{"id":"https://openalex.org/W7140234804","doi":"https://doi.org/10.48550/arxiv.2603.21289","title":"When Models Judge Themselves: Unsupervised Self-Evolution for Multimodal Reasoning","display_name":"When Models Judge Themselves: Unsupervised Self-Evolution for Multimodal Reasoning","publication_year":2026,"publication_date":"2026-03-22","ids":{"openalex":"https://openalex.org/W7140234804","doi":"https://doi.org/10.48550/arxiv.2603.21289"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.21289","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.21289","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.2603.21289","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Wu, Zhengxian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Zhengxian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Shi, Kai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Kai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zhang, Chuanrui","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Chuanrui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Liao, Zirui","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liao, Zirui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yang, Jun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Jun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yang, Ni","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Ni","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Peng, Qiuying","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peng, Qiuying","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zhang, Luyuan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Luyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Xu, Hangrui","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Hangrui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Su, Tianhuang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Su, Tianhuang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yang, Zhenyu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Zhenyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Lu, Haonan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Haonan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Wang, Haoqian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Haoqian","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":0.2280000001192093,"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":0.2280000001192093,"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":0.21160000562667847,"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/T11902","display_name":"Intelligent Tutoring Systems and Adaptive Learning","score":0.11379999667406082,"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/generalization","display_name":"Generalization","score":0.6929000020027161},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.4853000044822693},{"id":"https://openalex.org/keywords/sample","display_name":"Sample (material)","score":0.4244000017642975},{"id":"https://openalex.org/keywords/path","display_name":"Path (computing)","score":0.4000000059604645},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.3765999972820282},{"id":"https://openalex.org/keywords/bounded-function","display_name":"Bounded function","score":0.36739999055862427},{"id":"https://openalex.org/keywords/aggregate","display_name":"Aggregate (composite)","score":0.336899995803833}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7444999814033508},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6929000020027161},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6697999835014343},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.573199987411499},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.4853000044822693},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.4244000017642975},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.4000000059604645},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3765999972820282},{"id":"https://openalex.org/C34388435","wikidata":"https://www.wikidata.org/wiki/Q2267362","display_name":"Bounded function","level":2,"score":0.36739999055862427},{"id":"https://openalex.org/C4679612","wikidata":"https://www.wikidata.org/wiki/Q866298","display_name":"Aggregate (composite)","level":2,"score":0.336899995803833},{"id":"https://openalex.org/C2781311116","wikidata":"https://www.wikidata.org/wiki/Q83306","display_name":"Group (periodic table)","level":2,"score":0.31130000948905945},{"id":"https://openalex.org/C89288958","wikidata":"https://www.wikidata.org/wiki/Q7301504","display_name":"Reasoning system","level":2,"score":0.30869999527931213},{"id":"https://openalex.org/C20162079","wikidata":"https://www.wikidata.org/wiki/Q1151406","display_name":"Case-based reasoning","level":2,"score":0.3005000054836273},{"id":"https://openalex.org/C2780910867","wikidata":"https://www.wikidata.org/wiki/Q1952416","display_name":"Multimodality","level":2,"score":0.29330000281333923},{"id":"https://openalex.org/C83725634","wikidata":"https://www.wikidata.org/wiki/Q7268699","display_name":"Qualitative reasoning","level":2,"score":0.28769999742507935},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.2750999927520752},{"id":"https://openalex.org/C8038995","wikidata":"https://www.wikidata.org/wiki/Q1152135","display_name":"Unsupervised learning","level":2,"score":0.271699994802475},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.25769999623298645}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.21289","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.21289","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.2603.21289","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.21289","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":[{"score":0.48519039154052734,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"progress":[1],"in":[2],"multimodal":[3,47,156],"large":[4],"language":[5],"models":[6],"has":[7],"led":[8],"to":[9,34,94],"strong":[10],"performance":[11,52,142],"on":[12,20,134,145],"reasoning":[13,48,68,141,148],"tasks,":[14],"but":[15],"these":[16],"improvements":[17,53],"largely":[18],"rely":[19],"high-quality":[21],"annotated":[22],"data":[23],"or":[24,58],"teacher-model":[25],"distillation,":[26],"both":[27],"of":[28,98],"which":[29],"are":[30,160],"costly":[31],"and":[32,70,87,112,143],"difficult":[33],"scale.":[35],"To":[36],"address":[37],"this,":[38],"we":[39,65],"propose":[40],"an":[41],"unsupervised":[42],"self-evolution":[43],"training":[44,85],"framework":[45],"for":[46],"that":[49],"achieves":[50],"stable":[51],"without":[54],"using":[55],"human-annotated":[56],"answers":[57],"external":[59],"reward":[60],"models.":[61,157],"For":[62],"each":[63,120],"input,":[64],"sample":[66],"multiple":[67],"trajectories":[69,97],"jointly":[71],"model":[72,103],"their":[73],"within":[74,119],"group":[75,109],"structure.":[76],"We":[77,101],"use":[78],"the":[79,104],"Actor's":[80],"self-consistency":[81],"signal":[82],"as":[83,107],"a":[84,89,108,151],"prior,":[86],"introduce":[88],"bounded":[90],"Judge":[91],"based":[92],"modulation":[93],"continuously":[95],"reweight":[96],"different":[99],"quality.":[100],"further":[102],"modulated":[105],"scores":[106,115],"level":[110],"distribution":[111],"convert":[113],"absolute":[114],"into":[116],"relative":[117],"advantages":[118],"group,":[121],"enabling":[122],"more":[123],"robust":[124],"policy":[125],"updates.":[126],"Trained":[127],"with":[128],"Group":[129],"Relative":[130],"Policy":[131],"Optimization":[132],"(GRPO)":[133],"unlabeled":[135],"data,":[136],"our":[137],"method":[138],"consistently":[139],"improves":[140],"generalization":[144],"five":[146],"mathematical":[147],"benchmarks,":[149],"offering":[150],"scalable":[152],"path":[153],"toward":[154],"self-evolving":[155],"The":[158],"code":[159],"available":[161],"at":[162],"https://github.com/OPPO-Mente-Lab/LLM-Self-Judge.":[163]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-25T00:00:00"}