{"id":"https://openalex.org/W7164810449","doi":"https://doi.org/10.48550/arxiv.2606.14657","title":"HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities","display_name":"HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities","publication_year":2026,"publication_date":"2026-06-12","ids":{"openalex":"https://openalex.org/W7164810449","doi":"https://doi.org/10.48550/arxiv.2606.14657"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.14657","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.14657","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.2606.14657","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138632397","display_name":"Yijun Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Yijun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006022044","display_name":"J W Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Jie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5040083129","display_name":"Zeyue Xue","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xue, Zeyue","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138664933","display_name":"Yuming Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Yuming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136450802","display_name":"Ruizhe He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Ruizhe","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138674482","display_name":"Haoran Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Haoran","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5062661679","display_name":"Shijia Ge","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ge, Shijia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5083298839","display_name":"Siming Fu","orcid":"https://orcid.org/0000-0003-3257-1011"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fu, Siming","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.39070001244544983,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.39070001244544983,"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/T11605","display_name":"Visual Attention and Saliency Detection","score":0.16740000247955322,"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/T10203","display_name":"Recommender Systems and Techniques","score":0.06930000334978104,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/fidelity","display_name":"Fidelity","score":0.656000018119812},{"id":"https://openalex.org/keywords/discriminative-model","display_name":"Discriminative model","score":0.6442999839782715},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.621399998664856},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.49239999055862427},{"id":"https://openalex.org/keywords/consistency","display_name":"Consistency (knowledge bases)","score":0.46050000190734863},{"id":"https://openalex.org/keywords/projection","display_name":"Projection (relational algebra)","score":0.454800009727478},{"id":"https://openalex.org/keywords/preference","display_name":"Preference","score":0.45239999890327454},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.44449999928474426},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.3889999985694885}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7228999733924866},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.656000018119812},{"id":"https://openalex.org/C97931131","wikidata":"https://www.wikidata.org/wiki/Q5282087","display_name":"Discriminative model","level":2,"score":0.6442999839782715},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6363000273704529},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.621399998664856},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.49239999055862427},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4702000021934509},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.46050000190734863},{"id":"https://openalex.org/C57493831","wikidata":"https://www.wikidata.org/wiki/Q3134666","display_name":"Projection (relational algebra)","level":2,"score":0.454800009727478},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.45239999890327454},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.44449999928474426},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.3889999985694885},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.3847000002861023},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.3659000098705292},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3617999851703644},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.3443000018596649},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.31520000100135803},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.3147999942302704},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.31189998984336853},{"id":"https://openalex.org/C113364801","wikidata":"https://www.wikidata.org/wiki/Q26674","display_name":"High fidelity","level":2,"score":0.30820000171661377},{"id":"https://openalex.org/C181204326","wikidata":"https://www.wikidata.org/wiki/Q7239820","display_name":"Preference learning","level":3,"score":0.2978000044822693},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.290800005197525},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.29010000824928284},{"id":"https://openalex.org/C8038995","wikidata":"https://www.wikidata.org/wiki/Q1152135","display_name":"Unsupervised learning","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.25850000977516174},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.2547000050544739}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.14657","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.14657","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.2606.14657","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.14657","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":[{"score":0.7591015696525574,"id":"https://metadata.un.org/sdg/10","display_name":"Reduced inequalities"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reward":[0],"models":[1,15,143],"guide":[2],"text-to-image":[3],"(T2I)":[4],"systems":[5],"toward":[6],"outputs":[7],"aligned":[8],"with":[9,101,163],"human":[10,102,130],"preferences.":[11],"However,":[12],"typical":[13],"reward":[14,55,160,171],"such":[16],"as":[17],"HPSv3":[18,61,183],"are":[19],"trained":[20],"on":[21,185,188,193],"pre-annotated":[22],"data":[23,140],"from":[24,35,123,141],"earlier":[25],"T2I":[26,65,142,200,210],"models,":[27,211],"without":[28],"accounting":[29],"for":[30,63,89,158,199],"quality":[31,94],"discriminative":[32],"shifts":[33],"arising":[34],"evolving":[36],"model":[37,56,62,66,100,161,172],"capabilities":[38,67],"and":[39,68,92,148,151],"reinforcement":[40],"learning":[41],"(RL)":[42],"iterations,":[43,150],"limiting":[44],"their":[45,69],"broader":[46],"applicability.":[47],"In":[48],"this":[49],"work,":[50],"we":[51,79],"propose":[52,106],"HPSv3++,":[53],"a":[54,83,96,107,153,164],"framework":[57],"that":[58],"elevates":[59],"the":[60,74,127,159,174],"varying":[64],"RL":[70,149,201],"iteration":[71],"changes":[72],"across":[73,173,208],"full":[75],"capability-iteration":[76,175],"spectrum.":[77,176],"Specifically,":[78],"first":[80],"introduce":[81],"HPDv3++,":[82],"212K":[84],"dual-dimension":[85],"preference":[86,131,180],"dataset":[87],"annotated":[88],"text":[90],"fidelity":[91],"aesthetic":[93,121],"using":[95],"recent":[97],"high-capability":[98],"(Qwen-Image)":[99],"supervision.":[103],"We":[104],"then":[105],"two-stage":[108],"training":[109],"framework.":[110],"Stage":[111,135],"1":[112],"employs":[113],"data-aware":[114],"orthogonal":[115],"gradient":[116],"projection":[117],"to":[118],"incorporate":[119],"diverse":[120,209],"perception":[122],"HPDv3++":[124],"while":[125,190],"preserving":[126],"original":[128],"effective":[129],"knowledge":[132],"in":[133],"HPSv3.":[134],"2":[136],"further":[137],"leverages":[138],"unlabeled":[139],"spanning":[144],"different":[145],"capability":[146],"levels":[147],"introduces":[152],"joint":[154],"capability-iterations":[155],"conditioned":[156],"signal":[157],"together":[162],"standard":[165],"deviation-driven":[166],"unsupervised":[167],"guidance":[168],"mechanism,":[169],"strengthening":[170],"HPSv3++":[177],"achieves":[178],"state-of-the-art":[179],"prediction,":[181],"outperforming":[182],"9.8%":[184],"HPDv3,":[186],"5.5%":[187],"GenAI-Bench,":[189],"achieving":[191],"79.1%/88.1%":[192],"our":[194],"proposed":[195],"HPDv3++.":[196],"When":[197],"used":[198],"training,":[202],"it":[203],"consistently":[204],"improves":[205],"GenEval":[206],"scores":[207],"demonstrating":[212],"its":[213],"wide-range":[214],"capabilities.":[215],"The":[216],"code":[217],"is":[218],"available":[219],"at":[220],"https://github.com/PlantPotatoOnMoon/HPSv3-PlusPlus.":[221]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-16T00:00:00"}