776},{"id":"https://openalex.org/keywords/differentiable-function","display_name":"Differentiable function","score":0.42419999837875366},{"id":"https://openalex.org/keywords/entropy","display_name":"Entropy (arrow of time)","score":0.4077000021934509},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.33180001378059387}],"concepts":[{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.7253999710083008},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6462000012397766},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5967000126838684},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.53329998254776},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.47040000557899475},{"id":"https://openalex.org/C202615002","wikidata":"https://www.wikidata.org/wiki/Q783507","display_name":"Differentiable function","level":2,"score":0.42419999837875366},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.4077000021934509},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4023999869823456},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.33180001378059387},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.27160000801086426},{"id":"https://openalex.org/C115680565","wikidata":"https://www.wikidata.org/wiki/Q5977448","display_name":"Gradient method","level":2,"score":0.26820001006126404},{"id":"https://openalex.org/C167981619","wikidata":"https://www.wikidata.org/wiki/Q1685498","display_name":"Cross entropy","level":3,"score":0.2549999952316284},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.2531000077724457}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tsusc.2026.3666456","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tsusc.2026.3666456","pdf_url":null,"source":{"id":"https://openalex.org/S4210221417","display_name":"IEEE Transactions on Sustainable Computing","issn_l":"2377-3782","issn":["2377-3782","2377-3790"],"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 Sustainable Computing","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"based":[4],"on":[5,10,24,59,151],"transformers":[6],"achieve":[7],"cutting-edge":[8],"results":[9],"a":[11,79,117],"variety":[12],"of":[13,135],"applications.":[14],"However,":[15,41],"their":[16],"enormous":[17],"size":[18],"and":[19,31,54,92,133,172,195],"processing":[20],"requirements":[21],"hinder":[22],"deployment":[23],"constrained":[25],"resources.":[26],"To":[27,68],"enhance":[28,107],"efficiency,":[29],"binarization":[30,42],"Early":[32,75],"Exit":[33,76],"(EE)":[34],"have":[35],"proved":[36],"to":[37,45,106,121],"be":[38],"effective":[39],"solutions.":[40],"may":[43],"lead":[44],"performance":[46],"loss":[47,162],"as":[48],"reduced":[49],"precision":[50],"affects":[51],"gradient":[52,126],"estimation":[53],"parameter":[55],"updates.":[56],"Besides,":[57],"research":[58],"EE":[60,89,148],"mechanisms":[61],"is":[62],"still":[63],"in":[64,140,144,154,184],"its":[65,206],"early":[66],"stages.":[67],"address":[69],"these":[70],"challenges,":[71],"we":[72],"introduce":[73],"Binarized":[74],"Transformer":[77],"(BEExformer),":[78],"first-of-its-kind":[80],"selective":[81],"learning-based":[82],"transformer":[83,97,158],"integrating":[84],"Binarization-Aware":[85],"Training":[86],"(BAT)":[87],"with":[88,160,191],"for":[90],"efficient":[91],"fast":[93],"textual":[94],"inference.":[95],"Each":[96],"block":[98],"has":[99],"an":[100],"integrated":[101],"Selective-Learn":[102],"Forget":[103],"Network":[104],"(SLFN)":[105],"contextual":[108],"retention":[109],"while":[110],"eliminating":[111],"irrelevant":[112],"information.":[113],"The":[114,147],"BAT":[115],"employs":[116],"differentiable":[118],"second-order":[119],"approximation":[120],"the":[122,131,136,180,192],"sign":[123,132],"function,":[124],"enabling":[125],"computation":[127],"that":[128],"captures":[129],"both":[130],"magnitude":[134],"weights.":[137],"This":[138,164],"aids":[139],"21.30":[141],"\u00d7":[142],"reduction":[143,153],"model":[145],"size.":[146],"mechanism":[149],"hinges":[150],"fractional":[152],"entropy":[155],"among":[156],"intermediate":[157],"blocks":[159],"soft-routing":[161],"estimation.":[163],"accelerates":[165],"inference":[166],"by":[167,170,176,178],"reducing":[168],"FLOPs":[169],"52.27%":[171],"even":[173],"improves":[174],"accuracy":[175],"3.22%":[177],"resolving":[179],"\u201coverthinking\u201d":[181],"problem":[182],"inherent":[183],"deep":[185],"networks.":[186],"Extensive":[187],"evaluation":[188],"through":[189],"comparison":[190],"SOTA":[193],"methods":[194],"various":[196],"ablations":[197],"across":[198],"nine":[199],"datasets":[200],"covering":[201],"multiple":[202],"NLP":[203],"tasks":[204],"demonstrates":[205],"Pareto-optimal":[207],"performance-efficiency":[208],"trade-off.":[209]},"counts_by_year":[],"updated_date":"2026-04-04T06:10:10.580331","created_date":"2026-02-20T00:00:00"}