{"id":"https://openalex.org/W7140125607","doi":"https://doi.org/10.48550/arxiv.2603.19939","title":"Timestep-Aware Block Masking for Efficient Diffusion Model Inference","display_name":"Timestep-Aware Block Masking for Efficient Diffusion Model Inference","publication_year":2026,"publication_date":"2026-03-20","ids":{"openalex":"https://openalex.org/W7140125607","doi":"https://doi.org/10.48550/arxiv.2603.19939"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.19939","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19939","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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2603.19939","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5130388119","display_name":"Haodong He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Haodong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130326858","display_name":"Yuan Gao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Yuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130334845","display_name":"Weizhong Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Weizhong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130388582","display_name":"Gui-Song Xia","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xia, Gui-Song","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"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":0.9715999960899353,"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":0.9715999960899353,"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/T11019","display_name":"Image Enhancement Techniques","score":0.003000000026077032,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.0027000000700354576,"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/inference","display_name":"Inference","score":0.6636999845504761},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.5672000050544739},{"id":"https://openalex.org/keywords/fidelity","display_name":"Fidelity","score":0.558899998664856},{"id":"https://openalex.org/keywords/noise-reduction","display_name":"Noise reduction","score":0.526199996471405},{"id":"https://openalex.org/keywords/probabilistic-logic","display_name":"Probabilistic logic","score":0.42179998755455017},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.3991999924182892},{"id":"https://openalex.org/keywords/iterative-and-incremental-development","display_name":"Iterative and incremental development","score":0.3540000021457672},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.35199999809265137},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.34209999442100525}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7335000038146973},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6636999845504761},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.5672000050544739},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.558899998664856},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.526199996471405},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4553000032901764},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44449999928474426},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.42179998755455017},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.3991999924182892},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3824999928474426},{"id":"https://openalex.org/C143587482","wikidata":"https://www.wikidata.org/wiki/Q1543216","display_name":"Iterative and incremental development","level":2,"score":0.3540000021457672},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.35199999809265137},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.34209999442100525},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.3418000042438507},{"id":"https://openalex.org/C2777402240","wikidata":"https://www.wikidata.org/wiki/Q6783436","display_name":"Masking (illustration)","level":2,"score":0.3294999897480011},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.32589998841285706},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3255999982357025},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.31630000472068787},{"id":"https://openalex.org/C70518039","wikidata":"https://www.wikidata.org/wiki/Q16000077","display_name":"Dimensionality reduction","level":2,"score":0.28999999165534973},{"id":"https://openalex.org/C45347329","wikidata":"https://www.wikidata.org/wiki/Q5166604","display_name":"Convolution (computer science)","level":3,"score":0.28790000081062317},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2858000099658966},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.2784000039100647},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.2782999873161316},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2750000059604645},{"id":"https://openalex.org/C75291252","wikidata":"https://www.wikidata.org/wiki/Q1315756","display_name":"TRACE (psycholinguistics)","level":2,"score":0.2727000117301941},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.2720000147819519},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.266400009393692},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.2630999982357025},{"id":"https://openalex.org/C159694833","wikidata":"https://www.wikidata.org/wiki/Q2321565","display_name":"Iterative method","level":2,"score":0.25619998574256897},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.2558000087738037},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.251800000667572}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.19939","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19939","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":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.19939","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19939","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":"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":{"Diffusion":[0],"Probabilistic":[1],"Models":[2],"(DPMs)":[3],"have":[4],"achieved":[5],"great":[6],"success":[7],"in":[8],"image":[9],"generation":[10],"but":[11],"suffer":[12],"from":[13],"high":[14],"inference":[15,69],"latency":[16],"due":[17],"to":[18,38,60,122],"their":[19],"iterative":[20],"denoising":[21,31,113,155],"nature.":[22],"Motivated":[23],"by":[24,116,152],"the":[25,30,40,154],"evolving":[26],"feature":[27,65,109],"dynamics":[28],"across":[29,136],"trajectory,":[32],"we":[33,100],"propose":[34],"a":[35,47,92,102,117,137,158,167],"novel":[36],"framework":[37],"optimize":[39],"computational":[41,162],"graph":[42],"of":[43,140,160],"pre-trained":[44],"DPMs":[45],"on":[46],"per-timestep":[48],"basis.":[49],"By":[50],"learning":[51],"timestep-specific":[52],"masks,":[53],"our":[54,83,164],"method":[55,84,165],"dynamically":[56],"determines":[57],"which":[58],"blocks":[59],"execute":[61],"or":[62],"bypass":[63],"through":[64],"reuse":[66],"at":[67],"each":[68,88],"stage.":[70],"Unlike":[71],"global":[72],"optimization":[73],"methods":[74],"that":[75,107,151],"incur":[76],"prohibitive":[77],"memory":[78],"costs":[79],"via":[80],"full-chain":[81],"backpropagation,":[82],"optimizes":[85],"masks":[86],"for":[87],"timestep":[89],"independently,":[90],"ensuring":[91],"memory-efficient":[93],"training":[94],"process.":[95],"To":[96],"guide":[97],"this":[98],"process,":[99],"introduce":[101],"timestep-aware":[103],"loss":[104],"scaling":[105],"mechanism":[106],"prioritizes":[108],"fidelity":[110],"during":[111],"sensitive":[112],"phases,":[114],"complemented":[115],"knowledge-guided":[118],"mask":[119],"rectification":[120],"strategy":[121],"prune":[123],"redundant":[124],"spatial-temporal":[125],"dependencies.":[126],"Our":[127,176],"approach":[128],"is":[129],"architecture-agnostic":[130],"and":[131,146,173],"demonstrates":[132],"significant":[133],"efficiency":[134],"gains":[135],"broad":[138],"spectrum":[139],"models,":[141],"including":[142],"DDPM,":[143],"LDM,":[144],"DiT,":[145],"PixArt.":[147],"Experimental":[148],"results":[149],"show":[150],"treating":[153],"process":[156],"as":[157],"sequence":[159],"optimized":[161],"paths,":[163],"achieves":[166],"superior":[168],"balance":[169],"between":[170],"sampling":[171],"speed":[172],"generative":[174],"quality.":[175],"code":[177],"will":[178],"be":[179],"released.":[180]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-03-24T00:00:00"}