{"id":"https://openalex.org/W7211896151","doi":"https://doi.org/10.1145/3767308.3835988","title":"Step Back to Move Forward: Reflection-Aware Preference Optimization for Visual Generation","display_name":"Step Back to Move Forward: Reflection-Aware Preference Optimization for Visual Generation","publication_year":2026,"publication_date":"2026-11-05","ids":{"openalex":"https://openalex.org/W7211896151","doi":"https://doi.org/10.1145/3767308.3835988"},"language":"en","primary_location":{"id":"doi:10.1145/3767308.3835988","doi":"https://doi.org/10.1145/3767308.3835988","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3767308.3835988","pdf_url":null,"source":{"id":"https://openalex.org/S4363608757","display_name":"ACM International Conference on Multimedia (ACM MM)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 34th ACM International Conference on Multimedia","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.1145/3767308.3835988","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5150570475","display_name":"Junlong Wu","orcid":"https://orcid.org/0009-0000-9744-1045","observed_orcids":["https://orcid.org/0009-0000-9744-1045"]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Junlong Wu","raw_affiliation_strings":["Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0000-9744-1045","affiliations":[{"raw_affiliation_string":"Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136020620","display_name":"Jiuzhou Lin","orcid":"https://orcid.org/0009-0005-7730-0479","observed_orcids":["https://orcid.org/0009-0005-7730-0479"]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiuzhou Lin","raw_affiliation_strings":["Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0005-7730-0479","affiliations":[{"raw_affiliation_string":"Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126236846","display_name":"Jia Sun","orcid":"https://orcid.org/0009-0009-7688-5726","observed_orcids":["https://orcid.org/0009-0009-7688-5726"]},"institutions":[{"id":"https://openalex.org/I4401726859","display_name":"Kuaishou (China)","ror":"https://ror.org/0258as409","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401726859"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jia Sun","raw_affiliation_strings":["Kuaishou Technology, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0009-7688-5726","affiliations":[{"raw_affiliation_string":"Kuaishou Technology, Beijing, China","institution_ids":["https://openalex.org/I4401726859"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027860856","display_name":"Boheng Zhang","orcid":"https://orcid.org/0000-0001-6040-2210","observed_orcids":["https://orcid.org/0000-0001-6040-2210","https://orcid.org/0000-0003-1185-3239"]},"institutions":[{"id":"https://openalex.org/I4401726859","display_name":"Kuaishou (China)","ror":"https://ror.org/0258as409","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401726859"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Boheng Zhang","raw_affiliation_strings":["Kuaishou Technology, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0003-1185-3239","affiliations":[{"raw_affiliation_string":"Kuaishou Technology, Beijing, China","institution_ids":["https://openalex.org/I4401726859"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126253104","display_name":"Huaiqing Wang","orcid":"https://orcid.org/0009-0007-0429-0948","observed_orcids":["https://orcid.org/0009-0007-0429-0948"]},"institutions":[{"id":"https://openalex.org/I4401726859","display_name":"Kuaishou (China)","ror":"https://ror.org/0258as409","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401726859"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Huaiqing Wang","raw_affiliation_strings":["Kuaishou Technology, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0007-0429-0948","affiliations":[{"raw_affiliation_string":"Kuaishou Technology, Beijing, China","institution_ids":["https://openalex.org/I4401726859"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5016757091","display_name":"Dewen Fan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4401726859","display_name":"Kuaishou (China)","ror":"https://ror.org/0258as409","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401726859"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dewen Fan","raw_affiliation_strings":["Kuaishou Technology, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-6062-4208","affiliations":[{"raw_affiliation_string":"Kuaishou Technology, Beijing, China","institution_ids":["https://openalex.org/I4401726859"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5150613838","display_name":"Houde Liu","orcid":"https://orcid.org/0000-0002-7314-3366","observed_orcids":["https://orcid.org/0000-0002-7314-3366"]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Houde Liu","raw_affiliation_strings":["Tsinghua University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-7314-3366","affiliations":[{"raw_affiliation_string":"Tsinghua University, Beijing, China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017594071","display_name":"Qianqian Gan","orcid":"https://orcid.org/0009-0003-1812-8993","observed_orcids":["https://orcid.org/0009-0003-1812-8993"]},"institutions":[{"id":"https://openalex.org/I4401726859","display_name":"Kuaishou (China)","ror":"https://ror.org/0258as409","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401726859"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Qianqian Gan","raw_affiliation_strings":["Kuaishou Technology, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0003-1812-8993","affiliations":[{"raw_affiliation_string":"Kuaishou Technology, Beijing, China","institution_ids":["https://openalex.org/I4401726859"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140290394","display_name":"Fan Yang","orcid":"https://orcid.org/0009-0005-4570-5885","observed_orcids":["https://orcid.org/0009-0005-4570-5885"]},"institutions":[{"id":"https://openalex.org/I4401726859","display_name":"Kuaishou (China)","ror":"https://ror.org/0258as409","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401726859"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fan Yang","raw_affiliation_strings":["Kuaishou Technology, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0005-4570-5885","affiliations":[{"raw_affiliation_string":"Kuaishou Technology, Beijing, China","institution_ids":["https://openalex.org/I4401726859"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5125518543","display_name":"Tingting Gao","orcid":"https://orcid.org/0009-0003-0310-7751","observed_orcids":["https://orcid.org/0009-0003-0310-7751"]},"institutions":[{"id":"https://openalex.org/I4401726859","display_name":"Kuaishou (China)","ror":"https://ror.org/0258as409","country_code":"CN","type":"company","lineage":["https://openalex.org/I4401726859"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tingting Gao","raw_affiliation_strings":["Kuaishou Technology, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0003-0310-7751","affiliations":[{"raw_affiliation_string":"Kuaishou Technology, Beijing, China","institution_ids":["https://openalex.org/I4401726859"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.0015,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.92478528,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":null,"issue":null,"first_page":"11574","last_page":"11583"},"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":0.9994999766349792,"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":0.9994999766349792,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":9.999999747378752e-05,"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"}},{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":9.999999747378752e-05,"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/preference-optimization","display_name":"preference optimization","score":0.824999988079071},{"id":"https://openalex.org/keywords/diffusion-models","display_name":"diffusion models","score":0.8040000200271606},{"id":"https://openalex.org/keywords/text-to-image-generation","display_name":"text-to-image generation","score":0.8460000157356262},{"id":"https://openalex.org/keywords/text-to-video-generation","display_name":"text-to-video generation","score":0.9990000128746033},{"id":"https://openalex.org/keywords/reward-hacking","display_name":"reward hacking","score":0.6769999861717224},{"id":"https://openalex.org/keywords/backward-reflection","display_name":"backward reflection","score":0.4569999873638153},{"id":"https://openalex.org/keywords/preference-alignment","display_name":"preference alignment","score":0.414000004529953}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7311999797821045},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.5852000117301941},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5741000175476074},{"id":"https://openalex.org/C108650721","wikidata":"https://www.wikidata.org/wiki/Q1783253","display_name":"Counterfactual thinking","level":2,"score":0.5580000281333923},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5296000242233276},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.4796999990940094},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.47380000352859497},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.4595000147819519},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.45829999446868896},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.45590001344680786},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.36640000343322754},{"id":"https://openalex.org/C2776196576","wikidata":"https://www.wikidata.org/wiki/Q196113","display_name":"Camouflage","level":2,"score":0.3497999906539917},{"id":"https://openalex.org/C2779405079","wikidata":"https://www.wikidata.org/wiki/Q356040","display_name":"Jigsaw","level":2,"score":0.31459999084472656},{"id":"https://openalex.org/C123197309","wikidata":"https://www.wikidata.org/wiki/Q2882343","display_name":"Multi-armed bandit","level":3,"score":0.3102000057697296},{"id":"https://openalex.org/C65682993","wikidata":"https://www.wikidata.org/wiki/Q1056451","display_name":"Reflection (computer programming)","level":2,"score":0.28690001368522644},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.2799000144004822},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.27869999408721924},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2768000066280365},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.2736000120639801},{"id":"https://openalex.org/C73602740","wikidata":"https://www.wikidata.org/wiki/Q7795822","display_name":"Thompson sampling","level":3,"score":0.26829999685287476},{"id":"https://openalex.org/C2779321571","wikidata":"https://www.wikidata.org/wiki/Q7936605","display_name":"Visual learning","level":2,"score":0.2671999931335449},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.26649999618530273},{"id":"https://openalex.org/C43091099","wikidata":"https://www.wikidata.org/wiki/Q1067788","display_name":"Through-the-lens metering","level":3,"score":0.26489999890327454},{"id":"https://openalex.org/C2780440489","wikidata":"https://www.wikidata.org/wiki/Q5227278","display_name":"Data-driven","level":2,"score":0.26089999079704285},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.25440001487731934}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1145/3767308.3835988","doi":"https://doi.org/10.1145/3767308.3835988","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3767308.3835988","pdf_url":null,"source":{"id":"https://openalex.org/S4363608757","display_name":"ACM International Conference on Multimedia (ACM MM)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 34th ACM International Conference on Multimedia","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2609.04282","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.04282","pdf_url":"https://arxiv.org/pdf/2609.04282","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.04282","doi":"https://doi.org/10.48550/arxiv.2609.04282","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.04282","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,"listed_in":[],"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":"Preprint"}],"best_oa_location":{"id":"doi:10.1145/3767308.3835988","doi":"https://doi.org/10.1145/3767308.3835988","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3767308.3835988","pdf_url":null,"source":{"id":"https://openalex.org/S4363608757","display_name":"ACM International Conference on Multimedia (ACM MM)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 34th ACM International Conference on Multimedia","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7211896151.pdf","grobid_xml":"https://content.openalex.org/works/W7211896151.grobid-xml"},"referenced_works_count":39,"referenced_works":["https://openalex.org/W2736601468","https://openalex.org/W3092442149","https://openalex.org/W3110257065","https://openalex.org/W3153469116","https://openalex.org/W4223908421","https://openalex.org/W4303647933","https://openalex.org/W4378464471","https://openalex.org/W4380994608","https://openalex.org/W4383472654","https://openalex.org/W4384918448","https://openalex.org/W4385436553","https://openalex.org/W4391631327","https://openalex.org/W4396665924","https://openalex.org/W4402753790","https://openalex.org/W4405468733","https://openalex.org/W4406772739","https://openalex.org/W4407123638","https://openalex.org/W4407124205","https://openalex.org/W4409092337","https://openalex.org/W4409362098","https://openalex.org/W4414449548","https://openalex.org/W4414560664","https://openalex.org/W4414634804","https://openalex.org/W4415058126","https://openalex.org/W4415332679","https://openalex.org/W4415797331","https://openalex.org/W4416070669","https://openalex.org/W4417517408","https://openalex.org/W6967046348","https://openalex.org/W6967052438","https://openalex.org/W7101437281","https://openalex.org/W7133195264","https://openalex.org/W7133211372","https://openalex.org/W7133217041","https://openalex.org/W7133222210","https://openalex.org/W7133236450","https://openalex.org/W7160104918","https://openalex.org/W7160178070","https://openalex.org/W7167206815"],"related_works":["https://openalex.org/W7162408510","https://openalex.org/W4407184534","https://openalex.org/W4416189736","https://openalex.org/W7202150354","https://openalex.org/W7114793404","https://openalex.org/W7160537411","https://openalex.org/W4413144755","https://openalex.org/W7106695053","https://openalex.org/W7155209452","https://openalex.org/W4417517408"],"abstract_inverted_index":{"Diffusion":[0,103],"models":[1,29,165],"have":[2,12],"become":[3],"the":[4,112,126,143,146,150],"mainstream":[5],"paradigm":[6],"for":[7,82,195],"modern":[8],"visual":[9,65],"generation":[10,93],"and":[11,21,64,163,178,185],"substantially":[13],"advanced":[14],"multimedia":[15],"content":[16],"synthesis,":[17],"especially":[18],"in":[19,174],"text-to-image":[20],"text-to-video":[22],"tasks.":[23],"To":[24,67],"further":[25],"align":[26],"such":[27],"generative":[28,84],"with":[30,115,188],"human":[31],"preferences,":[32],"reinforcement":[33],"learning":[34],"(RL)":[35],"has":[36],"recently":[37],"shown":[38],"strong":[39],"potential":[40],"as":[41],"a":[42,76,116,192],"post-training":[43],"strategy.":[44],"Nevertheless,":[45],"existing":[46,171],"policy":[47],"gradient-based":[48],"methods":[49],"often":[50],"explore":[51],"inefficiently,":[52],"making":[53],"them":[54],"vulnerable":[55],"to":[56,90,136,148],"local":[57],"optima":[58],"that":[59,167],"may":[60],"degrade":[61],"semantic":[62],"faithfulness":[63],"realism.":[66],"address":[68],"these":[69,139],"challenges,":[70],"we":[71,131],"present":[72],"Reflection-Aware":[73],"GRPO":[74],"(RA-GRPO),":[75],"new":[77],"RL-based":[78],"preference":[79,197],"alignment":[80],"framework":[81],"diffusion":[83,113],"models.":[85],"The":[86,181],"core":[87],"idea":[88],"is":[89],"improve":[91],"'forward'":[92],"by":[94,110],"incorporating":[95],"'backward'":[96],"reflection":[97],"during":[98],"optimization.":[99],"We":[100],"first":[101],"introduce":[102,132],"Reflection,":[104],"which":[105],"rectifies":[106],"intermediate":[107],"sampling":[108],"trajectories":[109,141],"inverting":[111],"process":[114],"weak":[117],"estimator,":[118],"guiding":[119],"latent":[120],"states":[121],"toward":[122],"higher-probability":[123],"regions":[124],"of":[125,152],"true":[127],"data":[128],"manifold.":[129],"Furthermore,":[130],"Counterfactual":[133],"Path":[134],"Synthesis":[135],"implicitly":[137],"distill":[138],"rectified":[140],"into":[142],"policy,":[144],"enabling":[145],"model":[147],"internalize":[149],"benefits":[151],"search-based":[153],"exploration":[154],"without":[155],"incurring":[156],"inference-time":[157],"overhead.":[158],"Extensive":[159],"experiments":[160],"on":[161],"T2I":[162],"T2V":[164],"demonstrate":[166],"RA-GRPO":[168],"significantly":[169],"outperforms":[170],"methods,":[172],"particularly":[173],"mitigating":[175],"reward":[176],"hacking":[177],"improving":[179],"generalization.":[180],"method":[182],"remains":[183],"architecture-agnostic":[184],"integrates":[186],"seamlessly":[187],"standard":[189],"pipelines,":[190],"suggesting":[191],"promising":[193],"direction":[194],"stable":[196],"alignment.":[198]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-08T00:00:00"}