{"id":"https://openalex.org/W7135077501","doi":"https://doi.org/10.48550/arxiv.2603.10391","title":"Variance-Aware Adaptive Weighting for Diffusion Model Training","display_name":"Variance-Aware Adaptive Weighting for Diffusion Model Training","publication_year":2026,"publication_date":"2026-03-11","ids":{"openalex":"https://openalex.org/W7135077501","doi":"https://doi.org/10.48550/arxiv.2603.10391"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.10391","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.10391","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.2603.10391","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128877460","display_name":"Nanlong Sun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Nanlong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128808615","display_name":"Lei Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Lei","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":0.7785999774932861,"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":0.7785999774932861,"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/T11206","display_name":"Model Reduction and Neural Networks","score":0.03669999912381172,"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"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.028599999845027924,"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/variance","display_name":"Variance (accounting)","score":0.7447999715805054},{"id":"https://openalex.org/keywords/weighting","display_name":"Weighting","score":0.7444000244140625},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.6292999982833862},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.6011000275611877},{"id":"https://openalex.org/keywords/perspective","display_name":"Perspective (graphical)","score":0.4966999888420105},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.4496000111103058},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.4334000051021576},{"id":"https://openalex.org/keywords/generative-model","display_name":"Generative model","score":0.4178999960422516}],"concepts":[{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.7447999715805054},{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.7444000244140625},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.6292999982833862},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.6011000275611877},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5834000110626221},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.4966999888420105},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4496000111103058},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.4334000051021576},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.4178999960422516},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3977999985218048},{"id":"https://openalex.org/C70136482","wikidata":"https://www.wikidata.org/wiki/Q13583781","display_name":"A-weighting","level":3,"score":0.3912999927997589},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.384799987077713},{"id":"https://openalex.org/C2986577269","wikidata":"https://www.wikidata.org/wiki/Q11306265","display_name":"Random noise","level":2,"score":0.34689998626708984},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.32910001277923584},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.31869998574256897},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.31839999556541443},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.31450000405311584},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.301800012588501},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.28850001096725464},{"id":"https://openalex.org/C125014702","wikidata":"https://www.wikidata.org/wiki/Q4680749","display_name":"Adaptive learning","level":2,"score":0.2702000141143799},{"id":"https://openalex.org/C167393769","wikidata":"https://www.wikidata.org/wiki/Q4680768","display_name":"Adaptive strategies","level":2,"score":0.26179999113082886}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.10391","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.10391","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.2603.10391","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.10391","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":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Diffusion":[0],"models":[1],"have":[2],"recently":[3],"achieved":[4],"remarkable":[5],"success":[6],"in":[7],"generative":[8,92],"modeling,":[9],"yet":[10],"their":[11],"training":[12,62,96,130,139],"dynamics":[13],"across":[14,44,76,109],"different":[15],"noise":[16,77],"levels":[17,46],"remain":[18],"highly":[19],"imbalanced,":[20],"which":[21],"can":[22],"lead":[23],"to":[24,54],"inefficient":[25],"optimization":[26,74],"and":[27,47,83,119],"unstable":[28],"learning":[29],"behavior.":[30],"In":[31],"this":[32,36],"work,":[33],"we":[34],"investigate":[35],"imbalance":[37],"from":[38],"the":[39,66,87,125,135],"perspective":[40],"of":[41,137],"loss":[42],"variance":[43,68,108,117],"log-SNR":[45],"propose":[48],"a":[49,71],"variance-aware":[50,138],"adaptive":[51,126],"weighting":[52,127],"strategy":[53],"address":[55],"it.":[56],"The":[57],"proposed":[58,88],"approach":[59],"dynamically":[60],"adjusts":[61],"weights":[63],"based":[64],"on":[65,81],"observed":[67],"distribution,":[69],"encouraging":[70],"more":[72],"balanced":[73],"process":[75],"levels.":[78],"Extensive":[79],"experiments":[80],"CIFAR-10":[82],"CIFAR-100":[84],"demonstrate":[85],"that":[86,124],"method":[89],"consistently":[90],"improves":[91],"performance":[93,107],"over":[94],"standard":[95],"schemes,":[97],"achieving":[98],"lower":[99],"Fr\u00e9chet":[100],"Inception":[101],"Distance":[102],"(FID)":[103],"while":[104],"also":[105],"reducing":[106],"random":[110],"seeds.":[111],"Additional":[112],"analysis,":[113],"including":[114],"loss-log-SNR":[115],"visualization,":[116],"heatmaps,":[118],"ablation":[120],"studies,":[121],"further":[122],"reveal":[123],"effectively":[128],"stabilizes":[129],"dynamics.":[131],"These":[132],"results":[133],"highlight":[134],"potential":[136],"strategies":[140],"for":[141],"improving":[142],"diffusion":[143],"model":[144],"optimization.":[145]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-13T00:00:00"}