{"id":"https://openalex.org/W7164937413","doi":"https://doi.org/10.48550/arxiv.2606.16700","title":"Multi-Turn Reflective Masking Elicits Reasoning in Mask Diffusion Models","display_name":"Multi-Turn Reflective Masking Elicits Reasoning in Mask Diffusion Models","publication_year":2026,"publication_date":"2026-06-15","ids":{"openalex":"https://openalex.org/W7164937413","doi":"https://doi.org/10.48550/arxiv.2606.16700"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.16700","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.16700","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.2606.16700","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138699393","display_name":"Yanming Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yanming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031473965","display_name":"Yihan Bian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bian, Yihan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065738809","display_name":"Jingyuan Qi","orcid":"https://orcid.org/0009-0000-2747-2903"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qi, Jingyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138740774","display_name":"Yuguang Yao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yao, Yuguang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138718735","display_name":"Lifu Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Lifu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138734994","display_name":"Tianyi Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Tianyi","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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.510200023651123,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.510200023651123,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.1128000020980835,"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":0.03319999948143959,"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/masking","display_name":"Masking (illustration)","score":0.888700008392334},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.6098999977111816},{"id":"https://openalex.org/keywords/mechanism","display_name":"Mechanism (biology)","score":0.4749999940395355},{"id":"https://openalex.org/keywords/property","display_name":"Property (philosophy)","score":0.4740999937057495},{"id":"https://openalex.org/keywords/noise-reduction","display_name":"Noise reduction","score":0.4058000147342682},{"id":"https://openalex.org/keywords/visual-masking","display_name":"Visual masking","score":0.3758000135421753},{"id":"https://openalex.org/keywords/diffusion","display_name":"Diffusion","score":0.37139999866485596}],"concepts":[{"id":"https://openalex.org/C2777402240","wikidata":"https://www.wikidata.org/wiki/Q6783436","display_name":"Masking (illustration)","level":2,"score":0.888700008392334},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7767000198364258},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.6098999977111816},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.4749999940395355},{"id":"https://openalex.org/C189950617","wikidata":"https://www.wikidata.org/wiki/Q937228","display_name":"Property (philosophy)","level":2,"score":0.4740999937057495},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4629000127315521},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4081999957561493},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.4058000147342682},{"id":"https://openalex.org/C2779200073","wikidata":"https://www.wikidata.org/wiki/Q18395575","display_name":"Visual masking","level":4,"score":0.3758000135421753},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.37139999866485596},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.36970001459121704},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3553999960422516},{"id":"https://openalex.org/C2776817677","wikidata":"https://www.wikidata.org/wiki/Q4839818","display_name":"Backward masking","level":3,"score":0.3458000123500824},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.32739999890327454},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.3257000148296356},{"id":"https://openalex.org/C2779982483","wikidata":"https://www.wikidata.org/wiki/Q6094420","display_name":"Iterative refinement","level":2,"score":0.3165999948978424},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.28459998965263367},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.2759000062942505},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.2621999979019165}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.16700","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.16700","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.2606.16700","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.16700","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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"While":[0,62],"reasoning":[1,11,96,190],"on":[2,21,47,120,191],"autoregressive":[3],"(AR)":[4],"models":[5],"is":[6,154],"often":[7],"performed":[8],"by":[9,73],"chain-of-thought":[10],"and":[12,57,84,114,153,163,169,179],"reflection,":[13],"their":[14],"refinement":[15,52],"of":[16],"previous":[17,48,55,127],"outputs":[18,118],"still":[19],"relies":[20],"fully":[22],"sequential":[23],"generation,":[24,167],"even":[25],"when":[26],"only":[27],"local":[28,45,75],"edits":[29,46],"are":[30],"needed.":[31],"In":[32],"contrast,":[33],"the":[34],"masking":[35,83],"mechanism":[36,139],"in":[37,98],"Mask":[38],"Diffusion":[39],"Models":[40],"(MDMs)":[41],"naturally":[42],"supports":[43],"explicit":[44],"outputs,":[49],"allowing":[50],"selective":[51],"without":[53],"discarding":[54],"answers":[56],"generating":[58],"another":[59],"from":[60,126],"scratch.":[61],"this":[63],"property":[64],"more":[65],"closely":[66],"aligns":[67],"with":[68],"how":[69],"humans":[70],"correct":[71],"mistakes":[72],"iterative":[74],"refinement,":[76],"existing":[77,158],"MDMs":[78,99],"do":[79],"not":[80],"support":[81],"multi-turn":[82],"denoising.":[85],"We":[86],"propose":[87],"Reflective":[88,172],"Masking":[89,173],"(RM),":[90],"which":[91],"elicits":[92],"such":[93],"an":[94,110],"intrinsic":[95],"capability":[97],"via":[100],"lightweight":[101],"post-training.":[102],"RM":[103,184],"provides":[104],"a":[105,137,186],"native":[106],"test-time":[107],"scaling,":[108],"where":[109],"MDM":[111],"iteratively":[112],"revisits":[113],"revises":[115],"its":[116],"prior":[117],"based":[119],"evolving":[121],"context.":[122],"To":[123],"exploit":[124],"insights":[125],"turns":[128],"like":[129],"AR":[130],"reasoning,":[131],"we":[132],"further":[133],"introduce":[134],"History":[135],"Reference,":[136],"parameter-free":[138],"that":[140],"leverages":[141],"intermediate":[142],"denoising":[143],"states":[144],"during":[145],"revision.":[146],"Our":[147],"approach":[148],"requires":[149],"no":[150],"architectural":[151],"changes":[152],"easily":[155],"applicable":[156],"to":[157],"MDMs.":[159,192],"Across":[160],"diverse":[161],"tasks":[162],"modalities,":[164],"including":[165],"text":[166],"Sudoku,":[168],"image":[170],"editing,":[171],"consistently":[174],"outperforms":[175],"standard":[176],"masking-based":[177],"baselines":[178],"demonstrates":[180],"strong":[181],"generality,":[182],"positioning":[183],"as":[185],"fundamental":[187],"primitive":[188],"for":[189]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-17T00:00:00"}