{"id":"https://openalex.org/W7138467026","doi":"https://doi.org/10.1609/aaai.v40i39.40625","title":"Re-SpS: A Reinforcement Learning Approach to Speculative Sampling","display_name":"Re-SpS: A Reinforcement Learning Approach to Speculative Sampling","publication_year":2026,"publication_date":"2026-03-14","ids":{"openalex":"https://openalex.org/W7138467026","doi":"https://doi.org/10.1609/aaai.v40i39.40625"},"language":"en","primary_location":{"id":"doi:10.1609/aaai.v40i39.40625","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i39.40625","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/40625/44586","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://ojs.aaai.org/index.php/AAAI/article/download/40625/44586","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129720787","display_name":"Chenan Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I16285277","display_name":"William & Mary","ror":"https://ror.org/03hsf0573","country_code":"US","type":"education","lineage":["https://openalex.org/I16285277"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chenan Wang","raw_affiliation_strings":["William & Mary"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"William & Mary","institution_ids":["https://openalex.org/I16285277"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025403238","display_name":"Daniel Shi","orcid":"https://orcid.org/0000-0002-8689-007X"},"institutions":[{"id":"https://openalex.org/I16285277","display_name":"William & Mary","ror":"https://ror.org/03hsf0573","country_code":"US","type":"education","lineage":["https://openalex.org/I16285277"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Daniel H. Shi","raw_affiliation_strings":["William & Mary"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"William & Mary","institution_ids":["https://openalex.org/I16285277"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129693808","display_name":"Haipeng Chen","orcid":null},"institutions":[{"id":"https://openalex.org/I16285277","display_name":"William & Mary","ror":"https://ror.org/03hsf0573","country_code":"US","type":"education","lineage":["https://openalex.org/I16285277"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Haipeng Chen","raw_affiliation_strings":["William & Mary"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"William & Mary","institution_ids":["https://openalex.org/I16285277"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I16285277"],"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":"40","issue":"39","first_page":"33386","last_page":"33394"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.3199000060558319,"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":0.3199000060558319,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.13259999454021454,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},{"id":"https://openalex.org/T13702","display_name":"Machine Learning in Healthcare","score":0.09319999814033508,"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/speedup","display_name":"Speedup","score":0.8529000282287598},{"id":"https://openalex.org/keywords/hyperparameter","display_name":"Hyperparameter","score":0.8251000046730042},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.7738000154495239},{"id":"https://openalex.org/keywords/tree","display_name":"Tree (set theory)","score":0.4668000042438507},{"id":"https://openalex.org/keywords/flexibility","display_name":"Flexibility (engineering)","score":0.453900009393692},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.4535999894142151},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.4442000091075897},{"id":"https://openalex.org/keywords/thompson-sampling","display_name":"Thompson sampling","score":0.40700000524520874}],"concepts":[{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.8529000282287598},{"id":"https://openalex.org/C8642999","wikidata":"https://www.wikidata.org/wiki/Q4171168","display_name":"Hyperparameter","level":2,"score":0.8251000046730042},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8169000148773193},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7738000154495239},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5914999842643738},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5860999822616577},{"id":"https://openalex.org/C113174947","wikidata":"https://www.wikidata.org/wiki/Q2859736","display_name":"Tree (set theory)","level":2,"score":0.4668000042438507},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.453900009393692},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.4535999894142151},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.4442000091075897},{"id":"https://openalex.org/C73602740","wikidata":"https://www.wikidata.org/wiki/Q7795822","display_name":"Thompson sampling","level":3,"score":0.40700000524520874},{"id":"https://openalex.org/C140745168","wikidata":"https://www.wikidata.org/wiki/Q1210082","display_name":"Tree traversal","level":2,"score":0.373199999332428},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.3458999991416931},{"id":"https://openalex.org/C84525736","wikidata":"https://www.wikidata.org/wiki/Q831366","display_name":"Decision tree","level":2,"score":0.3330000042915344},{"id":"https://openalex.org/C79581498","wikidata":"https://www.wikidata.org/wiki/Q1367530","display_name":"Suite","level":2,"score":0.2743000090122223},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.25870001316070557},{"id":"https://openalex.org/C163797641","wikidata":"https://www.wikidata.org/wiki/Q2067937","display_name":"Tree structure","level":3,"score":0.25440001487731934},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.25369998812675476},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.25}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1609/aaai.v40i39.40625","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i39.40625","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/40625/44586","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},{"id":"pmh:oai:ojs.aaai.org:article/40625","is_oa":false,"landing_page_url":"https://ojs.aaai.org/index.php/AAAI/article/view/40625","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"2159-5399","raw_type":"info:eu-repo/semantics/article"}],"best_oa_location":{"id":"doi:10.1609/aaai.v40i39.40625","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i39.40625","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/40625/44586","source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7138467026.pdf","grobid_xml":"https://content.openalex.org/works/W7138467026.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Inference":[0],"time":[1],"latency":[2],"has":[3],"remained":[4],"an":[5],"open":[6],"challenge":[7],"for":[8,23,60,70,112],"real":[9],"world":[10],"applications":[11],"of":[12],"large":[13],"language":[14],"models":[15],"(LLMs).":[16],"State-of-the-art":[17],"(SOTA)":[18],"speculative":[19,92],"sampling":[20],"(SpS)":[21],"methods":[22],"LLMs,":[24],"like":[25],"EAGLE-3,":[26,129],"use":[27],"tree-based":[28],"drafting":[29],"to":[30,132,141,145],"explore":[31],"multiple":[32],"candidate":[33],"continuations":[34],"in":[35,81,154],"parallel.":[36],"However,":[37],"the":[38,41,64,126,136],"hyperparameters":[39,80],"controlling":[40],"tree":[42,72,79],"structure":[43],"are":[44],"static,":[45],"which":[46],"limits":[47],"flexibility":[48],"and":[49,54,107,139],"efficiency":[50],"across":[51,118,147],"diverse":[52,120,149],"contexts":[53],"domains.":[55],"We":[56],"introduce":[57],"Reinforcement":[58],"learning":[59,67,83],"Speculative":[61],"Sampling":[62],"(Re-SpS),":[63],"first":[65],"reinforcement":[66],"(RL)-based":[68],"framework":[69],"draft":[71,78],"hyperparameter":[73],"optimization.":[74],"Re-SpS":[75],"dynamically":[76],"adjusts":[77],"real-time,":[82],"context-aware":[84],"policies":[85],"that":[86],"maximize":[87],"generation":[88],"speed":[89],"by":[90],"balancing":[91],"aggression":[93],"with":[94,151],"computational":[95],"overhead.":[96],"It":[97],"leverages":[98],"efficient":[99],"state":[100],"representations":[101],"from":[102],"target":[103],"model":[104],"hidden":[105],"states":[106],"introduces":[108],"multi-step":[109],"action":[110],"persistence":[111],"better":[113],"context":[114],"modeling.":[115],"Evaluation":[116],"results":[117],"five":[119,148],"benchmarks":[121],"demonstrate":[122],"consistent":[123],"improvements":[124],"over":[125,135],"SOTA":[127],"method":[128],"achieving":[130],"up":[131,140],"5.45x":[133],"speedup":[134,143],"backbone":[137],"LLM":[138],"1.12x":[142],"compared":[144],"EAGLE-3":[146],"benchmarks,":[150],"no":[152],"loss":[153],"output":[155],"fidelity.":[156]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-03-18T00:00:00"}