{"id":"https://openalex.org/W7154625525","doi":"https://doi.org/10.48550/arxiv.2604.13540","title":"Free Lunch for Unified Multimodal Models: Enhancing Generation via Reflective Rectification with Inherent Understanding","display_name":"Free Lunch for Unified Multimodal Models: Enhancing Generation via Reflective Rectification with Inherent Understanding","publication_year":2026,"publication_date":"2026-04-15","ids":{"openalex":"https://openalex.org/W7154625525","doi":"https://doi.org/10.48550/arxiv.2604.13540"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.13540","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.13540","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.2604.13540","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5133817270","display_name":"Yibo Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Jiang, Yibo","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133750913","display_name":"Tao Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Tao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133792524","display_name":"Rui Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Rui","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113394962","display_name":"Yehao Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Yehao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101675201","display_name":"Chaoxiang Cai","orcid":"https://orcid.org/0000-0003-3322-4273"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cai, Chaoxiang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058879520","display_name":"Zequn Qin","orcid":"https://orcid.org/0000-0002-1632-8313"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qin, Zequn","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133745856","display_name":"Xi Li","orcid":"https://orcid.org/0009-0008-0832-7530"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Xi","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5133817270"],"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":0.35100001096725464,"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":0.35100001096725464,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.07670000195503235,"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/T11883","display_name":"Embodied and Extended Cognition","score":0.07119999825954437,"subfield":{"id":"https://openalex.org/subfields/2805","display_name":"Cognitive Neuroscience"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/rectification","display_name":"Rectification","score":0.8082000017166138},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.7009000182151794},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.5435000061988831},{"id":"https://openalex.org/keywords/signal","display_name":"SIGNAL (programming language)","score":0.3517000079154968},{"id":"https://openalex.org/keywords/visual-reasoning","display_name":"Visual reasoning","score":0.3312999904155731},{"id":"https://openalex.org/keywords/unified-model","display_name":"Unified Model","score":0.31470000743865967}],"concepts":[{"id":"https://openalex.org/C50942859","wikidata":"https://www.wikidata.org/wiki/Q4967193","display_name":"Rectification","level":3,"score":0.8082000017166138},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7102000117301941},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.7009000182151794},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.5435000061988831},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.48559999465942383},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.43540000915527344},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.3517000079154968},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.3312999904155731},{"id":"https://openalex.org/C45493050","wikidata":"https://www.wikidata.org/wiki/Q7884934","display_name":"Unified Model","level":2,"score":0.31470000743865967},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.30709999799728394},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.27230000495910645},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.27149999141693115}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.13540","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.13540","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.2604.13540","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.13540","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":[{"display_name":"Quality Education","score":0.43198826909065247,"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":{"Unified":[0],"Multimodal":[1],"Models":[2],"(UMMs)":[3],"aim":[4],"to":[5,64,97,141],"integrate":[6],"visual":[7,120],"understanding":[8,25,43,96],"and":[9,68,104,123],"generation":[10,158],"within":[11],"a":[12,19,78,138],"single":[13],"structure.":[14],"However,":[15],"these":[16],"models":[17],"exhibit":[18],"notable":[20],"capability":[21,26],"mismatch,":[22],"where":[23,60],"their":[24,29,66],"significantly":[27,156],"outperforms":[28],"generation.":[30,48],"This":[31],"mismatch":[32],"indicates":[33],"that":[34,147],"the":[35,56,87,92,111,125,129,134],"model's":[36],"rich":[37],"internal":[38,102],"knowledge,":[39],"while":[40],"effective":[41],"for":[42],"tasks,":[44],"remains":[45],"underactivated":[46],"during":[47,108],"To":[49],"address":[50],"this,":[51],"we":[52,75],"draw":[53],"inspiration":[54],"from":[55],"human":[57],"``Thinking-While-Drawing''":[58],"paradigm,":[59],"humans":[61],"continuously":[62,98],"reflect":[63],"activate":[65],"knowledge":[67,103],"rectify":[69,142],"intermediate":[70,106,126],"results.":[71],"In":[72],"this":[73],"paper,":[74],"propose":[76],"UniRect-CoT,":[77],"training-free":[79],"unified":[80],"rectification":[81],"chain-of-thought":[82],"framework.":[83],"Our":[84],"approach":[85],"unlocks":[86],"``free":[88],"lunch''":[89],"hidden":[90],"in":[91,115],"UMM's":[93],"powerful":[94],"inherent":[95],"reflect,":[99],"activating":[100],"its":[101],"rectifying":[105],"results":[107,127],"generation.We":[109],"regard":[110],"diffusion":[112],"denoising":[113],"process":[114,122],"UMMs":[116],"as":[117,137],"an":[118],"intrinsic":[119],"reasoning":[121],"align":[124],"with":[128],"target":[130],"instruction":[131],"understood":[132],"by":[133],"model,":[135],"serving":[136],"self-supervisory":[139],"signal":[140],"UMM":[143],"generation.Extensive":[144],"experiments":[145],"demonstrate":[146],"UniRect-CoT":[148],"can":[149],"be":[150],"easily":[151],"integrated":[152],"into":[153],"existing":[154],"UMMs,":[155],"enhancing":[157],"quality":[159],"across":[160],"diverse":[161],"complex":[162],"tasks.":[163]},"counts_by_year":[],"updated_date":"2026-04-17T06:04:52.305304","created_date":"2026-04-17T00:00:00"}