{"id":"https://openalex.org/W7128619172","doi":"https://doi.org/10.48550/arxiv.2602.09449","title":"Look-Ahead and Look-Back Flows: Training-Free Image Generation with Trajectory Smoothing","display_name":"Look-Ahead and Look-Back Flows: Training-Free Image Generation with Trajectory Smoothing","publication_year":2026,"publication_date":"2026-02-10","ids":{"openalex":"https://openalex.org/W7128619172","doi":"https://doi.org/10.48550/arxiv.2602.09449"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2602.09449","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":null,"license_id":null,"version":"submittedVersion","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":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125640271","display_name":"Yan Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Yan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Huang, Henry","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Henry","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125633857","display_name":"Todd Y. Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Todd Y.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5125651343","display_name":"Mengyu Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Mengyu","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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.8245999813079834,"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":0.8245999813079834,"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":0.1080000028014183,"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/T10481","display_name":"Computer Graphics and Visualization Techniques","score":0.010499999858438969,"subfield":{"id":"https://openalex.org/subfields/1704","display_name":"Computer Graphics and Computer-Aided Design"},"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/trajectory","display_name":"Trajectory","score":0.8177000284194946},{"id":"https://openalex.org/keywords/smoothing","display_name":"Smoothing","score":0.7462999820709229},{"id":"https://openalex.org/keywords/exponential-smoothing","display_name":"Exponential smoothing","score":0.5278000235557556},{"id":"https://openalex.org/keywords/path","display_name":"Path (computing)","score":0.4325999915599823},{"id":"https://openalex.org/keywords/flow","display_name":"Flow (mathematics)","score":0.42980000376701355},{"id":"https://openalex.org/keywords/field","display_name":"Field (mathematics)","score":0.4219000041484833},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.41179999709129333}],"concepts":[{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.8177000284194946},{"id":"https://openalex.org/C3770464","wikidata":"https://www.wikidata.org/wiki/Q775963","display_name":"Smoothing","level":2,"score":0.7462999820709229},{"id":"https://openalex.org/C133710760","wikidata":"https://www.wikidata.org/wiki/Q775837","display_name":"Exponential smoothing","level":2,"score":0.5278000235557556},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4925000071525574},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.47130000591278076},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.4325999915599823},{"id":"https://openalex.org/C38349280","wikidata":"https://www.wikidata.org/wiki/Q1434290","display_name":"Flow (mathematics)","level":2,"score":0.42980000376701355},{"id":"https://openalex.org/C9652623","wikidata":"https://www.wikidata.org/wiki/Q190109","display_name":"Field (mathematics)","level":2,"score":0.4219000041484833},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.41179999709129333},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.41179999709129333},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.384799987077713},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3562999963760376},{"id":"https://openalex.org/C91188154","wikidata":"https://www.wikidata.org/wiki/Q186247","display_name":"Vector field","level":2,"score":0.35530000925064087},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.31369999051094055},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.2964000105857849},{"id":"https://openalex.org/C151376022","wikidata":"https://www.wikidata.org/wiki/Q168698","display_name":"Exponential function","level":2,"score":0.29330000281333923},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.29010000824928284},{"id":"https://openalex.org/C148043351","wikidata":"https://www.wikidata.org/wiki/Q4456944","display_name":"Current (fluid)","level":2,"score":0.26460000872612},{"id":"https://openalex.org/C51544822","wikidata":"https://www.wikidata.org/wiki/Q465274","display_name":"Ordinary differential equation","level":3,"score":0.2639999985694885},{"id":"https://openalex.org/C93226319","wikidata":"https://www.wikidata.org/wiki/Q193137","display_name":"Differential (mechanical device)","level":2,"score":0.2549000084400177}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2602.09449","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2602.09449","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.09449","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:doi:10.48550/arxiv.2602.09449","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":null,"license_id":null,"version":"submittedVersion","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":{"Recent":[0],"advances":[1],"have":[2,32],"reformulated":[3],"diffusion":[4],"models":[5,161,166],"as":[6],"deterministic":[7],"ordinary":[8],"differential":[9],"equations":[10],"(ODEs)":[11],"through":[12,39,60,148],"the":[13,23,45,52,61,68,76,108,125,156],"framework":[14],"of":[15],"flow":[16,29,40],"matching,":[17],"providing":[18],"a":[19,131],"unified":[20],"formulation":[21],"for":[22,47],"noise-to-data":[24],"generative":[25,109],"process.":[26],"Various":[27],"training-free":[28,90,118,158],"matching":[30],"approaches":[31,93],"been":[33],"developed":[34],"to":[35,67],"improve":[36],"image":[37],"generation":[38,63],"velocity":[41,53,78,99],"field":[42,54],"adjustment,":[43],"eliminating":[44],"need":[46],"costly":[48],"retraining.":[49],"However,":[50],"Modifying":[51],"$v$":[55,100],"introduces":[56],"errors":[57],"that":[58,106,155],"propagate":[59],"full":[62],"path,":[64],"whereas":[65],"adjustments":[66],"latent":[69,102,113],"trajectory":[70,103,119,159],"$z$":[71,104],"are":[72],"naturally":[73],"corrected":[74],"by":[75],"pretrained":[77],"network,":[79],"reducing":[80],"error":[81],"accumulation.":[82],"In":[83],"this":[84],"paper,":[85],"we":[86],"propose":[87,116],"two":[88,117],"complementary":[89],"latent-trajectory":[91],"adjustment":[92],"based":[94],"on":[95],"future":[96],"and":[97,101,127,134,151,173],"past":[98],"information":[105],"refine":[107],"path":[110],"directly":[111],"in":[112],"space.":[114],"We":[115,146],"smoothing":[120,160],"schemes:":[121],"\\emph{Look-Ahead},":[122],"which":[123,136],"averages":[124],"current":[126],"next-step":[128],"latents":[129,138],"using":[130,139],"curvature-gated":[132],"weight,":[133],"\\emph{Look-Back},":[135],"smoothes":[137],"an":[140],"exponential":[141],"moving":[142],"average":[143],"with":[144],"decay.":[145],"demonstrate":[147],"extensive":[149],"experiments":[150],"comprehensive":[152],"evaluation":[153],"metrics":[154],"proposed":[157],"substantially":[162],"outperform":[163],"various":[164],"state-of-the-art":[165],"across":[167],"multiple":[168],"datasets":[169],"including":[170],"COCO17,":[171],"CUB-200,":[172],"Flickr30K.":[174]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-02-12T00:00:00"}