{"id":"https://openalex.org/W7225260336","doi":"https://doi.org/10.48550/arxiv.2611.04386","title":"StepX-OPD: Rethinking Multi-Teacher On-Policy Distillation in Diffusion Models","display_name":"StepX-OPD: Rethinking Multi-Teacher On-Policy Distillation in Diffusion Models","publication_year":2026,"publication_date":"2026-11-03","ids":{"openalex":"https://openalex.org/W7225260336","doi":"https://doi.org/10.48550/arxiv.2611.04386"},"language":"en","primary_location":{"id":"doi:10.48550/arxiv.2611.04386","doi":"https://doi.org/10.48550/arxiv.2611.04386","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2611.04386","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,"listed_in":[],"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":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"volume":null,"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2611.04386","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125932880","display_name":"Chenru Wang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Chenru","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041793389","display_name":"Xuanyang Zhang","orcid":"https://orcid.org/0000-0001-5092-8657","observed_orcids":["https://orcid.org/0000-0001-5092-8657"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Xuanyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160597527","display_name":"Yucheng Han","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han, Yucheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5030688626","display_name":"Wenbo Nie","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Nie, Wenbo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160725771","display_name":"Wei Cheng","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cheng, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160650890","display_name":"Xianfang Zeng","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zeng, Xianfang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160685980","display_name":"Gang YU","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"YU, Gang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5160592020","display_name":"Chi Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Chi","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":0.9980000257492065,"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":0.9980000257492065,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.00039999998989515007,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.0003000000142492354,"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":[],"concepts":[{"id":"https://openalex.org/C205711294","wikidata":"https://www.wikidata.org/wiki/Q176953","display_name":"Rendering (computer graphics)","level":2,"score":0.5788999795913696},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5443000197410583},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.5304999947547913},{"id":"https://openalex.org/C190839683","wikidata":"https://www.wikidata.org/wiki/Q2448197","display_name":"Train","level":2,"score":0.4936000108718872},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.4918000102043152},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.4629000127315521},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.4165000021457672},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.41449999809265137},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3905999958515167},{"id":"https://openalex.org/C176734034","wikidata":"https://www.wikidata.org/wiki/Q259432","display_name":"Worksheet","level":2,"score":0.3560999929904938},{"id":"https://openalex.org/C124681953","wikidata":"https://www.wikidata.org/wiki/Q339062","display_name":"Decomposition","level":2,"score":0.3515999913215637},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3467000126838684},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3319999873638153},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.30630001425743103},{"id":"https://openalex.org/C44154836","wikidata":"https://www.wikidata.org/wiki/Q45045","display_name":"Simulation","level":1,"score":0.28209999203681946},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.2797999978065491},{"id":"https://openalex.org/C2779346075","wikidata":"https://www.wikidata.org/wiki/Q7268763","display_name":"Quality Score","level":3,"score":0.2770000100135803},{"id":"https://openalex.org/C3020318244","wikidata":"https://www.wikidata.org/wiki/Q4812187","display_name":"Large sample","level":2,"score":0.2721000015735626},{"id":"https://openalex.org/C3018868555","wikidata":"https://www.wikidata.org/wiki/Q2918907","display_name":"Single camera","level":2,"score":0.26100000739097595},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.2603999972343445},{"id":"https://openalex.org/C55020928","wikidata":"https://www.wikidata.org/wiki/Q3813865","display_name":"Image quality","level":3,"score":0.25679999589920044},{"id":"https://openalex.org/C129848803","wikidata":"https://www.wikidata.org/wiki/Q2564360","display_name":"Sample size determination","level":2,"score":0.25380000472068787},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.2515999972820282}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2611.04386","doi":"https://doi.org/10.48550/arxiv.2611.04386","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2611.04386","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,"listed_in":[],"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":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2611.04386","doi":"https://doi.org/10.48550/arxiv.2611.04386","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2611.04386","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,"listed_in":[],"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":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"A":[0],"deployable":[1],"text-to-image":[2],"diffusion":[3],"model":[4,38,57],"must":[5,21],"hold":[6],"several":[7],"capabilities":[8,41],"at":[9],"once,":[10],"such":[11],"as":[12,48],"instruction":[13,180],"following,":[14,181],"text":[15,182],"rendering":[16],"and":[17,20,28,64,165,177,186,204,210],"human":[18,187],"preference,":[19],"sample":[22,90],"in":[23,33,91],"a":[24,36,49,92,110,117,168],"few":[25,93],"steps.":[26],"Capabilities":[27],"few-step":[29],"sampling":[30],"are":[31],"built":[32],"separate":[34],"stages:":[35,81],"full-step":[37,56],"first":[39,72],"consolidates":[40],"through":[42],"multi-teacher":[43],"on-policy":[44],"distillation,":[45],"then":[46],"serves":[47],"fixed":[50],"teacher":[51],"for":[52,163],"step":[53,225],"distillation.":[54,226],"Every":[55],"we":[58,132,146,166],"evaluate":[59,174],"scores":[60],"lower":[61],"once":[62],"compressed,":[63],"the":[65,71,79,101,120,129,135,139,143,194,200,206],"second":[66],"stage":[67],"cannot":[68],"begin":[69],"until":[70],"finishes.":[73],"We":[74,173],"introduce":[75],"StepX-OPD,":[76],"which":[77],"merges":[78],"two":[80],"one":[82],"student":[83],"acquires":[84],"every":[85],"capability":[86,149],"while":[87],"learning":[88],"to":[89],"steps,":[94],"so":[95],"frozen":[96],"specialists":[97],"supervise":[98],"it":[99,113,213],"on":[100,109,138,158,175,211],"states":[102],"its":[103],"own":[104],"rollout":[105],"visits":[106],"rather":[107],"than":[108,220],"longer":[111],"trajectory":[112],"never":[114,161],"runs.":[115],"Because":[116],"decomposition":[118],"of":[119,142],"rollout's":[121],"KL":[122],"weights":[123],"velocity":[124,137],"mismatch":[125],"most":[126],"heavily":[127],"near":[128],"clean":[130],"end,":[131],"also":[133],"regress":[134],"student's":[136],"low-noise":[140],"part":[141],"rollout.":[144],"Separately,":[145],"find":[147],"that":[148,160,170],"integration":[150],"makes":[151],"models":[152],"write":[153],"prompt":[154],"words":[155],"into":[156],"images":[157],"prompts":[159],"asked":[162],"text,":[164],"give":[167],"variant":[169],"lowers":[171],"it.":[172],"SD3.5-Medium":[176,212],"FLUX.1-dev":[178],"across":[179],"rendering,":[183],"image":[184],"quality":[185],"preference.":[188],"On":[189],"both":[190],"architectures,":[191],"StepX-OPD":[192],"attains":[193],"highest":[195],"normalized":[196],"average":[197],"score":[198],"among":[199],"four-step":[201],"methods":[202],"compared":[203],"exceeds":[205],"forty-step":[207],"Flow-OPD":[208],"model,":[209],"trains":[214],"with":[215],"about":[216],"37%":[217],"fewer":[218],"GPU-hours":[219],"sequential":[221],"DiffusionOPD":[222],"followed":[223],"by":[224]},"counts_by_year":[],"updated_date":"2026-11-05T12:25:35.769452","created_date":"2026-11-05T00:00:00"}