{"id":"https://openalex.org/W7161029974","doi":"https://doi.org/10.48550/arxiv.2605.11403","title":"fg-expo: Frontier-guided exploration-prioritized policy optimization via adaptive kl and gaussian curriculum","display_name":"fg-expo: Frontier-guided exploration-prioritized policy optimization via adaptive kl and gaussian curriculum","publication_year":2026,"publication_date":"2026-05-12","ids":{"openalex":"https://openalex.org/W7161029974","doi":"https://doi.org/10.48550/arxiv.2605.11403"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.11403","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.11403","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.11403","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135915021","display_name":"Mingxiong Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Lin, Mingxiong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135913516","display_name":"Zhangquan Gong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gong, Zhangquan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135958173","display_name":"Maowen Tang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tang, Maowen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136032477","display_name":"Qian Li","orcid":"https://orcid.org/0000-0001-7903-2838"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Qian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135997517","display_name":"Chuangchuang Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Chuangchuang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136035841","display_name":"Jian Ma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Jian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136040812","display_name":"Sutian Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Sutian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136007914","display_name":"Kai Tang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tang, Kai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136039906","display_name":"Haonan Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Haonan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":9,"corresponding_author_ids":["https://openalex.org/A5135915021"],"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.25589999556541443,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.25589999556541443,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.07819999754428864,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.05429999902844429,"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/benchmark","display_name":"Benchmark (surveying)","score":0.5989999771118164},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5960999727249146},{"id":"https://openalex.org/keywords/gaussian","display_name":"Gaussian","score":0.542900025844574},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.5339999794960022},{"id":"https://openalex.org/keywords/constraint","display_name":"Constraint (computer-aided design)","score":0.4966999888420105},{"id":"https://openalex.org/keywords/importance-sampling","display_name":"Importance sampling","score":0.48980000615119934},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.4300999939441681},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4219000041484833}],"concepts":[{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5989999771118164},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5960999727249146},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.5683000087738037},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.542900025844574},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5389999747276306},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.5339999794960022},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.4966999888420105},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.48980000615119934},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.4300999939441681},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4219000041484833},{"id":"https://openalex.org/C2781395549","wikidata":"https://www.wikidata.org/wiki/Q4680762","display_name":"Adaptive sampling","level":3,"score":0.41519999504089355},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.40709999203681946},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3871000111103058},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.37139999866485596},{"id":"https://openalex.org/C61326573","wikidata":"https://www.wikidata.org/wiki/Q1496376","display_name":"Gaussian process","level":3,"score":0.3709999918937683},{"id":"https://openalex.org/C73602740","wikidata":"https://www.wikidata.org/wiki/Q7795822","display_name":"Thompson sampling","level":3,"score":0.36739999055862427},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.365200012922287},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.3612000048160553},{"id":"https://openalex.org/C158622935","wikidata":"https://www.wikidata.org/wiki/Q660848","display_name":"Nonlinear system","level":2,"score":0.3522000014781952},{"id":"https://openalex.org/C115527620","wikidata":"https://www.wikidata.org/wiki/Q769909","display_name":"Nonlinear programming","level":3,"score":0.33090001344680786},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.33079999685287476},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.32710000872612},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.3009999990463257},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.28299999237060547},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.25850000977516174},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.25850000977516174},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.2563999891281128}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.11403","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.11403","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.11403","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.11403","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"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":[2,15],"Verifiable":[3],"Rewards":[4],"(RLVR)":[5],"has":[6],"become":[7],"the":[8,23,46,53,66,90,104,107,115,178,199,217],"standard":[9],"paradigm":[10],"for":[11,75],"LLM":[12],"mathematical":[13,158],"reasoning,":[14],"Group":[16],"Relative":[17],"Policy":[18,78],"Optimization":[19],"(GRPO)":[20],"serving":[21],"as":[22],"dominant":[24],"algorithm.":[25],"We":[26,71,148],"identify":[27],"two":[28,82],"overlooked":[29],"inefficiencies":[30],"inherent":[31],"in":[32],"GRPO.":[33,169],"First,":[34],"a":[35,95,130,135,223],"fixed":[36,224],"KL":[37,86,91],"coefficient":[38],"overly":[39],"restricts":[40],"policy":[41],"exploration":[42,220],"at":[43,134],"moments":[44],"when":[45,106,114],"model":[47,108,116,142],"needs":[48],"to":[49,127,187,211],"diverge":[50],"significantly":[51],"from":[52,184],"reference":[54],"policy.":[55],"Second,":[56],"uniform":[57],"question":[58],"sampling":[59,125],"overlooks":[60],"that":[61,164,214],"moderately":[62],"difficult":[63],"problems":[64],"produce":[65],"most":[67],"informative":[68],"gradient":[69],"signals.":[70],"propose":[72],"FG-ExPO,":[73],"short":[74],"Frontier-Guided":[76],"Exploration-Prioritized":[77],"Optimization,":[79],"which":[80],"integrates":[81],"lightweight":[83],"components.":[84],"Accuracy-Conditioned":[85],"Scaling":[87],"(AKL)":[88],"adjusts":[89],"penalty":[92],"strength":[93],"through":[94],"smooth":[96],"nonlinear":[97],"function":[98],"of":[99,175,196],"batch":[100],"average":[101,193],"accuracy,":[102],"loosening":[103],"constraint":[105],"performs":[109],"poorly":[110],"and":[111,153,190],"strengthening":[112],"it":[113],"achieves":[117],"satisfactory":[118],"results.":[119],"Gaussian":[120,131],"Curriculum":[121],"Sampling":[122],"(GCS)":[123],"assigns":[124],"weights":[126],"questions":[128],"following":[129],"distribution":[132],"centered":[133],"moderate":[136],"accuracy":[137],"level":[138],"around":[139],"0.5,":[140],"focusing":[141],"training":[143],"on":[144,151,177,198,208],"its":[145],"learning":[146],"frontier.":[147],"conduct":[149],"evaluations":[150],"DeepSeek-R1-Distill-Qwen-1.5B":[152],"Qwen3-8B-Base":[154],"across":[155],"six":[156],"mainstream":[157],"reasoning":[159],"benchmarks.":[160],"Experimental":[161],"results":[162],"demonstrate":[163],"FG-ExPO":[165,215],"consistently":[166],"outperforms":[167],"vanilla":[168],"It":[170],"delivers":[171],"an":[172,192],"absolute":[173],"improvement":[174],"13.34":[176],"AIME":[179],"2025":[180],"pass@32":[181,194,209],"metric,":[182],"rising":[183],"63.33":[185],"percent":[186],"76.67":[188],"percent,":[189],"obtains":[191],"gain":[195],"2.66":[197],"8B":[200],"model.":[201],"The":[202],"substantially":[203],"larger":[204],"performance":[205],"gains":[206],"observed":[207],"compared":[210],"pass@1":[212],"verify":[213],"enlarges":[216],"model's":[218],"effective":[219],"space":[221],"under":[222],"inference":[225],"budget.":[226]},"counts_by_year":[],"updated_date":"2026-05-14T06:16:12.342656","created_date":"2026-05-14T00:00:00"}