{"id":"https://openalex.org/W7135203466","doi":"https://doi.org/10.48550/arxiv.2603.11607","title":"DyWeight: Dynamic Gradient Weighting for Few-Step Diffusion Sampling","display_name":"DyWeight: Dynamic Gradient Weighting for Few-Step Diffusion Sampling","publication_year":2026,"publication_date":"2026-03-12","ids":{"openalex":"https://openalex.org/W7135203466","doi":"https://doi.org/10.48550/arxiv.2603.11607"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.11607","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11607","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2603.11607","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129042733","display_name":"Tong Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Zhao, Tong","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126784177","display_name":"Mingkun Lei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lei, Mingkun","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121607381","display_name":"Liangyu Yuan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan, Liangyu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128964406","display_name":"Yanming Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Yanming","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128971741","display_name":"Chenxi Song","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song, Chenxi","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129059870","display_name":"Yang Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009738035","display_name":"Beier Zhu","orcid":"https://orcid.org/0000-0002-7900-6979"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Beier","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128978507","display_name":"Chi Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Chi","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5129042733"],"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":0.7447999715805054,"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":0.7447999715805054,"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":0.046300001442432404,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.03689999878406525,"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/weighting","display_name":"Weighting","score":0.5972999930381775},{"id":"https://openalex.org/keywords/solver","display_name":"Solver","score":0.5582000017166138},{"id":"https://openalex.org/keywords/diffusion","display_name":"Diffusion","score":0.5406000018119812},{"id":"https://openalex.org/keywords/trajectory","display_name":"Trajectory","score":0.5394999980926514},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.5371000170707703},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.5256999731063843},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.4902999997138977},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.47200000286102295},{"id":"https://openalex.org/keywords/ode","display_name":"Ode","score":0.46389999985694885},{"id":"https://openalex.org/keywords/diffusion-process","display_name":"Diffusion process","score":0.4406000077724457}],"concepts":[{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.5972999930381775},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5680000185966492},{"id":"https://openalex.org/C2778770139","wikidata":"https://www.wikidata.org/wiki/Q1966904","display_name":"Solver","level":2,"score":0.5582000017166138},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.5406000018119812},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.5394999980926514},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.5371000170707703},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.5256999731063843},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5224999785423279},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.4902999997138977},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.47200000286102295},{"id":"https://openalex.org/C34862557","wikidata":"https://www.wikidata.org/wiki/Q178985","display_name":"Ode","level":2,"score":0.46389999985694885},{"id":"https://openalex.org/C68710425","wikidata":"https://www.wikidata.org/wiki/Q5275442","display_name":"Diffusion process","level":3,"score":0.4406000077724457},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.43650001287460327},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.4323999881744385},{"id":"https://openalex.org/C131584629","wikidata":"https://www.wikidata.org/wiki/Q4308705","display_name":"Coupling (piping)","level":2,"score":0.41780000925064087},{"id":"https://openalex.org/C4679612","wikidata":"https://www.wikidata.org/wiki/Q866298","display_name":"Aggregate (composite)","level":2,"score":0.3808000087738037},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.3700000047683716},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.3650999963283539},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.35249999165534973},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.34630000591278076},{"id":"https://openalex.org/C45347329","wikidata":"https://www.wikidata.org/wiki/Q5166604","display_name":"Convolution (computer science)","level":3,"score":0.34549999237060547},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3411000072956085},{"id":"https://openalex.org/C176321772","wikidata":"https://www.wikidata.org/wiki/Q1430640","display_name":"Numerical stability","level":3,"score":0.33559998869895935},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.32820001244544983},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.31049999594688416},{"id":"https://openalex.org/C57691317","wikidata":"https://www.wikidata.org/wiki/Q1289248","display_name":"Scalar (mathematics)","level":2,"score":0.2980000078678131},{"id":"https://openalex.org/C138673069","wikidata":"https://www.wikidata.org/wiki/Q322229","display_name":"Tracing","level":2,"score":0.2915000021457672},{"id":"https://openalex.org/C113364801","wikidata":"https://www.wikidata.org/wiki/Q26674","display_name":"High fidelity","level":2,"score":0.29109999537467957},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.289000004529953},{"id":"https://openalex.org/C208081375","wikidata":"https://www.wikidata.org/wiki/Q274502","display_name":"Degrees of freedom (physics and chemistry)","level":2,"score":0.2800999879837036},{"id":"https://openalex.org/C189950617","wikidata":"https://www.wikidata.org/wiki/Q937228","display_name":"Property (philosophy)","level":2,"score":0.2703000009059906},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.2669000029563904},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2637999951839447},{"id":"https://openalex.org/C75553542","wikidata":"https://www.wikidata.org/wiki/Q178161","display_name":"A priori and a posteriori","level":2,"score":0.259799987077713},{"id":"https://openalex.org/C106131492","wikidata":"https://www.wikidata.org/wiki/Q3072260","display_name":"Filter (signal processing)","level":2,"score":0.25929999351501465},{"id":"https://openalex.org/C174576160","wikidata":"https://www.wikidata.org/wiki/Q1183700","display_name":"Deconvolution","level":2,"score":0.25870001316070557},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.2574999928474426},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.2524000108242035}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.11607","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11607","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.11607","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11607","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"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":[0,138],"Models":[1],"(DMs)":[2],"have":[3],"achieved":[4],"state-of-the-art":[5,158],"generative":[6],"performance":[7],"across":[8],"multiple":[9],"modalities,":[10],"yet":[11],"their":[12],"sampling":[13],"process":[14],"remains":[15],"prohibitively":[16],"slow":[17],"due":[18],"to":[19,50,52],"the":[20,53,99,109,114],"need":[21],"for":[22],"hundreds":[23],"of":[24,56],"function":[25,153],"evaluations.":[26],"Recent":[27],"progress":[28],"in":[29],"multi-step":[30,72],"ODE":[31],"solvers":[32],"has":[33],"greatly":[34],"improved":[35],"efficiency":[36],"by":[37],"reusing":[38],"historical":[39,94],"gradients,":[40],"but":[41],"existing":[42],"methods":[43],"rely":[44],"on":[45,131],"handcrafted":[46],"coefficients":[47],"that":[48,74,91,142],"fail":[49],"adapt":[51],"non-stationary":[54],"dynamics":[55,118],"diffusion":[57,161],"sampling.":[58],"To":[59],"address":[60],"this":[61],"limitation,":[62],"we":[63],"propose":[64],"Dynamic":[65],"Gradient":[66],"Weighting":[67],"(DyWeight),":[68],"a":[69,76,156],"lightweight,":[70],"learning-based":[71],"solver":[73],"introduces":[75],"streamlined":[77],"implicit":[78,104],"coupling":[79],"paradigm.":[80],"By":[81],"relaxing":[82],"classical":[83],"numerical":[84,111],"constraints,":[85],"DyWeight":[86,143],"learns":[87],"unconstrained":[88],"time-varying":[89],"parameters":[90],"adaptively":[92],"aggregate":[93],"gradients":[95],"while":[96],"intrinsically":[97],"scaling":[98],"effective":[100],"step":[101],"size.":[102],"This":[103],"time":[105],"calibration":[106],"accurately":[107],"aligns":[108],"solver's":[110],"trajectory":[112],"with":[113,150],"model's":[115],"internal":[116],"denoising":[117],"under":[119],"large":[120],"integration":[121],"steps,":[122],"avoiding":[123],"complex":[124],"decoupled":[125],"parameterizations":[126],"and":[127,139,148],"optimizations.":[128],"Extensive":[129],"experiments":[130],"CIFAR-10,":[132],"FFHQ,":[133],"AFHQv2,":[134],"ImageNet64,":[135],"LSUN-Bedroom,":[136],"Stable":[137],"FLUX.1-dev":[140],"demonstrate":[141],"achieves":[144],"superior":[145],"visual":[146],"fidelity":[147],"stability":[149],"significantly":[151],"fewer":[152],"evaluations,":[154],"establishing":[155],"new":[157],"among":[159],"efficient":[160],"solvers.":[162],"Code":[163],"is":[164],"available":[165],"at":[166],"https://github.com/Westlake-AGI-Lab/DyWeight":[167]},"counts_by_year":[],"updated_date":"2026-03-14T06:46:50.379900","created_date":"2026-03-14T00:00:00"}