{"id":"https://openalex.org/W7147333421","doi":"https://doi.org/10.48550/arxiv.2603.27185","title":"MotionRFT: Unified Reinforcement Fine-Tuning for Text-to-Motion Generation","display_name":"MotionRFT: Unified Reinforcement Fine-Tuning for Text-to-Motion Generation","publication_year":2026,"publication_date":"2026-03-28","ids":{"openalex":"https://openalex.org/W7147333421","doi":"https://doi.org/10.48550/arxiv.2603.27185"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.27185","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.27185","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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","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.2603.27185","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5132599481","display_name":"Xiaofeng Tan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tan, Xiaofeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132628879","display_name":"Wanjiang Weng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Weng, Wanjiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132618514","display_name":"Hongsong Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Hongsong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132669585","display_name":"Fang Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Fang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132568407","display_name":"Xin Geng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Geng, Xin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5132573101","display_name":"Liang Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Liang","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/T12290","display_name":"Human Motion and Animation","score":0.8098999857902527,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12290","display_name":"Human Motion and Animation","score":0.8098999857902527,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.07150000333786011,"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/T10719","display_name":"3D Shape Modeling and Analysis","score":0.025200000032782555,"subfield":{"id":"https://openalex.org/subfields/2206","display_name":"Computational Mechanics"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.7103000283241272},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.6829000115394592},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.6040999889373779},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.45559999346733093},{"id":"https://openalex.org/keywords/reinforcement","display_name":"Reinforcement","score":0.3898000121116638},{"id":"https://openalex.org/keywords/motion","display_name":"Motion (physics)","score":0.3465999960899353},{"id":"https://openalex.org/keywords/space","display_name":"Space (punctuation)","score":0.3407999873161316},{"id":"https://openalex.org/keywords/generative-model","display_name":"Generative model","score":0.34060001373291016}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7547000050544739},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7103000283241272},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.6829000115394592},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.6040999889373779},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5604000091552734},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4724999964237213},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.45559999346733093},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.3898000121116638},{"id":"https://openalex.org/C104114177","wikidata":"https://www.wikidata.org/wiki/Q79782","display_name":"Motion (physics)","level":2,"score":0.3465999960899353},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.3407999873161316},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.34060001373291016},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.33799999952316284},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.337799996137619},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.33480000495910645},{"id":"https://openalex.org/C136389625","wikidata":"https://www.wikidata.org/wiki/Q334384","display_name":"Supervised learning","level":3,"score":0.3095000088214874},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.3093999922275543},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.3091000020503998},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.2581000030040741},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.2574999928474426},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.2538999915122986},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.25279998779296875}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.27185","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.27185","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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.27185","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.27185","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":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Text-to-motion":[0],"generation":[1],"has":[2],"advanced":[3],"with":[4,18,210],"diffusion-":[5],"and":[6,26,54,59,67,84,124,140,155,177,193,199],"flow-based":[7],"generative":[8],"models,":[9],"yet":[10],"supervised":[11],"pretraining":[12],"remains":[13],"insufficient":[14],"to":[15,180],"align":[16],"models":[17],"high-level":[19],"objectives":[20],"such":[21,42,50],"as":[22,43,51,136],"semantic":[23,104],"consistency,":[24],"realism,":[25],"human":[27],"preference.":[28],"Existing":[29],"post-training":[30],"methods":[31],"have":[32],"key":[33,138],"limitations:":[34],"they":[35],"(1)":[36],"target":[37],"a":[38,47,72,78,93,102,195],"specific":[39],"motion":[40],"representation,":[41,96],"joints,":[44],"(2)":[45],"optimize":[46],"particular":[48],"aspect,":[49],"text-motion":[52],"alignment,":[53],"may":[55],"compromise":[56],"other":[57],"factors;":[58],"(3)":[60],"incur":[61],"substantial":[62],"computational":[63],"overhead,":[64],"data":[65],"dependence,":[66],"coarse-grained":[68],"optimization.":[69],"We":[70],"present":[71],"reinforcement":[73],"fine-tuning":[74,88],"framework":[75],"that":[76],"comprises":[77],"heterogeneous-representation,":[79],"multi-dimensional":[80],"reward":[81,111],"model,":[82],"MotionReward,":[83],"an":[85],"efficient,":[86],"fine-grained":[87],"method,":[89],"EasyTune.":[90],"To":[91],"obtain":[92],"unified":[94],"semantics":[95,118],"MotionReward":[97],"maps":[98],"heterogeneous":[99],"motions":[100],"into":[101],"shared":[103],"space":[105],"anchored":[106],"by":[107,188],"text,":[108],"enabling":[109],"multidimensional":[110],"learning;":[112],"Self-refinement":[113],"Preference":[114],"Learning":[115],"further":[116],"enhances":[117],"without":[119],"additional":[120],"annotations.":[121],"For":[122],"efficient":[123],"effective":[125],"fine-tuning,":[126],"we":[127],"identify":[128],"the":[129,137,149,161],"recursive":[130],"gradient":[131],"dependence":[132],"across":[133],"denoising":[134],"steps":[135],"bottleneck,":[139],"propose":[141],"EasyTune,":[142],"which":[143],"optimizes":[144],"step-wise":[145],"rather":[146],"than":[147],"over":[148,183],"full":[150],"trajectory,":[151],"yielding":[152],"dense,":[153],"fine-grained,":[154],"memory-efficient":[156],"updates.":[157],"Extensive":[158],"experiments":[159],"validate":[160],"effectiveness":[162],"of":[163],"our":[164],"framework,":[165],"achieving":[166],"FID":[167,187,201],"0.132":[168],"at":[169],"22.10":[170],"GB":[171,182],"peak":[172],"memory":[173],"for":[174],"MLD":[175],"model":[176],"saving":[178],"up":[179],"15.22":[181],"DRaFT.":[184],"It":[185],"reduces":[186],"22.9%":[189],"on":[190,203],"joint-based":[191],"ACMDM,":[192],"achieves":[194],"12.6%":[196],"R-Precision":[197],"gain":[198],"23.3%":[200],"improvement":[202],"rotation-based":[204],"HY":[205],"Motion.":[206],"Our":[207],"project":[208],"page":[209],"code":[211],"is":[212],"publicly":[213],"available.":[214]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-02T00:00:00"}