{"id":"https://openalex.org/W7147634669","doi":"https://doi.org/10.48550/arxiv.2603.27027","title":"TAPS: Task Aware Proposal Distributions for Speculative Sampling","display_name":"TAPS: Task Aware Proposal Distributions for Speculative Sampling","publication_year":2026,"publication_date":"2026-03-27","ids":{"openalex":"https://openalex.org/W7147634669","doi":"https://doi.org/10.48550/arxiv.2603.27027"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.27027","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.27027","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,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2603.27027","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5132599195","display_name":"Mohamad Zbib","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zbib, Mohamad","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132622864","display_name":"Mohamad Bazzi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bazzi, Mohamad","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048867203","display_name":"Ammar Mohanna","orcid":"https://orcid.org/0000-0001-7622-317X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mohanna, Ammar","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029119136","display_name":"Hasan Abed Al Kader Hammoud","orcid":"https://orcid.org/0000-0002-2991-0804"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hammoud, Hasan Abed Al Kader","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5132555708","display_name":"Bernard Ghanem","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ghanem, Bernard","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":false,"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":0.2529999911785126,"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":0.2529999911785126,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.18629999458789825,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.11020000278949738,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.8389000296592712},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6733999848365784},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6581000089645386},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.5529000163078308},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.5202000141143799},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.45249998569488525},{"id":"https://openalex.org/keywords/routing","display_name":"Routing (electronic design automation)","score":0.4521999955177307},{"id":"https://openalex.org/keywords/downstream","display_name":"Downstream (manufacturing)","score":0.41670000553131104}],"concepts":[{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.8389000296592712},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7074999809265137},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6733999848365784},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6581000089645386},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.5529000163078308},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.5202000141143799},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.45249998569488525},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.4521999955177307},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4302999973297119},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.41670000553131104},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.39969998598098755},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.3758000135421753},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3508000075817108},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.3411000072956085},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.33970001339912415},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.33000001311302185},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3264000117778778},{"id":"https://openalex.org/C44249647","wikidata":"https://www.wikidata.org/wiki/Q208498","display_name":"Confidence interval","level":2,"score":0.30630001425743103},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.289900004863739},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.27459999918937683},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2711000144481659},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.27070000767707825},{"id":"https://openalex.org/C193969084","wikidata":"https://www.wikidata.org/wiki/Q7452500","display_name":"Sequential decoding","level":4,"score":0.26350000500679016},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.25099998712539673}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.27027","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.27027","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.27027","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.27027","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,"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"},"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":{"Speculative":[0],"decoding":[1,42,108,175],"accelerates":[2],"autoregressive":[3],"generation":[4],"by":[5,75],"letting":[6],"a":[7,15,148],"lightweight":[8,55],"draft":[9,26,47,181,189],"model":[10,18],"propose":[11],"future":[12],"tokens":[13,156],"that":[14,173,196],"larger":[16,102],"target":[17],"then":[19],"verifies":[20],"in":[21,206],"parallel.":[22],"In":[23],"practice,":[24],"however,":[25],"models":[27],"are":[28,85,93,199],"usually":[29],"trained":[30,60],"on":[31,45,61,68,87,95,180,185],"broad":[32],"generic":[33],"corpora,":[34],"which":[35],"leaves":[36],"it":[37],"unclear":[38],"how":[39,113],"much":[40,165],"speculative":[41,174],"quality":[43,176],"depends":[44,177],"the":[46,137,186],"training":[48,79,98,190],"distribution.":[49],"We":[50,110],"study":[51,112],"this":[52],"question":[53],"with":[54],"HASS":[56],"and":[57,64,72,133,192,195],"EAGLE-2":[58],"drafters":[59,117,198],"MathInstruct,":[62],"ShareGPT,":[63],"mixed-data":[65],"variants,":[66],"evaluated":[67],"MT-Bench,":[69],"GSM8K,":[70],"MATH-500,":[71],"SVAMP.":[73],"Measured":[74],"acceptance":[76,139],"length,":[77],"task-specific":[78],"yields":[80,136],"clear":[81],"specialization:":[82],"MathInstruct-trained":[83],"drafts":[84,92,132],"strongest":[86,94],"reasoning":[88],"benchmarks,":[89],"while":[90],"ShareGPT-trained":[91],"MT-Bench.":[96],"Mixed-data":[97],"improves":[99,129],"robustness,":[100],"but":[101,162,183],"mixtures":[103],"do":[104],"not":[105,178],"dominate":[106],"across":[107],"temperatures.":[109],"also":[111,184],"to":[114,158],"combine":[115],"specialized":[116,197],"at":[118,202],"inference":[119,203],"time.":[120],"Naive":[121],"checkpoint":[122],"averaging":[123],"performs":[124],"poorly,":[125],"whereas":[126],"confidence-based":[127],"routing":[128,151,168],"over":[130],"single-domain":[131],"merged-tree":[134],"verification":[135],"highest":[138],"length":[140],"overall":[141],"for":[142],"both":[143],"backbones.":[144],"Finally,":[145],"confidence":[146,163],"is":[147],"more":[149],"useful":[150],"signal":[152],"than":[153,205],"entropy:":[154],"rejected":[155],"tend":[157],"have":[159],"higher":[160],"entropy,":[161],"produces":[164],"clearer":[166],"benchmark-level":[167],"decisions.":[169],"These":[170],"results":[171],"show":[172],"only":[179],"architecture,":[182],"match":[187],"between":[188],"data":[191],"downstream":[193],"workload,":[194],"better":[200],"combined":[201],"time":[204],"weight":[207],"space.":[208]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-04-02T00:00:00"}