{"id":"https://openalex.org/W7164204658","doi":"https://doi.org/10.48550/arxiv.2606.11119","title":"TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning","display_name":"TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning","publication_year":2026,"publication_date":"2026-06-09","ids":{"openalex":"https://openalex.org/W7164204658","doi":"https://doi.org/10.48550/arxiv.2606.11119"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.11119","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.11119","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.11119","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138384285","display_name":"Heming Zou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zou, Heming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138349936","display_name":"Qi Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Qi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138383324","display_name":"Yun Qu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qu, Yun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138393019","display_name":"Yuhang Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Yuhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138374786","display_name":"Lizhou Cai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cai, Lizhou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138302953","display_name":"Yixiu Mao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mao, Yixiu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138349862","display_name":"Ru Peng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peng, Ru","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138355944","display_name":"Xin Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Xin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138298735","display_name":"Weijie Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Weijie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138300875","display_name":"Kai Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Kai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027918344","display_name":"Saiyong Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Saiyong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138381426","display_name":"Xiangyang Ji","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ji, Xiangyang","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/T10028","display_name":"Topic Modeling","score":0.3158000111579895,"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":0.3158000111579895,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.10649999976158142,"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":0.0885000005364418,"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/reinforcement-learning","display_name":"Reinforcement learning","score":0.7242000102996826},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.6687999963760376},{"id":"https://openalex.org/keywords/prefix","display_name":"Prefix","score":0.508899986743927},{"id":"https://openalex.org/keywords/hindsight-bias","display_name":"Hindsight bias","score":0.45170000195503235},{"id":"https://openalex.org/keywords/trace","display_name":"TRACE (psycholinguistics)","score":0.4293999969959259},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.4178999960422516},{"id":"https://openalex.org/keywords/sample","display_name":"Sample (material)","score":0.4090999960899353},{"id":"https://openalex.org/keywords/contrast","display_name":"Contrast (vision)","score":0.35589998960494995},{"id":"https://openalex.org/keywords/tree","display_name":"Tree (set theory)","score":0.353300005197525}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7318000197410583},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7242000102996826},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.6687999963760376},{"id":"https://openalex.org/C141603448","wikidata":"https://www.wikidata.org/wiki/Q134830","display_name":"Prefix","level":2,"score":0.508899986743927},{"id":"https://openalex.org/C10347200","wikidata":"https://www.wikidata.org/wiki/Q1960297","display_name":"Hindsight bias","level":2,"score":0.45170000195503235},{"id":"https://openalex.org/C75291252","wikidata":"https://www.wikidata.org/wiki/Q1315756","display_name":"TRACE (psycholinguistics)","level":2,"score":0.4293999969959259},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.4178999960422516},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.4090999960899353},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.35589998960494995},{"id":"https://openalex.org/C113174947","wikidata":"https://www.wikidata.org/wiki/Q2859736","display_name":"Tree (set theory)","level":2,"score":0.353300005197525},{"id":"https://openalex.org/C2776007630","wikidata":"https://www.wikidata.org/wiki/Q2798912","display_name":"Accountability","level":2,"score":0.34850001335144043},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3474999964237213},{"id":"https://openalex.org/C8505890","wikidata":"https://www.wikidata.org/wiki/Q605095","display_name":"Budget constraint","level":2,"score":0.33489999175071716},{"id":"https://openalex.org/C2778334786","wikidata":"https://www.wikidata.org/wiki/Q1586270","display_name":"Variation (astronomy)","level":2,"score":0.33079999685287476},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.32600000500679016},{"id":"https://openalex.org/C7149132","wikidata":"https://www.wikidata.org/wiki/Q1377840","display_name":"Forgetting","level":2,"score":0.30149999260902405},{"id":"https://openalex.org/C29202148","wikidata":"https://www.wikidata.org/wiki/Q287260","display_name":"Resource allocation","level":2,"score":0.2992999851703644},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.2939000129699707},{"id":"https://openalex.org/C42475967","wikidata":"https://www.wikidata.org/wiki/Q194292","display_name":"Operations research","level":1,"score":0.28999999165534973},{"id":"https://openalex.org/C88626702","wikidata":"https://www.wikidata.org/wiki/Q1128903","display_name":"Continuation","level":2,"score":0.2766999900341034},{"id":"https://openalex.org/C166052673","wikidata":"https://www.wikidata.org/wiki/Q83021","display_name":"Empirical evidence","level":2,"score":0.27559998631477356},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.2621000111103058},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.26170000433921814},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.259799987077713},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.25209999084472656},{"id":"https://openalex.org/C2780148112","wikidata":"https://www.wikidata.org/wiki/Q1432581","display_name":"Proxy (statistics)","level":2,"score":0.2515000104904175},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.25029999017715454}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.11119","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.11119","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.11119","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.11119","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"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"},"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":{"Reinforcement":[0],"learning":[1],"with":[2,119],"verifiable":[3],"rewards":[4,44],"(RLVR)":[5],"is":[6,24],"a":[7,54,104,136,146],"promising":[8,67],"approach":[9],"for":[10,132],"enhancing":[11],"reasoning":[12],"and":[13,41,79,159,197,207],"agentic":[14,95,212],"behavior":[15],"in":[16,53,82],"large":[17],"language":[18],"models.":[19],"However,":[20],"rollout-intensive":[21],"policy":[22],"optimization":[23],"often":[25],"limited":[26],"by":[27,97,221],"insufficient":[28],"reward":[29,143],"contrast,":[30],"arising":[31],"when":[32,42],"overly":[33],"simple":[34],"or":[35],"complex":[36],"prompts":[37],"generate":[38],"low-variance":[39],"feedback":[40,196],"outcome-only":[43,195],"assign":[45],"the":[46,76,88,199],"same":[47,89],"terminal":[48,169],"assessment":[49],"to":[50,66,111,116,155,166,185],"every":[51],"decision":[52],"multi-turn":[55,94],"rollout.":[56,90],"Past":[57],"efforts":[58],"have":[59],"focused":[60],"on":[61,210],"allocating":[62],"available":[63],"rollout":[64,138,153],"resources":[65],"prompts,":[68],"yet":[69],"they":[70],"only":[71],"leverage":[72],"sample":[73],"informativeness":[74,84],"at":[75,179,227],"prompt":[77,114,157],"level":[78],"neglect":[80],"variation":[81],"prefix-level":[83],"across":[85],"turns":[86],"within":[87,145],"This":[91],"work":[92],"targets":[93],"RL":[96],"modeling":[98],"each":[99],"ReAct-style":[100],"thought-action-observation":[101],"turn":[102],"as":[103],"semantically":[105],"distinct":[106],"node,":[107],"allowing":[108],"budget":[109,154],"allocation":[110,139],"extend":[112],"from":[113,182],"roots":[115,158],"turn-level":[117],"prefixes":[118,161],"further":[120],"continuations,":[121],"which":[122],"naturally":[123],"forms":[124],"tree-structured":[125],"rollouts.":[126],"We":[127],"introduce":[128],"Tree":[129],"Rollout":[130],"Allocation":[131],"Contrastive":[133],"Exploration":[134],"(TRACE),":[135],"unified":[137],"framework":[140],"that":[141,162],"enhances":[142],"contrast":[144],"fixed":[147],"sampling":[148,229],"budget.":[149],"Technically,":[150],"TRACE":[151,203],"allocates":[152],"both":[156],"intermediate":[160],"are":[163],"most":[164],"likely":[165],"yield":[167],"mixed":[168],"rewards.":[170],"A":[171],"shared":[172],"generalizable":[173],"predictor":[174],"estimates":[175],"conditional":[176],"success":[177],"probability":[178],"these":[180],"anchors":[181],"prefix":[183],"histories":[184],"guide":[186],"this":[187],"allocation.":[188],"The":[189],"resulting":[190],"adaptive":[191],"tree":[192],"structure":[193],"enriches":[194],"amplifies":[198],"policy-update":[200],"signal.":[201],"Empirically,":[202],"achieves":[204],"competitive":[205,225],"performance":[206],"efficiency":[208],"gains":[209],"typical":[211],"benchmarks,":[213],"e.g.,":[214],"improving":[215],"Qwen3-14B":[216],"Multi-Hop":[217],"QA":[218],"average":[219],"accuracy":[220],"2.8":[222],"points":[223],"over":[224],"baselines":[226],"equal":[228],"cost.":[230]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-11T00:00:00"}