{"id":"https://openalex.org/W7225162803","doi":"https://doi.org/10.48550/arxiv.2611.02921","title":"Unlocking the Potential of Contrastive Regularization for Diffusion Transformer Training","display_name":"Unlocking the Potential of Contrastive Regularization for Diffusion Transformer Training","publication_year":2026,"publication_date":"2026-11-02","ids":{"openalex":"https://openalex.org/W7225162803","doi":"https://doi.org/10.48550/arxiv.2611.02921"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2611.02921","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2611.02921","pdf_url":"https://arxiv.org/pdf/2611.02921","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2611.02921","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5160540867","display_name":"Yi Tan","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tan, Yi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009738035","display_name":"Beier Zhu","orcid":"https://orcid.org/0000-0002-7900-6979","observed_orcids":["https://orcid.org/0000-0002-7900-6979"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Beier","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160397625","display_name":"Sirui Chen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I116953780","display_name":"Tongji University","ror":"https://ror.org/03rc6as71","country_code":"CN","type":"education","lineage":["https://openalex.org/I116953780"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chen, Sirui","raw_affiliation_strings":["Tongji University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tongji University","institution_ids":["https://openalex.org/I116953780"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5053667307","display_name":"\u6797\u5141\u54f2","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I16365422","display_name":"Hefei University of Technology","ror":"https://ror.org/02czkny70","country_code":"CN","type":"education","lineage":["https://openalex.org/I16365422"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lin, Yunzhe","raw_affiliation_strings":["HFUT"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"HFUT","institution_ids":["https://openalex.org/I16365422"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160497922","display_name":"Yuan Zhou","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Yuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160467954","display_name":"Cheng Chen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Cheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160543830","display_name":"Xingyu Zhu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I165932596","display_name":"National University of Singapore","ror":"https://ror.org/01tgyzw49","country_code":"SG","type":"education","lineage":["https://openalex.org/I165932596"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Zhu, Xingyu","raw_affiliation_strings":["NUS"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NUS","institution_ids":["https://openalex.org/I165932596"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160519669","display_name":"Tianrun Chen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Tianrun","raw_affiliation_strings":["Moxin Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Moxin Technology","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160497776","display_name":"Yanbin Hao","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I16365422","display_name":"Hefei University of Technology","ror":"https://ror.org/02czkny70","country_code":"CN","type":"education","lineage":["https://openalex.org/I16365422"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hao, Yanbin","raw_affiliation_strings":["HFUT"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"HFUT","institution_ids":["https://openalex.org/I16365422"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5160529624","display_name":"Hanwang Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Hanwang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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":0.6470000147819519,"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":0.6470000147819519,"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":0.164900004863739,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.07150000333786011,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/contrastive-regularization","display_name":"contrastive regularization","score":0.8299999833106995},{"id":"https://openalex.org/keywords/diffusion-transformer","display_name":"diffusion transformer","score":0.8169999718666077},{"id":"https://openalex.org/keywords/infonce","display_name":"InfoNCE","score":0.8040000200271606},{"id":"https://openalex.org/keywords/representation-learning","display_name":"representation learning","score":0.7239999771118164},{"id":"https://openalex.org/keywords/image-classification","display_name":"image classification","score":0.4429999887943268}],"concepts":[{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.6843000054359436},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5738999843597412},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5361999869346619},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4763999879360199},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.34630000591278076},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3449000120162964},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.34220001101493835},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.31459999084472656},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.3066999912261963},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.3057999908924103},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.28439998626708984},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.2818000018596649},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.2662999927997589},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2502000033855438}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2611.02921","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2611.02921","pdf_url":"https://arxiv.org/pdf/2611.02921","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2611.02921","doi":"https://doi.org/10.48550/arxiv.2611.02921","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2611.02921","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2611.02921","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2611.02921","pdf_url":"https://arxiv.org/pdf/2611.02921","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7225162803.pdf","grobid_xml":"https://content.openalex.org/works/W7225162803.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4403345601","https://openalex.org/W4415202797","https://openalex.org/W4416901645","https://openalex.org/W4417511319","https://openalex.org/W7214790027","https://openalex.org/W7125800819","https://openalex.org/W7168176385","https://openalex.org/W7164015409","https://openalex.org/W7221783811","https://openalex.org/W7172426404"],"abstract_inverted_index":{"Representation":[0],"regularization":[1,50,149],"has":[2],"emerged":[3],"as":[4,37,162],"an":[5,38,163],"effective":[6,39,164],"way":[7],"to":[8,117,121],"improve":[9],"diffusion":[10,42,71,119,135,167],"transformer":[11,136],"training.":[12],"Yet":[13],"one":[14],"of":[15,93],"the":[16,59,91,123],"most":[17],"canonical":[18],"regularizers--InfoNCE-style":[19],"contrastive":[20,34,94,106,160],"learning--remains":[21],"largely":[22],"underexplored":[23],"in":[24,70],"this":[25],"setting.":[26],"We":[27],"naturally":[28],"ask":[29],"a":[30,65,74],"question:":[31],"can":[32],"standard":[33],"learning":[35,161],"serve":[36],"regularizer":[40,165],"for":[41,76,166],"transformers?":[43],"Surprisingly,":[44],"we":[45,96],"find":[46],"that":[47,58,131],"naive":[48],"InfoNCE":[49],"brings":[51],"only":[52],"marginal":[53],"gains.":[54],"Our":[55],"analysis":[56],"reveals":[57],"objective":[60],"is":[61,173],"easily":[62],"dominated":[63],"by":[64],"shortcut:":[66],"near-orthogonal":[67],"noise":[68,87,124],"components":[69],"inputs":[72],"provide":[73],"bypass":[75],"enlarging":[77],"negative-pair":[78],"distances":[79],"while":[80,151],"weakening":[81],"positive-pair":[82],"alignment,":[83],"especially":[84],"at":[85],"high":[86],"levels.":[88],"To":[89],"unlock":[90],"benefit":[92],"regularization,":[95],"propose":[97],"Clean":[98],"Representation-ANchored":[99],"Contrastive":[100],"(CRAN)":[101],"regularization.":[102],"Specifically,":[103],"CRAN":[104,132],"anchors":[105],"dispersion":[107],"with":[108],"clean-sample":[109],"representation":[110],"distances,":[111],"transferring":[112],"clean":[113],"relational":[114],"semantic":[115],"structure":[116],"noisy":[118],"representations":[120],"suppress":[122],"shortcut.":[125],"Extensive":[126],"experiments":[127],"on":[128],"ImageNet":[129],"show":[130],"consistently":[133],"improves":[134],"training":[137,142],"across":[138],"model":[139],"scales":[140],"and":[141,147],"budgets,":[143],"outperforming":[144],"internal":[145],"feature-alignment":[146],"dispersive":[148],"baselines":[150],"matching":[152],"external":[153],"DINO-guided":[154],"REPA.":[155],"These":[156],"results":[157],"re-establish":[158],"InfoNCE-based":[159],"training,":[168],"once":[169],"its":[170],"noise-driven":[171],"over-dispersion":[172],"properly":[174],"constrained.":[175]},"counts_by_year":[],"updated_date":"2026-11-06T12:05:46.655746","created_date":"2026-11-04T00:00:00"}