{"id":"https://openalex.org/W7117159477","doi":"https://doi.org/10.48550/arxiv.2512.19311","title":"MixFlow Training: Alleviating Exposure Bias with Slowed Interpolation Mixture","display_name":"MixFlow Training: Alleviating Exposure Bias with Slowed Interpolation Mixture","publication_year":2025,"publication_date":"2025-12-22","ids":{"openalex":"https://openalex.org/W7117159477","doi":"https://doi.org/10.48550/arxiv.2512.19311"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2512.19311","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.19311","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2512.19311","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5121159346","display_name":"Hui Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Li, Hui","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121215931","display_name":"Jiayue Lyu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lyu, Jiayue","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121184996","display_name":"Fu-Yun Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Fu-Yun","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121179366","display_name":"Kaihui Cheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cheng, Kaihui","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121191687","display_name":"Siyu Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Siyu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5121173504","display_name":"Jingdong Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Jingdong","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5121159346"],"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":0.8163999915122986,"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":0.8163999915122986,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0738999992609024,"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/T11206","display_name":"Model Reduction and Neural Networks","score":0.01269999984651804,"subfield":{"id":"https://openalex.org/subfields/3109","display_name":"Statistical and Nonlinear Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/interpolation","display_name":"Interpolation (computer graphics)","score":0.8202000260353088},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.6158000230789185},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.6021999716758728},{"id":"https://openalex.org/keywords/linear-interpolation","display_name":"Linear interpolation","score":0.4196000099182129},{"id":"https://openalex.org/keywords/nearest-neighbor-interpolation","display_name":"Nearest-neighbor interpolation","score":0.3878999948501587},{"id":"https://openalex.org/keywords/noise-measurement","display_name":"Noise measurement","score":0.3856000006198883},{"id":"https://openalex.org/keywords/multivariate-interpolation","display_name":"Multivariate interpolation","score":0.3630000054836273}],"concepts":[{"id":"https://openalex.org/C137800194","wikidata":"https://www.wikidata.org/wiki/Q11713455","display_name":"Interpolation (computer graphics)","level":3,"score":0.8202000260353088},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.6158000230789185},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.6021999716758728},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5472000241279602},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.5253999829292297},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.47110000252723694},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44699999690055847},{"id":"https://openalex.org/C171836373","wikidata":"https://www.wikidata.org/wiki/Q2266329","display_name":"Linear interpolation","level":3,"score":0.4196000099182129},{"id":"https://openalex.org/C207214200","wikidata":"https://www.wikidata.org/wiki/Q4202129","display_name":"Nearest-neighbor interpolation","level":4,"score":0.3878999948501587},{"id":"https://openalex.org/C29265498","wikidata":"https://www.wikidata.org/wiki/Q7047719","display_name":"Noise measurement","level":3,"score":0.3856000006198883},{"id":"https://openalex.org/C203332170","wikidata":"https://www.wikidata.org/wiki/Q6334079","display_name":"Multivariate interpolation","level":3,"score":0.3630000054836273},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3393999934196472},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.33469998836517334},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.3346000015735626},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.2994000017642975},{"id":"https://openalex.org/C119768884","wikidata":"https://www.wikidata.org/wiki/Q7597108","display_name":"Stairstep interpolation","level":4,"score":0.2879999876022339},{"id":"https://openalex.org/C188325010","wikidata":"https://www.wikidata.org/wiki/Q861176","display_name":"Trilinear interpolation","level":4,"score":0.27709999680519104},{"id":"https://openalex.org/C2781170535","wikidata":"https://www.wikidata.org/wiki/Q30587856","display_name":"Noisy data","level":2,"score":0.26159998774528503},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.25679999589920044},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.2529999911785126}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2512.19311","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.19311","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2512.19311","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2512.19311","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"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":{"This":[0],"paper":[1],"studies":[2],"the":[3,12,17,28,38,41,46,49,63,70,74,79,82,103,110,115,118,127,148,157],"training-testing":[4],"discrepancy":[5],"(a.k.a.":[6],"exposure":[7],"bias)":[8],"problem":[9],"for":[10,61,125,130],"improving":[11,62],"diffusion":[13],"models.":[14],"During":[15],"training,":[16],"input":[18,47],"of":[19,37,150],"a":[20,55,87,96],"prediction":[21,128],"network":[22,129],"at":[23,86,117,174,187],"one":[24],"training":[25,57,132],"timestep":[26,90,98,106],"is":[27,34,48,67,78,91,107],"corresponding":[29,104],"ground-truth":[30,75,105],"noisy":[31,51,84],"data":[32,85],"that":[33,77],"an":[35],"interpolation":[36,76,123],"noise":[39],"and":[40,43,142,144,170,178,183],"data,":[42],"during":[44],"testing,":[45],"generated":[50,83],"data.":[52],"We":[53],"present":[54],"novel":[56],"approach,":[58],"named":[59,121],"MixFlow,":[60],"performance.":[64],"Our":[65,153],"approach":[66,154],"motivated":[68],"by":[69],"Slow":[71],"Flow":[72],"phenomenon:":[73],"nearest":[80],"to":[81,93,95],"given":[88],"sampling":[89,111],"observed":[92],"correspond":[94],"higher-noise":[97],"(termed":[99],"slowed":[100,119,122],"timestep),":[101],"i.e.,":[102],"slower":[108],"than":[109],"timestep.":[112,133],"MixFlow":[113,155],"leverages":[114],"interpolations":[116],"timesteps,":[120],"mixture,":[124],"post-training":[126],"each":[131],"Experiments":[134],"over":[135,156],"class-conditional":[136],"image":[137],"generation":[138,146,162],"(including":[139],"SiT,":[140],"REPA,":[141],"RAE)":[143],"text-to-image":[145],"validate":[147],"effectiveness":[149],"our":[151],"approach.":[152],"RAE":[158],"models":[159],"achieve":[160],"strong":[161],"results":[163],"on":[164],"ImageNet:":[165],"1.43":[166],"FID":[167,180],"(without":[168,181],"guidance)":[169,173,182,186],"1.10":[171,184],"(with":[172,185],"256":[175],"x":[176,189],"256,":[177],"1.55":[179],"512":[188],"512.":[190]},"counts_by_year":[],"updated_date":"2025-12-24T23:14:05.333182","created_date":"2025-12-24T00:00:00"}