{"id":"https://openalex.org/W7161808386","doi":"https://doi.org/10.48550/arxiv.2605.19317","title":"Inference-Time Scaling in Diffusion Models through Iterative Partial Refinement","display_name":"Inference-Time Scaling in Diffusion Models through Iterative Partial Refinement","publication_year":2026,"publication_date":"2026-05-19","ids":{"openalex":"https://openalex.org/W7161808386","doi":"https://doi.org/10.48550/arxiv.2605.19317"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.19317","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.19317","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,"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.2605.19317","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136568340","display_name":"Taegu Kang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kang, Taegu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110659834","display_name":"Jaesik Yoon","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yoon, Jaesik","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136511047","display_name":"Sungjin Ahn","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ahn, Sungjin","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":0.5594000220298767,"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":0.5594000220298767,"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":0.10750000178813934,"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/T12303","display_name":"Tensor decomposition and applications","score":0.021900000050663948,"subfield":{"id":"https://openalex.org/subfields/2605","display_name":"Computational Mathematics"},"field":{"id":"https://openalex.org/fields/26","display_name":"Mathematics"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.7073000073432922},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.6150000095367432},{"id":"https://openalex.org/keywords/diffusion","display_name":"Diffusion","score":0.6007000207901001},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.49309998750686646},{"id":"https://openalex.org/keywords/constraint","display_name":"Constraint (computer-aided design)","score":0.4884999990463257},{"id":"https://openalex.org/keywords/iterative-method","display_name":"Iterative method","score":0.48190000653266907},{"id":"https://openalex.org/keywords/iterative-refinement","display_name":"Iterative refinement","score":0.41760000586509705},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.41280001401901245},{"id":"https://openalex.org/keywords/limiting","display_name":"Limiting","score":0.3555999994277954}],"concepts":[{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.7073000073432922},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.6150000095367432},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.6007000207901001},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5809999704360962},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.49309998750686646},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.4884999990463257},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.48660001158714294},{"id":"https://openalex.org/C159694833","wikidata":"https://www.wikidata.org/wiki/Q2321565","display_name":"Iterative method","level":2,"score":0.48190000653266907},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.4309999942779541},{"id":"https://openalex.org/C2779982483","wikidata":"https://www.wikidata.org/wiki/Q6094420","display_name":"Iterative refinement","level":2,"score":0.41760000586509705},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.41280001401901245},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.3555999994277954},{"id":"https://openalex.org/C190502265","wikidata":"https://www.wikidata.org/wiki/Q17069496","display_name":"MNIST database","level":3,"score":0.34470000863075256},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.34310001134872437},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.33799999952316284},{"id":"https://openalex.org/C2780069185","wikidata":"https://www.wikidata.org/wiki/Q7977945","display_name":"Equivalence (formal languages)","level":2,"score":0.33480000495910645},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.33009999990463257},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.3188000023365021},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.2980000078678131},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.28529998660087585},{"id":"https://openalex.org/C143587482","wikidata":"https://www.wikidata.org/wiki/Q1543216","display_name":"Iterative and incremental development","level":2,"score":0.2802000045776367},{"id":"https://openalex.org/C121864883","wikidata":"https://www.wikidata.org/wiki/Q677916","display_name":"Statistical physics","level":1,"score":0.27300000190734863},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.2671999931335449},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.26159998774528503},{"id":"https://openalex.org/C132459708","wikidata":"https://www.wikidata.org/wiki/Q744069","display_name":"Extrapolation","level":2,"score":0.2596000134944916},{"id":"https://openalex.org/C57869625","wikidata":"https://www.wikidata.org/wiki/Q1783502","display_name":"Rate of convergence","level":3,"score":0.2522999942302704},{"id":"https://openalex.org/C2776029896","wikidata":"https://www.wikidata.org/wiki/Q3935810","display_name":"Relaxation (psychology)","level":2,"score":0.2508000135421753}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.19317","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.19317","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,"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.2605.19317","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.19317","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,"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":{"Inference-time":[0],"scaling":[1,23,66,82,178],"has":[2,13],"emerged":[3],"as":[4,174],"a":[5,99,119],"major":[6],"approach":[7],"for":[8,25,84,180],"improving":[9],"reasoning":[10,142],"capabilities,":[11],"and":[12,38,51,103],"been":[14],"increasingly":[15],"applied":[16],"to":[17,36,44,68,114,162],"diffusion":[18,26,56,86,181],"models.":[19],"However,":[20],"existing":[21],"inference-time":[22,65,81,177],"methods":[24],"models":[27,35,57,182],"typically":[28],"rely":[29],"on":[30,107,152],"external":[31,90,139],"verifiers":[32],"or":[33],"reward":[34],"rank":[37],"select":[39],"samples,":[40],"limiting":[41],"their":[42],"scalability":[43],"settings":[45],"where":[46],"such":[47],"evaluators":[48],"are":[49],"available":[50,124,189],"reliable.":[52],"Moreover,":[53],"while":[54],"recent":[55],"perform":[58],"sequential":[59,85],"inference":[60],"with":[61],"region-wise,":[62],"mixed-noise":[63,185],"conditioning,":[64],"tailored":[67],"this":[69],"setting":[70],"remains":[71],"relatively":[72],"underexplored.":[73],"We":[74],"propose":[75],"Iterative":[76],"Partial":[77],"Refinement":[78],"(IPR),":[79],"an":[80,94,175],"method":[83],"that":[87,167],"requires":[88],"no":[89],"verifier.":[91],"Starting":[92],"from":[93,160],"already-generated":[95],"sample,":[96],"IPR":[97,148],"re-noises":[98],"subset":[100],"of":[101],"regions":[102],"regenerates":[104],"them":[105],"conditioned":[106],"the":[108,112,126,155],"remaining":[109],"regions,":[110],"enabling":[111],"model":[113],"revise":[115],"earlier":[116],"decisions":[117],"under":[118],"richer":[120],"context":[121],"than":[122],"was":[123],"during":[125],"initial":[127],"generation.":[128],"This":[129],"iterative":[130,168],"partial":[131,169],"refinement":[132,170],"produces":[133],"more":[134],"globally":[135],"consistent":[136],"samples":[137],"without":[138],"verification.":[140],"On":[141],"tasks":[143],"requiring":[144],"global":[145],"constraint":[146],"satisfaction,":[147],"consistently":[149],"improves":[150],"performance:":[151],"MNIST":[153],"Sudoku,":[154],"valid":[156],"solution":[157],"rate":[158],"increases":[159],"55.8%":[161],"75.0%.":[163],"These":[164],"results":[165],"show":[166],"alone":[171],"can":[172],"serve":[173],"effective":[176],"strategy":[179],"in":[183],"sequential,":[184],"settings.":[186],"Code":[187],"is":[188],"at:":[190],"https://github.com/ahn-ml/IPR":[191]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-21T00:00:00"}