{"id":"https://openalex.org/W4412944498","doi":"https://doi.org/10.18653/v1/2025.findings-acl.1346","title":"Fuzzy Speculative Decoding for a Tunable Accuracy-Runtime Tradeoff","display_name":"Fuzzy Speculative Decoding for a Tunable Accuracy-Runtime Tradeoff","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412944498","doi":"https://doi.org/10.18653/v1/2025.findings-acl.1346"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.1346","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.1346","pdf_url":"https://aclanthology.org/2025.findings-acl.1346.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-acl.1346.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5119204268","display_name":"Maximilian Holsman","orcid":null},"institutions":[{"id":"https://openalex.org/I170897317","display_name":"Duke University","ror":"https://ror.org/00py81415","country_code":"US","type":"education","lineage":["https://openalex.org/I170897317"]},{"id":"https://openalex.org/I36194607","display_name":"Duke Energy (United States)","ror":"https://ror.org/030r69d22","country_code":"US","type":"company","lineage":["https://openalex.org/I36194607"]},{"id":"https://openalex.org/I4210153043","display_name":"Duke University Hospital","ror":"https://ror.org/04bct7p84","country_code":"US","type":"healthcare","lineage":["https://openalex.org/I4210126298","https://openalex.org/I4210144876","https://openalex.org/I4210144876","https://openalex.org/I4210153043"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Maximilian Holsman","raw_affiliation_strings":["Duke University Duke University Duke University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Duke University Duke University Duke University","institution_ids":["https://openalex.org/I36194607","https://openalex.org/I170897317","https://openalex.org/I4210153043"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010418955","display_name":"Yukun Huang","orcid":"https://orcid.org/0000-0002-9662-2773"},"institutions":[{"id":"https://openalex.org/I170897317","display_name":"Duke University","ror":"https://ror.org/00py81415","country_code":"US","type":"education","lineage":["https://openalex.org/I170897317"]},{"id":"https://openalex.org/I36194607","display_name":"Duke Energy (United States)","ror":"https://ror.org/030r69d22","country_code":"US","type":"company","lineage":["https://openalex.org/I36194607"]},{"id":"https://openalex.org/I4210153043","display_name":"Duke University Hospital","ror":"https://ror.org/04bct7p84","country_code":"US","type":"healthcare","lineage":["https://openalex.org/I4210126298","https://openalex.org/I4210144876","https://openalex.org/I4210144876","https://openalex.org/I4210153043"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yukun Huang","raw_affiliation_strings":["Duke University Duke University Duke University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Duke University Duke University Duke University","institution_ids":["https://openalex.org/I36194607","https://openalex.org/I170897317","https://openalex.org/I4210153043"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5055033421","display_name":"Bhuwan Dhingra","orcid":"https://orcid.org/0000-0002-6874-9515"},"institutions":[{"id":"https://openalex.org/I170897317","display_name":"Duke University","ror":"https://ror.org/00py81415","country_code":"US","type":"education","lineage":["https://openalex.org/I170897317"]},{"id":"https://openalex.org/I36194607","display_name":"Duke Energy (United States)","ror":"https://ror.org/030r69d22","country_code":"US","type":"company","lineage":["https://openalex.org/I36194607"]},{"id":"https://openalex.org/I4210153043","display_name":"Duke University Hospital","ror":"https://ror.org/04bct7p84","country_code":"US","type":"healthcare","lineage":["https://openalex.org/I4210126298","https://openalex.org/I4210144876","https://openalex.org/I4210144876","https://openalex.org/I4210153043"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bhuwan Dhingra","raw_affiliation_strings":["Duke University Duke University Duke University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Duke University Duke University Duke University","institution_ids":["https://openalex.org/I36194607","https://openalex.org/I170897317","https://openalex.org/I4210153043"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"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":"26257","last_page":"26273"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10142","display_name":"Formal Methods in Verification","score":0.8458999991416931,"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/T10142","display_name":"Formal Methods in Verification","score":0.8458999991416931,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7989397048950195},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.7561625838279724},{"id":"https://openalex.org/keywords/fuzzy-logic","display_name":"Fuzzy logic","score":0.5495270490646362},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3904193341732025},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.29794949293136597},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.20715725421905518}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7989397048950195},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.7561625838279724},{"id":"https://openalex.org/C58166","wikidata":"https://www.wikidata.org/wiki/Q224821","display_name":"Fuzzy logic","level":2,"score":0.5495270490646362},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3904193341732025},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.29794949293136597},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.20715725421905518}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.1346","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.1346","pdf_url":"https://aclanthology.org/2025.findings-acl.1346.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.1346","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.1346","pdf_url":"https://aclanthology.org/2025.findings-acl.1346.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G2086134654","display_name":null,"funder_award_id":"IIS-2211526","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G3333846949","display_name":"III: Medium: Ask the Experts: Generating Question-Answer Pairs for Addressing Information Deficits about Vaccines","funder_award_id":"2211526","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"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412944498.pdf","grobid_xml":"https://content.openalex.org/works/W4412944498.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Speculative":[0,54],"Decoding":[1,55],"(SD)":[2],"enforces":[3],"strict":[4,23],"distributional":[5,147],"equivalence":[6,24,148],"to":[7,89,103,133,152,172,183],"the":[8,17,26,38,69,72,83],"target":[9,18,39,73,84,154],"model":[10,76,155],"when":[11],"accepting":[12,64],"candidate":[13,65],"tokens.While":[14],"it":[15,182],"maintains":[16],"model's":[19],"generation":[20,92],"quality,":[21],"this":[22,168,176],"limits":[25],"speedup":[27],"achievable":[28],"by":[29,63,169],"SD":[30,62,116,135,164],"and":[31,74],"prevents":[32],"users":[33,88],"from":[34,37,82],"trading":[35],"deviations":[36],"distribution":[40],"in":[41,124],"exchange":[42],"for":[43,79,94],"further":[44],"inference":[45,95],"speed":[46],"gains.To":[47],"address":[48],"these":[49],"limitations,":[50],"we":[51,166],"introduce":[52],"Fuzzy":[53],"(FSD)":[56],"-a":[57],"decoding":[58],"algorithm":[59],"that":[60,146],"generalizes":[61],"tokens":[66,111,141],"based":[67],"on":[68],"divergences":[70],"between":[71],"draft":[75],"distributions.By":[77],"allowing":[78,181],"controlled":[80],"divergence":[81],"model,":[85],"FSD":[86,129,157,171],"enables":[87],"flexibly":[90],"trade":[91],"quality":[93],"speed.Across":[96],"several":[97],"benchmarks,":[98],"our":[99],"method":[100],"is":[101,130,149],"able":[102,132],"achieve":[104],"significant":[105],"runtime":[106],"improvements":[107],"of":[108],"over":[109,139],"5":[110],"per":[112,142],"second":[113,143],"faster":[114],"than":[115],"at":[117,138],"only":[118],"an":[119],"approximate":[120],"2%":[121],"absolute":[122],"reduction":[123],"benchmark":[125,136],"accuracy.In":[126],"many":[127],"cases,":[128],"even":[131],"match":[134],"accuracy":[137],"2":[140],"faster,":[144],"demonstrating":[145],"not":[150],"necessary":[151],"maintain":[153],"performance.Furthermore,":[156],"can":[158],"be":[159],"seamlessly":[160],"integrated":[161],"into":[162],"existing":[163,177],"extensions;":[165],"demonstrate":[167],"applying":[170],"EAGLE-2,":[173],"greatly":[174],"enhancing":[175],"extension's":[178],"efficiency":[179],"while":[180],"leverage":[184],"FSD's":[185],"tunable":[186],"quality-speed":[187],"tradeoff.":[188]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}