{"id":"https://openalex.org/W4404571223","doi":"https://doi.org/10.1109/tpami.2026.3665753","title":"Aligning Few-Step Diffusion Models With Dense Reward Difference Learning","display_name":"Aligning Few-Step Diffusion Models With Dense Reward Difference Learning","publication_year":2026,"publication_date":"2026-02-23","ids":{"openalex":"https://openalex.org/W4404571223","doi":"https://doi.org/10.1109/tpami.2026.3665753","pmid":"https://pubmed.ncbi.nlm.nih.gov/41729671"},"language":"en","primary_location":{"id":"doi:10.1109/tpami.2026.3665753","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpami.2026.3665753","pdf_url":null,"source":{"id":"https://openalex.org/S199944782","display_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","issn_l":"0162-8828","issn":["0162-8828","1939-3539","2160-9292"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","raw_type":"journal-article"},"type":"article","indexed_in":["arxiv","crossref","datacite","pubmed"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2411.11727","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100352957","display_name":"Ziyi Zhang","orcid":"https://orcid.org/0000-0003-1927-8451"},"institutions":[{"id":"https://openalex.org/I117357259","display_name":"National Institute of Japanese Literature","ror":"https://ror.org/01464wm64","country_code":"JP","type":"archive","lineage":["https://openalex.org/I117357259","https://openalex.org/I1319490839","https://openalex.org/I4210148627"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Ziyi Zhang","raw_affiliation_strings":["School of Computer Science, National Engineering Research Center for Multimedia Software, Hubei Key Laboratory of Multimedia and Network Communication Engineering and Institute of Artificial Intelligence, Wuhan University, Wuhan, China"],"raw_orcid":"https://orcid.org/0000-0003-1728-2588","affiliations":[{"raw_affiliation_string":"School of Computer Science, National Engineering Research Center for Multimedia Software, Hubei Key Laboratory of Multimedia and Network Communication Engineering and Institute of Artificial Intelligence, Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I117357259"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100333326","display_name":"Shen Li","orcid":"https://orcid.org/0000-0002-9482-3045"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li Shen","raw_affiliation_strings":["School of Cyber Science and Technology, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China"],"raw_orcid":"https://orcid.org/0000-0001-5659-3464","affiliations":[{"raw_affiliation_string":"School of Cyber Science and Technology, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100378880","display_name":"Sen Zhang","orcid":"https://orcid.org/0000-0002-0077-2196"},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"The University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Sen Zhang","raw_affiliation_strings":["TikTok, ByteDance, Sydney, NSW, Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"TikTok, ByteDance, Sydney, NSW, Australia","institution_ids":["https://openalex.org/I129604602"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019789719","display_name":"Deheng Ye","orcid":null},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Deheng Ye","raw_affiliation_strings":["Tencent Inc., Shenzhen, China"],"raw_orcid":"https://orcid.org/0000-0002-1754-1837","affiliations":[{"raw_affiliation_string":"Tencent Inc., Shenzhen, China","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102207172","display_name":"Yong Luo","orcid":null},"institutions":[{"id":"https://openalex.org/I117357259","display_name":"National Institute of Japanese Literature","ror":"https://ror.org/01464wm64","country_code":"JP","type":"archive","lineage":["https://openalex.org/I117357259","https://openalex.org/I1319490839","https://openalex.org/I4210148627"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Yong Luo","raw_affiliation_strings":["School of Computer Science, National Engineering Research Center for Multimedia Software, Hubei Key Laboratory of Multimedia and Network Communication Engineering and Institute of Artificial Intelligence, Wuhan University, Wuhan, China"],"raw_orcid":"https://orcid.org/0000-0002-2296-6370","affiliations":[{"raw_affiliation_string":"School of Computer Science, National Engineering Research Center for Multimedia Software, Hubei Key Laboratory of Multimedia and Network Communication Engineering and Institute of Artificial Intelligence, Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I117357259"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101675323","display_name":"Miaojing Shi","orcid":"https://orcid.org/0000-0002-4933-0073"},"institutions":[{"id":"https://openalex.org/I116953780","display_name":"Tongji University","ror":"https://ror.org/03rc6as71","country_code":"CN","type":"education","lineage":["https://openalex.org/I116953780"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Miaojing Shi","raw_affiliation_strings":["College of Electronic and Information Engineering, Tongji University, Shanghai, China"],"raw_orcid":"https://orcid.org/0000-0002-4933-0073","affiliations":[{"raw_affiliation_string":"College of Electronic and Information Engineering, Tongji University, Shanghai, China","institution_ids":["https://openalex.org/I116953780"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Dongjing Shan","orcid":"https://orcid.org/0000-0001-6386-6096"},"institutions":[{"id":"https://openalex.org/I3017480383","display_name":"Southwest Medical University","ror":"https://ror.org/00g2rqs52","country_code":"CN","type":"funder","lineage":["https://openalex.org/I3017480383"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dongjing Shan","raw_affiliation_strings":["School of Medical Information and Engineering, Southwest Medical University, Luzhou, China"],"raw_orcid":"https://orcid.org/0000-0001-6386-6096","affiliations":[{"raw_affiliation_string":"School of Medical Information and Engineering, Southwest Medical University, Luzhou, China","institution_ids":["https://openalex.org/I3017480383"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100745020","display_name":"Boxue Du","orcid":"https://orcid.org/0000-0002-7500-1134"},"institutions":[{"id":"https://openalex.org/I117357259","display_name":"National Institute of Japanese Literature","ror":"https://ror.org/01464wm64","country_code":"JP","type":"archive","lineage":["https://openalex.org/I117357259","https://openalex.org/I1319490839","https://openalex.org/I4210148627"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Bo Du","raw_affiliation_strings":["School of Computer Science, National Engineering Research Center for Multimedia Software, Hubei Key Laboratory of Multimedia and Network Communication Engineering and Institute of Artificial Intelligence, Wuhan University, Wuhan, China"],"raw_orcid":"https://orcid.org/0000-0002-0059-8458","affiliations":[{"raw_affiliation_string":"School of Computer Science, National Engineering Research Center for Multimedia Software, Hubei Key Laboratory of Multimedia and Network Communication Engineering and Institute of Artificial Intelligence, Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I117357259"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5074103823","display_name":"Dacheng Tao","orcid":"https://orcid.org/0000-0001-7225-5449"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Dacheng Tao","raw_affiliation_strings":["College of Computing and Data Science and the Generative AI Lab, Nanyang Technological University, Singapore"],"raw_orcid":"https://orcid.org/0000-0001-7225-5449","affiliations":[{"raw_affiliation_string":"College of Computing and Data Science and the Generative AI Lab, Nanyang Technological University, Singapore","institution_ids":["https://openalex.org/I172675005"]}]}],"institutions":[],"countries_distinct_count":4,"institutions_distinct_count":7,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.00839962,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":"48","issue":"7","first_page":"7375","last_page":"7386"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11206","display_name":"Model Reduction and Neural Networks","score":0.7813000082969666,"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"}},"topics":[{"id":"https://openalex.org/T11206","display_name":"Model Reduction and Neural Networks","score":0.7813000082969666,"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/diffusion","display_name":"Diffusion","score":0.6025652885437012},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.4789046049118042},{"id":"https://openalex.org/keywords/cognitive-psychology","display_name":"Cognitive psychology","score":0.37500593066215515},{"id":"https://openalex.org/keywords/statistical-physics","display_name":"Statistical physics","score":0.37149709463119507},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.36472785472869873},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.33749687671661377},{"id":"https://openalex.org/keywords/econometrics","display_name":"Econometrics","score":0.3244786858558655},{"id":"https://openalex.org/keywords/economics","display_name":"Economics","score":0.2795264720916748},{"id":"https://openalex.org/keywords/physics","display_name":"Physics","score":0.14383801817893982},{"id":"https://openalex.org/keywords/thermodynamics","display_name":"Thermodynamics","score":0.09058594703674316}],"concepts":[{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.6025652885437012},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.4789046049118042},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.37500593066215515},{"id":"https://openalex.org/C121864883","wikidata":"https://www.wikidata.org/wiki/Q677916","display_name":"Statistical physics","level":1,"score":0.37149709463119507},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.36472785472869873},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.33749687671661377},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.3244786858558655},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.2795264720916748},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.14383801817893982},{"id":"https://openalex.org/C97355855","wikidata":"https://www.wikidata.org/wiki/Q11473","display_name":"Thermodynamics","level":1,"score":0.09058594703674316}],"mesh":[],"locations_count":6,"locations":[{"id":"doi:10.1109/tpami.2026.3665753","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpami.2026.3665753","pdf_url":null,"source":{"id":"https://openalex.org/S199944782","display_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","issn_l":"0162-8828","issn":["0162-8828","1939-3539","2160-9292"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","raw_type":"journal-article"},{"id":"pmid:41729671","is_oa":false,"landing_page_url":"https://pubmed.ncbi.nlm.nih.gov/41729671","pdf_url":null,"source":{"id":"https://openalex.org/S4306525036","display_name":"PubMed","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE transactions on pattern analysis and machine intelligence","raw_type":null},{"id":"pmh:oai:arXiv.org:2411.11727","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.11727","pdf_url":"https://arxiv.org/pdf/2411.11727","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"text"},{"id":"pmh:doi:10.48550/arxiv.2411.11727","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"pmh:oai:dr.ntu.edu.sg:10356/212207","is_oa":false,"landing_page_url":"https://hdl.handle.net/10356/212207","pdf_url":null,"source":{"id":"https://openalex.org/S4306402609","display_name":"DR-NTU (Nanyang Technological University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I172675005","host_organization_name":"Nanyang Technological University","host_organization_lineage":["https://openalex.org/I172675005"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Journal Article"},{"id":"doi:10.48550/arxiv.2411.11727","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2411.11727","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":"pmh:oai:arXiv.org:2411.11727","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.11727","pdf_url":"https://arxiv.org/pdf/2411.11727","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"text"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1163755948","display_name":null,"funder_award_id":"62576364","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G1284973545","display_name":null,"funder_award_id":"62276195","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G1463586485","display_name":null,"funder_award_id":"2024AFA017","funder_id":"https://openalex.org/F4320327654","funder_display_name":"Foundation for Innovative Research Groups of Hubei Province of China"},{"id":"https://openalex.org/G1836220556","display_name":null,"funder_award_id":"62225113","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8276535520","display_name":null,"funder_award_id":"U23A20318","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320327654","display_name":"Foundation for Innovative Research Groups of Hubei Province of China","ror":null}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4404571223.pdf"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W3188962172","https://openalex.org/W2772917594","https://openalex.org/W4312825515","https://openalex.org/W4306742369","https://openalex.org/W4303457083","https://openalex.org/W2386796262","https://openalex.org/W2131146434","https://openalex.org/W2951359407","https://openalex.org/W4376623224","https://openalex.org/W4387849428"],"abstract_inverted_index":{"Few-step":[0],"diffusion":[1,53],"models":[2],"enable":[3,78],"efficient":[4],"high-resolution":[5],"image":[6],"synthesis":[7],"but":[8],"struggle":[9],"to":[10,17,72,94],"align":[11],"with":[12,28],"specific":[13],"downstream":[14],"objectives":[15],"due":[16],"limitations":[18],"of":[19],"existing":[20],"reinforcement":[21],"learning":[22,110],"(RL)":[23],"methods":[24],"in":[25],"low-step":[26,137],"regimes":[27],"limited":[29],"state":[30],"spaces":[31],"and":[32,65,77,116,129,139,156],"suboptimal":[33],"sample":[34],"quality.":[35],"To":[36],"address":[37],"this,":[38],"we":[39,85],"propose":[40],"Stepwise":[41],"Diffusion":[42],"Policy":[43],"Optimization":[44],"(SDPO),":[45],"a":[46,57,87,106],"novel":[47],"RL":[48],"framework":[49],"tailored":[50],"for":[51],"few-step":[52,154],"models.":[54],"SDPO":[55,104,146],"introduces":[56],"dual-state":[58],"trajectory":[59],"sampling":[60],"mechanism,":[61],"tracking":[62],"both":[63],"noisy":[64],"predicted":[66],"clean":[67],"states":[68],"at":[69],"each":[70],"step":[71],"provide":[73],"dense":[74,90,97,102,107],"reward":[75,91,98,108],"feedback":[76],"low-variance,":[79],"mixed-step":[80],"optimization.":[81],"For":[82],"further":[83,133],"efficiency,":[84],"develop":[86],"latent":[88],"similarity-based":[89],"prediction":[92],"strategy":[93],"minimize":[95],"costly":[96],"queries.":[99],"Leveraging":[100],"these":[101],"rewards,":[103],"optimizes":[105],"difference":[109],"objective":[111],"that":[112,145],"enables":[113],"more":[114],"frequent":[115],"granular":[117],"policy":[118],"updates.":[119],"Additional":[120],"refinements,":[121],"including":[122],"stepwise":[123],"advantage":[124],"estimates,":[125],"temporal":[126],"importance":[127],"weighting,":[128],"step-shuffled":[130],"gradient":[131,140],"updates,":[132],"enhance":[134],"long-term":[135],"dependency,":[136],"priority,":[138],"stability.":[141],"Our":[142],"experiments":[143],"demonstrate":[144],"consistently":[147],"delivers":[148],"superior":[149],"reward-aligned":[150],"results":[151],"across":[152],"diverse":[153],"settings":[155],"tasks.":[157]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-07-15T18:14:33.161393","created_date":"2024-11-21T00:00:00"}
