{"id":"https://openalex.org/W7155992286","doi":"https://doi.org/10.48550/arxiv.2604.22379","title":"Efficient Diffusion Distillation via Embedding Loss","display_name":"Efficient Diffusion Distillation via Embedding Loss","publication_year":2026,"publication_date":"2026-04-24","ids":{"openalex":"https://openalex.org/W7155992286","doi":"https://doi.org/10.48550/arxiv.2604.22379"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.22379","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.22379","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.2604.22379","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5023233603","display_name":"\u6afb \u91d1\u57ce","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ying, Jincheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101768547","display_name":"Yitao Chen","orcid":"https://orcid.org/0009-0007-7454-1802"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Yitao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134668064","display_name":"Li Wenlin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wenlin, Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103077343","display_name":"Minghui Xu","orcid":"https://orcid.org/0000-0003-3675-3461"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Minghui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5070580609","display_name":"Yinhao Xiao","orcid":"https://orcid.org/0000-0002-9871-9552"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiao, Yinhao","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":0.718500018119812,"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":0.718500018119812,"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":0.06270000338554382,"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":0.04619999974966049,"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/distillation","display_name":"Distillation","score":0.7343999743461609},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.6229000091552734},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.5303000211715698},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.506600022315979},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.40720000863075256},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.3984000086784363},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.3750999867916107}],"concepts":[{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.7343999743461609},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6621000170707703},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.6229000091552734},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.5303000211715698},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.506600022315979},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.40720000863075256},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.3984000086784363},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.3750999867916107},{"id":"https://openalex.org/C2780992000","wikidata":"https://www.wikidata.org/wiki/Q17016113","display_name":"Generator (circuit theory)","level":3,"score":0.3472999930381775},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3395000100135803},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.3239000141620636},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3206999897956848},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.29739999771118164},{"id":"https://openalex.org/C117251300","wikidata":"https://www.wikidata.org/wiki/Q1849855","display_name":"Parametric statistics","level":2,"score":0.29440000653266907},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.28679999709129333},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.2766000032424927},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.2727999985218048},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.27090001106262207},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.2685000002384186},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.2678000032901764},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.26750001311302185}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.22379","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.22379","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.2604.22379","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.22379","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":{"Recent":[0],"advances":[1],"in":[2,129,235],"distilling":[3],"expensive":[4],"diffusion":[5,83],"models":[6,234],"into":[7],"efficient":[8],"few-step":[9,117],"generators":[10],"show":[11],"significant":[12],"promise.":[13],"However,":[14],"these":[15],"methods":[16,85],"typically":[17],"demand":[18],"substantial":[19],"computational":[20],"resources":[21],"and":[22,31,47,63,90,119,143,175,189,195,201,227],"extended":[23],"training":[24,46,61,92,217],"periods,":[25],"limiting":[26],"accessibility":[27],"for":[28,157,172,177,230],"resource-constrained":[29,236],"researchers,":[30],"existing":[32,82,209],"supplementary":[33,77],"loss":[34,40,78],"functions":[35],"have":[36],"notable":[37],"limitations.":[38],"Regression":[39],"requires":[41],"pre-generating":[42],"large":[43],"datasets":[44],"before":[45],"limits":[48],"the":[49,53,111,115,120,130,161],"student":[50],"model":[51],"to":[52,86,221],"teacher's":[54],"performance,":[55],"while":[56],"GAN-based":[57],"losses":[58],"suffer":[59],"from":[60,100],"instability":[62],"require":[64],"careful":[65],"tuning.":[66],"In":[67],"this":[68],"paper,":[69],"we":[70,182],"propose":[71],"Embedding":[72],"Loss":[73],"(EL),":[74],"a":[75,101,224],"novel":[76],"function":[79],"that":[80],"complements":[81],"distillation":[84,150,190,203,211],"enhance":[87],"generation":[88,174],"quality":[89],"accelerate":[91],"with":[93],"smaller":[94],"batch":[95],"sizes.":[96],"Leveraging":[97],"feature":[98,112,132],"embeddings":[99],"diverse":[102],"set":[103],"of":[104,170],"randomly":[105],"initialized":[106],"networks,":[107],"EL":[108,134,152,185],"effectively":[109],"aligns":[110],"distributions":[113],"between":[114],"distilled":[116],"generator":[118],"original":[121],"data.":[122],"By":[123],"computing":[124],"Maximum":[125],"Mean":[126],"Discrepancy":[127],"(MMD)":[128],"embedded":[131],"space,":[133],"ensures":[135],"robust":[136],"distribution":[137,148],"matching,":[138],"thereby":[139],"preserving":[140],"sample":[141],"fidelity":[142],"diversity":[144],"during":[145],"distillation.":[146],"Within":[147],"matching":[149],"frameworks,":[151,204],"demonstrates":[153],"strong":[154],"empirical":[155],"performance":[156],"one-step":[158,210],"generators.":[159],"On":[160],"CIFAR-10":[162],"dataset,":[163],"our":[164],"approach":[165],"achieves":[166],"state-of-the-art":[167],"FID":[168],"values":[169],"1.475":[171],"unconditional":[173],"1.380":[176],"conditional":[178],"generation.":[179],"Beyond":[180],"CIFAR-10,":[181],"further":[183],"validate":[184],"across":[186],"multiple":[187],"benchmarks":[188],"methods,":[191],"including":[192],"ImageNet,":[193],"AFHQ-v2,":[194],"FFHQ":[196],"datasets,":[197],"using":[198],"DMD,":[199],"DI,":[200],"CM":[202],"demonstrating":[205],"consistent":[206],"improvements":[207],"over":[208],"methods.":[212],"Our":[213],"method":[214],"also":[215],"reduces":[216],"iterations":[218],"by":[219],"up":[220],"80%,":[222],"offering":[223],"more":[225],"practical":[226],"scalable":[228],"solution":[229],"deploying":[231],"diffusion-based":[232],"generative":[233],"environments.":[237]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-28T00:00:00"}