{"id":"https://openalex.org/W6948439803","doi":"https://doi.org/10.48550/arxiv.2504.07615","title":"VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model","display_name":"VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model","publication_year":2025,"publication_date":"2025-04-10","ids":{"openalex":"https://openalex.org/W6948439803","doi":"https://doi.org/10.48550/arxiv.2504.07615"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2504.07615","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.07615","pdf_url":"https://arxiv.org/pdf/2504.07615","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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2504.07615","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5058869386","display_name":"Haozhan Shen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Haozhan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Liu, Peng","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Peng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Li, Jingcheng","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Jingcheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5083090126","display_name":"Chunxin Fang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fang, Chunxin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Ma, Yibo","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Yibo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5053066840","display_name":"Jiajia Liao","orcid":"https://orcid.org/0000-0002-8562-2743","observed_orcids":["https://orcid.org/0000-0002-8562-2743"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liao, Jiajia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Shen, Qiaoli","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Qiaoli","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013896311","display_name":"Zilun Zhang","orcid":"https://orcid.org/0009-0008-5961-5970","observed_orcids":["https://orcid.org/0009-0008-5961-5970"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Zilun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zhao, Kangjia","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Kangjia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zhang, Qianqian","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Qianqian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Xu, Ruochen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Ruochen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Zhao, Tiancheng","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Tiancheng","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":true,"cited_by_count":4,"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":true,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9984999895095825,"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":0.9984999895095825,"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":0.0003000000142492354,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.00019999999494757503,"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/vision-language-models","display_name":"vision-language models","score":0.8209999799728394},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"reinforcement learning","score":0.7730000019073486},{"id":"https://openalex.org/keywords/deepseek-r1","display_name":"DeepSeek-R1","score":0.8349999785423279},{"id":"https://openalex.org/keywords/visual-reasoning","display_name":"visual reasoning","score":0.8939999938011169},{"id":"https://openalex.org/keywords/visual-understanding","display_name":"visual understanding","score":0.5680000185966492},{"id":"https://openalex.org/keywords/reward-hacking","display_name":"reward hacking","score":0.7639999985694885},{"id":"https://openalex.org/keywords/object-detection","display_name":"object detection","score":0.9810000061988831},{"id":"https://openalex.org/keywords/generalization","display_name":"generalization","score":0.7080000042915344},{"id":"https://openalex.org/keywords/training-data-quality","display_name":"training data quality","score":0.8379999995231628},{"id":"https://openalex.org/keywords/model-scaling","display_name":"model scaling","score":0.6449999809265137}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.8144000172615051},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7530999779701233},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6247000098228455},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5985000133514404},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.42660000920295715},{"id":"https://openalex.org/C189950617","wikidata":"https://www.wikidata.org/wiki/Q937228","display_name":"Property (philosophy)","level":2,"score":0.4169999957084656},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.4138999879360199},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.41280001401901245},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.3874000012874603},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.34209999442100525},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.3294000029563904},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.31380000710487366},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.3124000132083893},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.2955000102519989},{"id":"https://openalex.org/C64876066","wikidata":"https://www.wikidata.org/wiki/Q5141226","display_name":"Cognitive neuroscience of visual object recognition","level":3,"score":0.289000004529953},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.2793000042438507},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C2778029271","wikidata":"https://www.wikidata.org/wiki/Q5421931","display_name":"Extension (predicate logic)","level":2,"score":0.26809999346733093},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.26660001277923584},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.26170000433921814},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.26100000739097595}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2504.07615","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.07615","pdf_url":"https://arxiv.org/pdf/2504.07615","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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2504.07615","doi":"https://doi.org/10.48550/arxiv.2504.07615","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2504.07615","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2504.07615","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.07615","pdf_url":"https://arxiv.org/pdf/2504.07615","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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W6948439803.pdf","grobid_xml":"https://content.openalex.org/works/W6948439803.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7160284950","https://openalex.org/W4415975221","https://openalex.org/W4415317420","https://openalex.org/W7078146995","https://openalex.org/W4415311095","https://openalex.org/W7137815919","https://openalex.org/W4416238817","https://openalex.org/W4415953283","https://openalex.org/W7156927051","https://openalex.org/W6979280327"],"abstract_inverted_index":{"Recently":[0],"DeepSeek":[1],"R1":[2,29],"has":[3],"shown":[4],"that":[5,57,143,171],"reinforcement":[6,92,221],"learning":[7,93,222],"(RL)":[8],"can":[9],"substantially":[10],"improve":[11],"the":[12,51,88,132,144,179,187,190,194,201,217,224,242],"reasoning":[13,103],"capabilities":[14,225],"of":[15,28,61,90,134,175,181,189,196,204,219,226],"Large":[16],"Language":[17],"Models":[18,96],"(LLMs)":[19],"through":[20],"a":[21,58,111,173],"simple":[22],"yet":[23],"effective":[24],"design.":[25],"The":[26],"core":[27],"lies":[30],"in":[31,162,184,241],"its":[32],"rule-based":[33,79],"reward":[34,48,80,182],"formulation,":[35],"which":[36],"leverages":[37],"tasks":[38,64,155],"with":[39,68,78],"deterministic":[40],"ground-truth":[41,70],"answers":[42],"to":[43,94,99,115,137,215],"enable":[44],"precise":[45],"and":[46,200,229,234,248],"stable":[47],"computation.":[49],"In":[50],"visual":[52,62,102,138,153],"domain,":[53],"we":[54,86,108,129,166,213,230],"similarly":[55],"observe":[56],"wide":[59],"range":[60],"understanding":[63,154,218],"are":[65,250],"inherently":[66],"equipped":[67],"well-defined":[69],"annotations.":[71],"This":[72],"property":[73],"makes":[74],"them":[75],"naturally":[76],"compatible":[77],"mechanisms.":[81],"Motivated":[82],"by":[83],"this":[84,106,127],"observation,":[85],"investigate":[87],"extension":[89],"R1-style":[91],"Vision-Language":[95],"(VLMs),":[97],"aiming":[98],"enhance":[100],"their":[101],"capabilities.":[104],"To":[105],"end,":[107],"develop":[109],"VLM-R1,":[110],"dedicated":[112],"framework":[113],"designed":[114],"harness":[116],"RL":[117,136,205,244],"for":[118],"improving":[119],"VLMs'":[120],"performance":[121,151],"on":[122,152],"general":[123],"vision-language":[124,227,243],"tasks.":[125],"Using":[126],"framework,":[128],"further":[130],"explore":[131],"feasibility":[133],"applying":[135],"domain.":[139],"Experimental":[140],"results":[141],"indicate":[142],"RL-based":[145],"model":[146,208,249],"not":[147],"only":[148],"delivers":[149],"competitive":[150],"but":[156],"also":[157],"surpasses":[158],"Supervised":[159],"Fine-Tuning":[160],"(SFT)":[161],"generalization":[163],"ability.":[164],"Furthermore,":[165],"conduct":[167],"comprehensive":[168],"ablation":[169],"studies":[170],"uncover":[172],"series":[174],"noteworthy":[176],"insights,":[177],"including":[178],"presence":[180],"hacking":[183],"object":[185],"detection,":[186],"emergence":[188],"\"OD":[191],"aha":[192],"moment\",":[193],"impact":[195],"training":[197],"data":[198],"quality,":[199],"scaling":[202],"behavior":[203],"across":[206],"different":[207],"sizes.":[209],"Through":[210],"these":[211],"analyses,":[212],"aim":[214],"deepen":[216],"how":[220],"enhances":[223],"models,":[228],"hope":[231],"our":[232],"findings":[233],"open-source":[235],"contributions":[236],"will":[237],"support":[238],"continued":[239],"progress":[240],"community.":[245],"Our":[246],"code":[247],"available":[251],"at":[252],"https://github.com/om-ai-lab/VLM-R1":[253]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":1}],"updated_date":"2026-10-14T16:39:01.820549","created_date":"2025-10-10T00:00:00"}