{"id":"https://openalex.org/W7163899384","doi":"https://doi.org/10.48550/arxiv.2606.06828","title":"AdaGRPO: A Capability-Aware Adaptive Enhancement for Flow-based GRPO","display_name":"AdaGRPO: A Capability-Aware Adaptive Enhancement for Flow-based GRPO","publication_year":2026,"publication_date":"2026-06-05","ids":{"openalex":"https://openalex.org/W7163899384","doi":"https://doi.org/10.48550/arxiv.2606.06828"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.06828","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.06828","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":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.06828","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138107823","display_name":"Jiazi Bu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bu, Jiazi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138148382","display_name":"Pengyang Ling","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ling, Pengyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138177850","display_name":"Yujie Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Yujie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138122142","display_name":"Yibin Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yibin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138146321","display_name":"Yuhang Zang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zang, Yuhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138181458","display_name":"Tianyi Wei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei, Tianyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5040988693","display_name":"Xiaohang Zhan","orcid":"https://orcid.org/0000-0003-2136-7592"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhan, Xiaohang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138162525","display_name":"Jiaqi Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Jiaqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138140088","display_name":"Tong Wu","orcid":"https://orcid.org/0000-0002-1920-5044"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Tong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138168086","display_name":"Xingang Pan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pan, Xingang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138107735","display_name":"Dahua Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Dahua","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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.266400009393692,"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":0.266400009393692,"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":0.10999999940395355,"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/T11704","display_name":"Mobile Crowdsensing and Crowdsourcing","score":0.06790000200271606,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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":0.6205000281333923},{"id":"https://openalex.org/keywords/sample","display_name":"Sample (material)","score":0.5662999749183655},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.5516999959945679},{"id":"https://openalex.org/keywords/principal","display_name":"Principal (computer security)","score":0.4374000132083893},{"id":"https://openalex.org/keywords/perspective","display_name":"Perspective (graphical)","score":0.41850000619888306},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.4180000126361847},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.3799000084400177},{"id":"https://openalex.org/keywords/meta-learning","display_name":"Meta learning (computer science)","score":0.37770000100135803},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.3702000081539154}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.729200005531311},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6205000281333923},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.5662999749183655},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.5516999959945679},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5149999856948853},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5095999836921692},{"id":"https://openalex.org/C144559511","wikidata":"https://www.wikidata.org/wiki/Q2986279","display_name":"Principal (computer security)","level":2,"score":0.4374000132083893},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.41850000619888306},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.4180000126361847},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.3799000084400177},{"id":"https://openalex.org/C2781002164","wikidata":"https://www.wikidata.org/wiki/Q6822311","display_name":"Meta learning (computer science)","level":3,"score":0.37770000100135803},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.3702000081539154},{"id":"https://openalex.org/C2780626000","wikidata":"https://www.wikidata.org/wiki/Q5936775","display_name":"Human-in-the-loop","level":2,"score":0.3449999988079071},{"id":"https://openalex.org/C184670325","wikidata":"https://www.wikidata.org/wiki/Q512604","display_name":"Loop (graph theory)","level":2,"score":0.33869999647140503},{"id":"https://openalex.org/C2986087404","wikidata":"https://www.wikidata.org/wiki/Q15946010","display_name":"Online learning","level":2,"score":0.33329999446868896},{"id":"https://openalex.org/C2781039887","wikidata":"https://www.wikidata.org/wiki/Q1391724","display_name":"Factor (programming language)","level":2,"score":0.3240000009536743},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.31940001249313354},{"id":"https://openalex.org/C38349280","wikidata":"https://www.wikidata.org/wiki/Q1434290","display_name":"Flow (mathematics)","level":2,"score":0.3188999891281128},{"id":"https://openalex.org/C2779136372","wikidata":"https://www.wikidata.org/wiki/Q10283002","display_name":"Information flow","level":2,"score":0.3100999891757965},{"id":"https://openalex.org/C186886427","wikidata":"https://www.wikidata.org/wiki/Q5441213","display_name":"Feedback loop","level":2,"score":0.2957000136375427},{"id":"https://openalex.org/C148043351","wikidata":"https://www.wikidata.org/wiki/Q4456944","display_name":"Current (fluid)","level":2,"score":0.2696000039577484},{"id":"https://openalex.org/C77967617","wikidata":"https://www.wikidata.org/wiki/Q4677561","display_name":"Active learning (machine learning)","level":2,"score":0.2639000117778778},{"id":"https://openalex.org/C2778334786","wikidata":"https://www.wikidata.org/wiki/Q1586270","display_name":"Variation (astronomy)","level":2,"score":0.26350000500679016},{"id":"https://openalex.org/C2779436431","wikidata":"https://www.wikidata.org/wiki/Q30672407","display_name":"Policy learning","level":2,"score":0.257999986410141}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.06828","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.06828","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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.06828","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.06828","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":"Preprint"},"sustainable_development_goals":[{"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4","score":0.429476797580719}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Group":[0],"Relative":[1],"Policy":[2],"Optimization":[3],"(GRPO)":[4],"has":[5],"demonstrated":[6],"remarkable":[7],"success":[8],"in":[9,71],"aligning":[10],"text-to-image":[11],"(T2I)":[12],"flow":[13,113,197],"models":[14],"with":[15,152,173],"human":[16],"preferences.":[17],"However,":[18],"we":[19,101],"have":[20],"identified":[21],"that":[22,136,185],"the":[23,34,57,129],"learning":[24,65,141],"loop":[25],"of":[26,60,118],"current":[27,36,140],"flow-based":[28],"GRPO":[29,72,104,194],"is":[30],"fundamentally":[31],"decoupled":[32],"from":[33,39],"learner's":[35],"capability,":[37],"suffering":[38],"critical":[40],"blind":[41],"spots":[42],"at":[43],"both":[44],"prompt":[45],"selection":[46,62],"and":[47,132,159,180],"advantage":[48],"estimation:":[49],"(i)":[50,122],"Existing":[51],"methods":[52],"sample":[53,80],"prompts":[54,135],"randomly,":[55],"overlooking":[56],"substantial":[58],"impact":[59],"data":[61],"on":[63,84],"reinforcement":[64],"(RL)":[66],"efficacy--a":[67],"factor":[68],"proven":[69],"crucial":[70],"for":[73,112,196],"large":[74],"language":[75],"models;":[76],"(ii)":[77,143],"They":[78],"evaluate":[79],"quality":[81],"solely":[82],"relying":[83],"intra-group":[85,150],"statistics,":[86],"lacking":[87],"a":[88,106,157,164],"global":[89,154],"perspective":[90],"to":[91],"accurately":[92],"measure":[93],"true":[94],"policy":[95,161],"improvement.":[96],"To":[97],"address":[98],"these":[99],"issues,":[100],"propose":[102],"Adaptive":[103],"(AdaGRPO),":[105],"novel":[107],"capability-aware":[108],"RL":[109],"algorithm":[110],"tailored":[111],"models.":[114,198],"Specifically,":[115],"AdaGRPO":[116,168,186],"consists":[117],"two":[119],"principal":[120],"components:":[121],"Online":[123],"Curriculum":[124],"Filtering":[125],"Strategy:":[126],"Dynamically":[127],"tracks":[128],"model's":[130],"proficiency":[131],"adaptively":[133],"selects":[134],"best":[137],"match":[138],"its":[139],"boundary;":[142],"Cross-Level":[144],"Advantage":[145],"Fusion:":[146],"Synergistically":[147],"integrates":[148],"fine-grained":[149],"advantages":[151],"macro-level":[153],"advantages,":[155],"providing":[156],"comprehensive":[158],"unbiased":[160],"evaluation.":[162],"As":[163],"lightweight,":[165],"plug-and-play":[166],"module,":[167],"can":[169],"be":[170],"seamlessly":[171],"integrated":[172],"existing":[174],"frameworks":[175],"such":[176],"as":[177],"Flow-GRPO,":[178],"DanceGRPO,":[179],"Flow-CPS.":[181],"Extensive":[182],"experiments":[183],"demonstrate":[184],"consistently":[187],"drives":[188],"performance":[189],"gains":[190],"while":[191],"significantly":[192],"stabilizes":[193],"training":[195]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-09T00:00:00"}