{"id":"https://openalex.org/W4414359318","doi":"https://doi.org/10.24963/ijcai.2025/951","title":"Semi-Clairvoyant Scheduling of Speculative Decoding Requests to Minimize LLM Inference Latency","display_name":"Semi-Clairvoyant Scheduling of Speculative Decoding Requests to Minimize LLM Inference Latency","publication_year":2025,"publication_date":"2025-09-01","ids":{"openalex":"https://openalex.org/W4414359318","doi":"https://doi.org/10.24963/ijcai.2025/951"},"language":"en","primary_location":{"id":"doi:10.24963/ijcai.2025/951","is_oa":false,"landing_page_url":"https://doi.org/10.24963/ijcai.2025/951","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"type":"conference-paper","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/A5101479594","display_name":"Ruixiao Li","orcid":"https://orcid.org/0009-0001-4804-4649"},"institutions":[{"id":"https://openalex.org/I87445476","display_name":"Xi'an Jiaotong University","ror":"https://ror.org/017zhmm22","country_code":"CN","type":"education","lineage":["https://openalex.org/I87445476"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ruixiao Li","raw_affiliation_strings":["Xi'an Jiaotong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xi'an Jiaotong University","institution_ids":["https://openalex.org/I87445476"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088461411","display_name":"Fahao Chen","orcid":"https://orcid.org/0000-0002-4345-1296"},"institutions":[{"id":"https://openalex.org/I141591182","display_name":"University of Aizu","ror":"https://ror.org/02pg0e883","country_code":"JP","type":"education","lineage":["https://openalex.org/I141591182"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Fahao Chen","raw_affiliation_strings":["The University of Aizu"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The University of Aizu","institution_ids":["https://openalex.org/I141591182"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100432795","display_name":"Peng Li","orcid":"https://orcid.org/0000-0003-4981-0496"},"institutions":[{"id":"https://openalex.org/I87445476","display_name":"Xi'an Jiaotong University","ror":"https://ror.org/017zhmm22","country_code":"CN","type":"education","lineage":["https://openalex.org/I87445476"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Peng Li","raw_affiliation_strings":["Xi'an Jiaotong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xi'an Jiaotong University","institution_ids":["https://openalex.org/I87445476"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"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":"8554","last_page":"8562"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.965399980545044,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.965399980545044,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/security-token","display_name":"Security token","score":0.8044000267982483},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.7092999815940857},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.6843000054359436},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.6191999912261963},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.5565999746322632},{"id":"https://openalex.org/keywords/queue","display_name":"Queue","score":0.4778999984264374},{"id":"https://openalex.org/keywords/preemption","display_name":"Preemption","score":0.39340001344680786},{"id":"https://openalex.org/keywords/execution-time","display_name":"Execution time","score":0.3849000036716461}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8675000071525574},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.8044000267982483},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7092999815940857},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6843000054359436},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.6191999912261963},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.5565999746322632},{"id":"https://openalex.org/C160403385","wikidata":"https://www.wikidata.org/wiki/Q220543","display_name":"Queue","level":2,"score":0.4778999984264374},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.43630000948905945},{"id":"https://openalex.org/C206952183","wikidata":"https://www.wikidata.org/wiki/Q1193100","display_name":"Preemption","level":2,"score":0.39340001344680786},{"id":"https://openalex.org/C2989134064","wikidata":"https://www.wikidata.org/wiki/Q288510","display_name":"Execution time","level":2,"score":0.3849000036716461},{"id":"https://openalex.org/C68387754","wikidata":"https://www.wikidata.org/wiki/Q7271585","display_name":"Schedule","level":2,"score":0.3813000023365021},{"id":"https://openalex.org/C2984822820","wikidata":"https://www.wikidata.org/wiki/Q1123036","display_name":"Processor scheduling","level":3,"score":0.33550000190734863},{"id":"https://openalex.org/C122141398","wikidata":"https://www.wikidata.org/wiki/Q5456330","display_name":"Fixed-priority pre-emptive scheduling","level":5,"score":0.33469998836517334},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.33399999141693115},{"id":"https://openalex.org/C46743427","wikidata":"https://www.wikidata.org/wiki/Q1341685","display_name":"Inference engine","level":3,"score":0.3296999931335449},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.313400000333786},{"id":"https://openalex.org/C115067241","wikidata":"https://www.wikidata.org/wiki/Q1639854","display_name":"Token passing","level":3,"score":0.305400013923645},{"id":"https://openalex.org/C107568181","wikidata":"https://www.wikidata.org/wiki/Q5319000","display_name":"Dynamic priority scheduling","level":3,"score":0.3019999861717224},{"id":"https://openalex.org/C127456818","wikidata":"https://www.wikidata.org/wiki/Q238879","display_name":"Rate-monotonic scheduling","level":4,"score":0.27410000562667847},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.26759999990463257},{"id":"https://openalex.org/C31689143","wikidata":"https://www.wikidata.org/wiki/Q733809","display_name":"Fair-share scheduling","level":3,"score":0.2651999890804291},{"id":"https://openalex.org/C2777131613","wikidata":"https://www.wikidata.org/wiki/Q18394147","display_name":"Retard","level":2,"score":0.26409998536109924},{"id":"https://openalex.org/C32310161","wikidata":"https://www.wikidata.org/wiki/Q666365","display_name":"Earliest deadline first scheduling","level":5,"score":0.259799987077713}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.24963/ijcai.2025/951","is_oa":false,"landing_page_url":"https://doi.org/10.24963/ijcai.2025/951","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Speculative":[0,104],"decoding":[1],"accelerates":[2],"Large":[3],"Language":[4],"Model":[5],"(LLM)":[6],"inference":[7,36,40,111,118,180],"by":[8,88,120,182],"employing":[9],"a":[10,49,96,108],"small":[11],"speculative":[12],"model":[13],"(SSM)":[14],"to":[15,125,141,186],"generate":[16],"multiple":[17,145],"candidate":[18],"tokens":[19],"and":[20,82,136,148,170],"verify":[21],"them":[22],"using":[23],"the":[24,89,157,167],"LLM":[25,35],"in":[26,56],"parallel.":[27],"This":[28],"technique":[29],"has":[30],"been":[31],"widely":[32],"integrated":[33],"into":[34],"serving":[37],"systems.":[38,58],"However,":[39],"requests":[41,55,123,172],"typically":[42],"exhibit":[43],"uncertain":[44],"execution":[45,62,75,137,151,168],"time,":[46],"which":[47,70],"poses":[48],"significant":[50],"challenge":[51],"of":[52,86,110],"efficiently":[53],"scheduling":[54,99,122,188],"these":[57],"Existing":[59],"work":[60],"estimates":[61],"time":[63,76,138,169],"based":[64],"solely":[65],"on":[66,78],"predicted":[67],"output":[68,80],"length,":[69],"could":[71],"be":[72],"inaccurate":[73],"because":[74],"depends":[77],"both":[79],"length":[81],"token":[83,131,158],"acceptance":[84,132,159],"rate":[85,133,160],"verification":[87],"LLM.":[90],"In":[91],"this":[92],"paper,":[93],"we":[94],"propose":[95],"semi-clairvoyant":[97],"request":[98,150],"algorithm":[100],"called":[101],"Least-Attained/Perceived-Service":[102],"for":[103],"Decoding":[105],"(LAPS-SD).":[106],"Given":[107],"number":[109],"requests,":[112],"LAPS-SD":[113,143,163,178],"can":[114,164],"effectively":[115],"minimize":[116],"average":[117],"latency":[119,181],"adaptively":[121],"according":[124],"their":[126],"features":[127],"during":[128],"decoding.":[129],"When":[130],"is":[134,139],"dynamic":[135],"difficult":[140],"estimate,":[142],"maintains":[144],"priority":[146],"queues":[147],"allows":[149],"preemption":[152],"across":[153],"different":[154],"queues.":[155],"Once":[156],"becomes":[161],"stable,":[162],"accurately":[165],"estimate":[166],"schedule":[171],"accordingly.":[173],"Extensive":[174],"experiments":[175],"show":[176],"that":[177],"reduces":[179],"approximately":[183],"39%":[184],"compared":[185],"state-of-the-art":[187],"methods.":[189]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}