{"id":"https://openalex.org/W7223899403","doi":"https://doi.org/10.48550/arxiv.2610.33368","title":"Multiple Importance Sampling for Guided Speculative Inference","display_name":"Multiple Importance Sampling for Guided Speculative Inference","publication_year":2026,"publication_date":"2026-10-23","ids":{"openalex":"https://openalex.org/W7223899403","doi":"https://doi.org/10.48550/arxiv.2610.33368"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.33368","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.33368","pdf_url":"https://arxiv.org/pdf/2610.33368","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2610.33368","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5123181985","display_name":"Joseph Rowan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I185261750","display_name":"University of Toronto","ror":"https://ror.org/03dbr7087","country_code":"CA","type":"education","lineage":["https://openalex.org/I185261750"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Rowan, Joseph","raw_affiliation_strings":["Department of Electrical and Computer Engineering , University of Toronto"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering , University of Toronto","institution_ids":["https://openalex.org/I185261750"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5046253142","display_name":"Ashish J. Khisti","orcid":"https://orcid.org/0000-0002-2331-8965","observed_orcids":["https://orcid.org/0000-0002-2331-8965"]},"institutions":[{"id":"https://openalex.org/I185261750","display_name":"University of Toronto","ror":"https://ror.org/03dbr7087","country_code":"CA","type":"education","lineage":["https://openalex.org/I185261750"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Khisti, Ashish J.","raw_affiliation_strings":["Department of Electrical and Computer Engineering , University of Toronto"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Electrical and Computer Engineering , University of Toronto","institution_ids":["https://openalex.org/I185261750"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I185261750"],"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/T10028","display_name":"Topic Modeling","score":0.40450000762939453,"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":0.40450000762939453,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.23829999566078186,"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/T12056","display_name":"Markov Chains and Monte Carlo Methods","score":0.12950000166893005,"subfield":{"id":"https://openalex.org/subfields/2613","display_name":"Statistics and Probability"},"field":{"id":"https://openalex.org/fields/26","display_name":"Mathematics"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/multiple-importance-sampling","display_name":"multiple importance sampling","score":0.9779999852180481},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.7889999747276306},{"id":"https://openalex.org/keywords/test-time-scaling","display_name":"test-time scaling","score":0.828000009059906},{"id":"https://openalex.org/keywords/reward-models","display_name":"reward models","score":0.5649999976158142},{"id":"https://openalex.org/keywords/sample-weighting","display_name":"sample weighting","score":0.5509999990463257}],"concepts":[{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.6872000098228455},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6862000226974487},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.633400022983551},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.5393000245094299},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.5257999897003174},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.45820000767707825},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4472000002861023},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44429999589920044},{"id":"https://openalex.org/C2780428219","wikidata":"https://www.wikidata.org/wiki/Q16952335","display_name":"Cover (algebra)","level":2,"score":0.430400013923645},{"id":"https://openalex.org/C2778029271","wikidata":"https://www.wikidata.org/wiki/Q5421931","display_name":"Extension (predicate logic)","level":2,"score":0.4300000071525574},{"id":"https://openalex.org/C4679612","wikidata":"https://www.wikidata.org/wiki/Q866298","display_name":"Aggregate (composite)","level":2,"score":0.4275999963283539},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.40290001034736633},{"id":"https://openalex.org/C2780522230","wikidata":"https://www.wikidata.org/wiki/Q1140419","display_name":"Ambiguity","level":2,"score":0.39149999618530273},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.37950000166893005},{"id":"https://openalex.org/C73602740","wikidata":"https://www.wikidata.org/wiki/Q7795822","display_name":"Thompson sampling","level":3,"score":0.3635999858379364},{"id":"https://openalex.org/C207390915","wikidata":"https://www.wikidata.org/wiki/Q1230525","display_name":"Divergence (linguistics)","level":2,"score":0.3614000082015991},{"id":"https://openalex.org/C70136482","wikidata":"https://www.wikidata.org/wiki/Q13583781","display_name":"A-weighting","level":3,"score":0.36010000109672546},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.3467999994754791},{"id":"https://openalex.org/C167723999","wikidata":"https://www.wikidata.org/wiki/Q3773214","display_name":"Sampling distribution","level":2,"score":0.33009999990463257},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.32089999318122864},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.29649999737739563},{"id":"https://openalex.org/C168031717","wikidata":"https://www.wikidata.org/wiki/Q1530280","display_name":"Balance (ability)","level":2,"score":0.29249998927116394},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.2881999909877777},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.2858999967575073},{"id":"https://openalex.org/C45235069","wikidata":"https://www.wikidata.org/wiki/Q278425","display_name":"Table (database)","level":2,"score":0.2854999899864197},{"id":"https://openalex.org/C194051981","wikidata":"https://www.wikidata.org/wiki/Q1337691","display_name":"Economic shortage","level":3,"score":0.2791000008583069},{"id":"https://openalex.org/C2984391234","wikidata":"https://www.wikidata.org/wiki/Q195771","display_name":"Sequential sampling","level":3,"score":0.27090001106262207},{"id":"https://openalex.org/C110121322","wikidata":"https://www.wikidata.org/wiki/Q865811","display_name":"Distribution (mathematics)","level":2,"score":0.2531000077724457},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.25279998779296875}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.33368","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.33368","pdf_url":"https://arxiv.org/pdf/2610.33368","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2610.33368","doi":"https://doi.org/10.48550/arxiv.2610.33368","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.33368","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.33368","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.33368","pdf_url":"https://arxiv.org/pdf/2610.33368","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7223899403.pdf","grobid_xml":"https://content.openalex.org/works/W7223899403.grobid-xml"},"referenced_works_count":5,"referenced_works":["https://openalex.org/W2136325069","https://openalex.org/W3001279689","https://openalex.org/W4415025609","https://openalex.org/W4417142985","https://openalex.org/W6910895196"],"related_works":["https://openalex.org/W4414801957","https://openalex.org/W4417228524","https://openalex.org/W4411120068","https://openalex.org/W7222277050","https://openalex.org/W4415035614","https://openalex.org/W7166877517","https://openalex.org/W7204542309","https://openalex.org/W7138467026","https://openalex.org/W7125114978","https://openalex.org/W7139093401"],"abstract_inverted_index":{"We":[0,69],"study":[1],"an":[2,23],"extension":[3],"of":[4,13,25,50,66,115,131,140],"the":[5,47,74,78,84,91,111,129,141],"Guided":[6],"Speculative":[7],"Inference":[8],"(GSI)":[9],"algorithm":[10],"for":[11],"test-time-scaling":[12],"large":[14],"language":[15],"models":[16],"(LLMs)":[17],"to":[18,59,98,126],"a":[19,37,42,63,101,132],"novel":[20],"setting":[21],"where":[22,128,143],"ensemble":[24],"smaller":[26],"LLMs":[27,108],"proposes":[28],"draft":[29,134],"reasoning":[30],"steps":[31],"which":[32],"are":[33],"later":[34],"verified":[35],"by":[36],"larger":[38],"base":[39],"model":[40,135],"and":[41,83],"process":[43],"reward":[44,86],"model.":[45],"Using":[46],"established":[48],"framework":[49],"multiple":[51],"importance":[52],"sampling":[53],"(MIS),":[54],"we":[55],"introduce":[56],"weighting":[57],"schemes":[58],"aggregate":[60],"samples":[61,116],"from":[62,77],"diverse":[64,107],"collection":[65],"proposal":[67],"distributions.":[68],"provide":[70],"new":[71],"bounds":[72],"on":[73],"KL":[75],"divergence":[76],"optimal":[79],"reward-tilted":[80],"target":[81],"distribution":[82],"expected":[85],"degradation,":[87],"then":[88],"demonstrate":[89],"that":[90,139],"well-known":[92],"balance":[93],"heuristic":[94],"is":[95],"optimal.":[96],"Compared":[97],"GSI":[99,144],"with":[100],"single":[102,133],"well-tuned":[103],"drafter,":[104],"combining":[105],"several":[106],"while":[109,122],"retaining":[110],"same":[112],"total":[113],"number":[114],"maintains":[117],"competitive":[118],"pass-at-$N$":[119],"downstream":[120],"performance":[121],"being":[123],"mathematically":[124],"robust":[125],"situations":[127],"support":[130],"does":[136],"not":[137],"cover":[138],"target,":[142],"may":[145],"fail.":[146]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-28T00:00:00"}