{"id":"https://openalex.org/W7171549984","doi":"https://doi.org/10.48550/arxiv.2607.23700","title":"Offline-Online Curriculum RL for Multimodal Reasoning","display_name":"Offline-Online Curriculum RL for Multimodal Reasoning","publication_year":2026,"publication_date":"2026-07-26","ids":{"openalex":"https://openalex.org/W7171549984","doi":"https://doi.org/10.48550/arxiv.2607.23700"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.23700","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.23700","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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.2607.23700","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136069865","display_name":"Wendi Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deng, Wendi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143847637","display_name":"Hang Du","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Du, Hang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020360628","display_name":"Guoshun Nan","orcid":"https://orcid.org/0000-0002-1987-2736"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nan, Guoshun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107126595","display_name":"Haokun Tian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tian, Haokun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143745611","display_name":"Jiaqi Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu, Jiaqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000377914","display_name":"Xinlei Cao","orcid":"https://orcid.org/0000-0001-6854-2154"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cao, Xinlei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143746350","display_name":"Jaile Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Jaile","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143844463","display_name":"Jingfeng Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Jingfeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101130399","display_name":"Ling Deng","orcid":"https://orcid.org/0009-0002-5607-8192"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deng, Ling","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143831581","display_name":"Ting Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Ting","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143811948","display_name":"Hao Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Hao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143794500","display_name":"Jun Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Jun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143839150","display_name":"Xudong Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Xudong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5143819986","display_name":"Sicong Leng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Leng, Sicong","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.6263999938964844,"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"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.6263999938964844,"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/T10028","display_name":"Topic Modeling","score":0.11509999632835388,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.053700000047683716,"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/interpretability","display_name":"Interpretability","score":0.8037999868392944},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5697000026702881},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5479000210762024},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.5314000248908997},{"id":"https://openalex.org/keywords/spurious-relationship","display_name":"Spurious relationship","score":0.48750001192092896},{"id":"https://openalex.org/keywords/filter","display_name":"Filter (signal processing)","score":0.4471000134944916},{"id":"https://openalex.org/keywords/abstraction","display_name":"Abstraction","score":0.43459999561309814}],"concepts":[{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.8037999868392944},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7488999962806702},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5697000026702881},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5681999921798706},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5479000210762024},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.5314000248908997},{"id":"https://openalex.org/C97256817","wikidata":"https://www.wikidata.org/wiki/Q1462316","display_name":"Spurious relationship","level":2,"score":0.48750001192092896},{"id":"https://openalex.org/C106131492","wikidata":"https://www.wikidata.org/wiki/Q3072260","display_name":"Filter (signal processing)","level":2,"score":0.4471000134944916},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.43459999561309814},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.39079999923706055},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.3424000144004822},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.2996000051498413},{"id":"https://openalex.org/C155911762","wikidata":"https://www.wikidata.org/wiki/Q422321","display_name":"Blueprint","level":2,"score":0.287200003862381},{"id":"https://openalex.org/C47177190","wikidata":"https://www.wikidata.org/wiki/Q207137","display_name":"Curriculum","level":2,"score":0.2833999991416931},{"id":"https://openalex.org/C2780660688","wikidata":"https://www.wikidata.org/wiki/Q25052564","display_name":"Multimodal learning","level":2,"score":0.2759999930858612},{"id":"https://openalex.org/C2779982483","wikidata":"https://www.wikidata.org/wiki/Q6094420","display_name":"Iterative refinement","level":2,"score":0.27070000767707825},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.26100000739097595},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.26100000739097595},{"id":"https://openalex.org/C9357733","wikidata":"https://www.wikidata.org/wiki/Q6878417","display_name":"Missing data","level":2,"score":0.25459998846054077}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.23700","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.23700","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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.2607.23700","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.23700","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":0.773186206817627}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Multimodal":[0],"large":[1],"language":[2],"models":[3],"exhibit":[4],"capabilities":[5],"on":[6,28,125,137],"reasoning":[7,88,139],"tasks,":[8],"yet":[9],"often":[10],"produce":[11],"flawed":[12],"intermediate":[13],"steps":[14,43,89,116,127],"while":[15,148],"yielding":[16],"correct":[17],"final":[18],"answers.":[19],"This":[20],"behavior":[21],"undermines":[22],"interpretability":[23],"and":[24,90,117,128,152],"reliability,":[25],"suggesting":[26],"reliance":[27],"spurious":[29],"shortcuts":[30],"rather":[31],"than":[32],"faithful":[33],"reasoning.":[34],"Although":[35],"efforts":[36],"have":[37],"explored":[38],"step-level":[39,80,103],"supervision,":[40],"distinguishing":[41],"decisive":[42],"from":[44],"redundant":[45,93],"ones":[46],"remains":[47],"challenging.":[48],"We":[49],"propose":[50],"$O^2$-CritiCuRL,":[51],"a":[52,101],"novel":[53],"curriculum":[54],"reinforcement":[55,104],"learning":[56,105],"framework":[57,84],"that":[58,142],"introduces":[59],"critical-step":[60],"awareness":[61],"through":[62],"an":[63],"iterative":[64],"offline-online":[65],"paradigm.":[66],"In":[67,95],"the":[68,83,96,111,130],"offline":[69],"stage,":[70,98],"$O^2$-CritiCuRL":[71],"conducts":[72],"multi-rollout":[73],"analysis":[74],"over":[75],"step-annotated":[76],"trajectories":[77],"to":[78,85,113],"estimate":[79],"importance,":[81],"allowing":[82],"distill":[86],"critical":[87,126],"filter":[91],"out":[92],"ones.":[94],"online":[97],"we":[99],"employ":[100],"progressive":[102],"strategy,":[106],"where":[107],"truncated":[108],"chains":[109],"guide":[110],"model":[112],"infer":[114],"missing":[115],"refine":[118],"its":[119,123],"reasoning,":[120],"thereby":[121],"sharpening":[122],"focus":[124],"overcoming":[129],"limitations":[131],"of":[132],"static":[133],"supervision.":[134],"Extensive":[135],"experiments":[136],"multimodal":[138],"benchmarks":[140],"show":[141],"our":[143],"method":[144],"achieves":[145],"state-of-the-art":[146],"performance":[147],"delivering":[149],"superior":[150],"training":[151],"inference":[153],"efficiency.":[154],"Code":[155],"is":[156],"available":[157],"at":[158],"https://github.com/kk0013/CritiCuRL.":[159]},"counts_by_year":[],"updated_date":"2026-07-29T09:40:50.615796","created_date":"2026-07-29T00:00:00"}