{"id":"https://openalex.org/W7166802065","doi":"https://doi.org/10.18653/v1/2026.findings-acl.724","title":"How to Allocate, How to Learn? Dynamic Rollout Allocation and Advantage Modulation for Policy Optimization","display_name":"How to Allocate, How to Learn? Dynamic Rollout Allocation and Advantage Modulation for Policy Optimization","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166802065","doi":"https://doi.org/10.18653/v1/2026.findings-acl.724"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.724","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.724","pdf_url":"https://aclanthology.org/2026.findings-acl.724.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.724.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139762830","display_name":"Yangyi Fang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yangyi Fang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139804640","display_name":"Jiaye Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiaye Lin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139716355","display_name":"Xiaoliang Fu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaoliang Fu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028740631","display_name":"Cong Qin","orcid":"https://orcid.org/0000-0002-1276-0985"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cong Qin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139769042","display_name":"Haolin Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Haolin Shi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139775111","display_name":"Chaowen Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chaowen Hu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139852782","display_name":"Lu Pan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu Pan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139839379","display_name":"Ke Zeng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ke Zeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139817245","display_name":"Xunliang Cai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xunliang Cai","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.85743146,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"14727","last_page":"14744"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10357","display_name":"Efficiency Analysis Using DEA","score":0.04230000078678131,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},"topics":[{"id":"https://openalex.org/T10357","display_name":"Efficiency Analysis Using DEA","score":0.04230000078678131,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T13825","display_name":"Economic Policies and Impacts","score":0.031700000166893005,"subfield":{"id":"https://openalex.org/subfields/2002","display_name":"Economics and Econometrics"},"field":{"id":"https://openalex.org/fields/20","display_name":"Economics, Econometrics and Finance"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T11043","display_name":"Merger and Competition Analysis","score":0.028699999675154686,"subfield":{"id":"https://openalex.org/subfields/2002","display_name":"Economics and Econometrics"},"field":{"id":"https://openalex.org/fields/20","display_name":"Economics, Econometrics and Finance"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/modulation","display_name":"Modulation (music)","score":0.4205000102519989},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.3239000141620636},{"id":"https://openalex.org/keywords/control","display_name":"Control (management)","score":0.31040000915527344},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.2989000082015991},{"id":"https://openalex.org/keywords/production","display_name":"Production (economics)","score":0.296099990606308},{"id":"https://openalex.org/keywords/resource-allocation","display_name":"Resource allocation","score":0.2750999927520752}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5486999750137329},{"id":"https://openalex.org/C123079801","wikidata":"https://www.wikidata.org/wiki/Q750240","display_name":"Modulation (music)","level":2,"score":0.4205000102519989},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3239000141620636},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.31040000915527344},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2989000082015991},{"id":"https://openalex.org/C2778348673","wikidata":"https://www.wikidata.org/wiki/Q739302","display_name":"Production (economics)","level":2,"score":0.296099990606308},{"id":"https://openalex.org/C29202148","wikidata":"https://www.wikidata.org/wiki/Q287260","display_name":"Resource allocation","level":2,"score":0.2750999927520752},{"id":"https://openalex.org/C11930861","wikidata":"https://www.wikidata.org/wiki/Q181417","display_name":"Frequency modulation","level":3,"score":0.2615000009536743},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.25920000672340393},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2513999938964844}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.724","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.724","pdf_url":"https://aclanthology.org/2026.findings-acl.724.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.724","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.724","pdf_url":"https://aclanthology.org/2026.findings-acl.724.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166802065.pdf","grobid_xml":"https://content.openalex.org/works/W7166802065.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reinforcement":[0],"Learning":[1],"with":[2],"Verifiable":[3],"Rewards":[4],"(RLVR)":[5],"has":[6],"proven":[7],"effective":[8],"for":[9,47,91,112],"Large":[10],"Language":[11],"Model":[12],"(LLM)":[13],"reasoning,":[14],"yet":[15],"current":[16],"methods":[17],"face":[18],"key":[19],"challenges":[20],"in":[21,103],"resource":[22],"allocation":[23,31,73,79],"and":[24,38,76],"policy":[25,42],"optimization":[26,64],"dynamics:":[27],"(i)":[28],"uniform":[29,72],"rollout":[30],"ignores":[32],"gradient":[33,45,53,92,107,113],"variance":[34,86],"heterogeneity":[35],"across":[36],"problems,":[37],"(ii)":[39],"the":[40,66,81,94],"softmax":[41],"structure":[43],"causes":[44],"attenuation":[46,114],"high-confidence":[48,116],"correct":[49,117],"actions,":[50],"while":[51,119],"excessive":[52,128],"updates":[54],"may":[55],"destabilize":[56],"training.Therefore,":[57],"we":[58,69,97],"propose":[59],"DynaMO,":[60],"a":[61,88,134],"theoretically-grounded":[62],"dual-pronged":[63],"framework.At":[65],"sequence":[67],"level,":[68,96],"prove":[70],"that":[71],"is":[74,149],"suboptimal":[75],"derive":[77],"variance-minimizing":[78],"from":[80],"first":[82],"principle,":[83],"establishing":[84],"Bernoulli":[85],"as":[87,123],"computable":[89,124],"proxy":[90],"informativeness.At":[93],"token":[95],"develop":[98],"gradientaware":[99],"advantage":[100],"modulation":[101],"grounded":[102],"theoretical":[104],"analysis":[105],"of":[106,115,137],"magnitude":[108],"bounds.Our":[109],"framework":[110],"compensates":[111],"actions":[118],"utilizing":[120],"entropy":[121],"changes":[122],"indicators":[125],"to":[126],"stabilize":[127],"update":[129],"magnitudes.Extensive":[130],"experiments":[131],"conducted":[132],"on":[133],"diverse":[135],"range":[136],"mathematical":[138],"reasoning":[139],"benchmarks":[140],"demonstrate":[141],"consistent":[142],"improvements":[143],"over":[144],"strong":[145],"RLVR":[146],"baselines.Our":[147],"implementation":[148],"available":[150],"at:":[151]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}