{"id":"https://openalex.org/W7138220373","doi":"https://doi.org/10.48550/arxiv.2603.14526","title":"LatSearch: Latent Reward-Guided Search for Faster Inference-Time Scaling in Video Diffusion","display_name":"LatSearch: Latent Reward-Guided Search for Faster Inference-Time Scaling in Video Diffusion","publication_year":2026,"publication_date":"2026-03-15","ids":{"openalex":"https://openalex.org/W7138220373","doi":"https://doi.org/10.48550/arxiv.2603.14526"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.14526","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.14526","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.2603.14526","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5032446974","display_name":"Zengqun Zhao","orcid":"https://orcid.org/0000-0002-7135-1946"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Zhao, Zengqun","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Liu, Ziquan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Ziquan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129671268","display_name":"Yu Cao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cao, Yu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129694573","display_name":"Shaogang Gong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gong, Shaogang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129742186","display_name":"Zhensong Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Zhensong","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5046874089","display_name":"Jifei Song","orcid":"https://orcid.org/0000-0002-3381-6685"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song, Jifei","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129729785","display_name":"Jiankang Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deng, Jiankang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5031205865","display_name":"Ioannis Patras","orcid":"https://orcid.org/0000-0003-3913-4738"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Patras, Ioannis","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5032446974"],"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":0.3610000014305115,"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":0.3610000014305115,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.07859999686479568,"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/T10860","display_name":"Speech and Audio Processing","score":0.06069999933242798,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/resampling","display_name":"Resampling","score":0.6765999794006348},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.5590999722480774},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.5565000176429749},{"id":"https://openalex.org/keywords/prior-probability","display_name":"Prior probability","score":0.5544000267982483},{"id":"https://openalex.org/keywords/noise-reduction","display_name":"Noise reduction","score":0.49900001287460327},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.4708999991416931},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4535999894142151},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.4422000050544739},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.4413999915122986}],"concepts":[{"id":"https://openalex.org/C150921843","wikidata":"https://www.wikidata.org/wiki/Q1170431","display_name":"Resampling","level":2,"score":0.6765999794006348},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6614999771118164},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.5590999722480774},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5565000176429749},{"id":"https://openalex.org/C177769412","wikidata":"https://www.wikidata.org/wiki/Q278090","display_name":"Prior probability","level":3,"score":0.5544000267982483},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5087000131607056},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.49900001287460327},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.4708999991416931},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4535999894142151},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.4422000050544739},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.4413999915122986},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.43869999051094055},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.41909998655319214},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.40700000524520874},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.38370001316070557},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.3801000118255615},{"id":"https://openalex.org/C103910844","wikidata":"https://www.wikidata.org/wiki/Q2631256","display_name":"Video quality","level":3,"score":0.37389999628067017},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.36469998955726624},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.36399999260902405},{"id":"https://openalex.org/C30814859","wikidata":"https://www.wikidata.org/wiki/Q4119603","display_name":"Video denoising","level":5,"score":0.3495999872684479},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3490000069141388},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3271999955177307},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.31290000677108765},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3109000027179718},{"id":"https://openalex.org/C202474056","wikidata":"https://www.wikidata.org/wiki/Q1931635","display_name":"Video tracking","level":3,"score":0.29030001163482666},{"id":"https://openalex.org/C3770464","wikidata":"https://www.wikidata.org/wiki/Q775963","display_name":"Smoothing","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C62644790","wikidata":"https://www.wikidata.org/wiki/Q3454689","display_name":"Variance reduction","level":3,"score":0.2667999863624573},{"id":"https://openalex.org/C137800194","wikidata":"https://www.wikidata.org/wiki/Q11713455","display_name":"Interpolation (computer graphics)","level":3,"score":0.2662999927997589},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.26510000228881836},{"id":"https://openalex.org/C4069607","wikidata":"https://www.wikidata.org/wiki/Q868732","display_name":"Aliasing","level":3,"score":0.25519999861717224},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.2515000104904175}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.14526","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.14526","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.2603.14526","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.14526","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":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"recent":[1],"success":[2],"of":[3,23,60,92],"inference-time":[4,132,185],"scaling":[5,133],"in":[6,14,107,126],"large":[7],"language":[8],"models":[9],"has":[10,32],"inspired":[11],"similar":[12],"explorations":[13],"video":[15,28,115,135,251],"diffusion.":[16],"In":[17,195,213],"particular,":[18],"motivated":[19],"by":[20,37],"the":[21,58,69,90,149,196,210,214,219,224,227,242,259],"existence":[22],"\"golden":[24],"noise\"":[25],"that":[26,158,188,247],"enhances":[27],"quality,":[29,170,172],"prior":[30],"work":[31],"attempted":[33],"to":[34,76,168,203,206,258],"improve":[35],"inference":[36],"optimising":[38],"or":[39,63],"searching":[40],"for":[41,114,134],"better":[42],"initial":[43],"noise.":[44],"However,":[45],"these":[46],"approaches":[47],"have":[48],"notable":[49],"limitations:":[50],"they":[51],"either":[52],"rely":[53],"on":[54,64,68,177,209,241],"priors":[55],"imposed":[56],"at":[57,163,218],"beginning":[59],"noise":[61],"sampling":[62],"rewards":[65],"evaluated":[66],"only":[67,223],"denoised":[70,161],"and":[71,80,84,111,145,173,192,236,245],"decoded":[72],"videos.":[73],"This":[74],"leads":[75],"error":[77],"accumulation,":[78],"delayed":[79],"sparse":[81],"reward":[82,139,156,211,230],"signals,":[83],"prohibitive":[85],"computational":[86,119],"cost,":[87],"which":[88,141],"prevents":[89],"use":[91],"stronger":[93,97],"search":[94,98,186],"algorithms.":[95],"Crucially,":[96],"algorithms":[99],"are":[100,200],"precisely":[101],"what":[102],"could":[103],"unlock":[104],"substantial":[105],"gains":[106],"controllability,":[108],"sample":[109],"efficiency":[110],"generation":[112,252],"quality":[113,235],"diffusion,":[116],"provided":[117],"their":[118],"cost":[120],"can":[121],"be":[122],"reduced.":[123],"To":[124],"fill":[125],"this":[127,178],"gap,":[128],"we":[129,180],"enable":[130],"efficient":[131,146],"diffusion":[136],"through":[137],"latent":[138,155],"guidance,":[140],"provides":[142],"intermediate,":[143],"informative":[144],"feedback":[147],"along":[148],"denoising":[150],"trajectory.":[151],"We":[152,238],"introduce":[153],"a":[154,183],"model":[157],"scores":[159],"partially":[160],"latents":[162],"arbitrary":[164],"timesteps":[165],"with":[166,226],"respect":[167],"visual":[169],"motion":[171],"text":[174],"alignment.":[175],"Building":[176],"model,":[179],"propose":[181],"LatSearch,":[182],"novel":[184],"mechanism":[187],"performs":[189],"Reward-Guided":[190],"Resampling":[191],"Pruning":[193],"(RGRP).":[194],"resampling":[197],"stage,":[198,216],"candidates":[199],"sampled":[201],"according":[202],"reward-normalised":[204],"probabilities":[205],"reduce":[207],"over-reliance":[208],"model.":[212,262],"pruning":[215],"applied":[217],"final":[220],"scheduled":[221],"step,":[222],"candidate":[225],"highest":[228],"cumulative":[229],"is":[231],"retained,":[232],"improving":[233],"both":[234],"efficiency.":[237],"evaluate":[239],"LatSearch":[240],"VBench-2.0":[243],"benchmark":[244],"demonstrate":[246],"it":[248],"consistently":[249],"improves":[250],"across":[253],"multiple":[254],"evaluation":[255],"dimensions":[256],"compared":[257],"baseline":[260],"Wan2.1":[261]},"counts_by_year":[],"updated_date":"2026-04-21T08:09:41.155169","created_date":"2026-03-18T00:00:00"}