{"id":"https://openalex.org/W7128036410","doi":"https://doi.org/10.48550/arxiv.2602.04145","title":"Training Data Efficiency in Multimodal Process Reward Models","display_name":"Training Data Efficiency in Multimodal Process Reward Models","publication_year":2026,"publication_date":"2026-02-04","ids":{"openalex":"https://openalex.org/W7128036410","doi":"https://doi.org/10.48550/arxiv.2602.04145"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2602.04145","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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/A5125112691","display_name":"Jinyuan Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Li, Jinyuan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125124756","display_name":"Chengsong Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Chengsong","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092240623","display_name":"Langlin Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Langlin","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102676471","display_name":"Shaoyang Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Shaoyang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125239160","display_name":"Haolin Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Haolin","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125104807","display_name":"Wenxuan Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Wenxuan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5125140040","display_name":"Jiaxin Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Jiaxin","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5125112691"],"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.7074999809265137,"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":0.7074999809265137,"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/T10799","display_name":"Data Visualization and Analytics","score":0.0502999983727932,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.04780000075697899,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/reliability","display_name":"Reliability (semiconductor)","score":0.6549000144004822},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.6126000285148621},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.5787000060081482},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.5303999781608582},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.5248000025749207},{"id":"https://openalex.org/keywords/monte-carlo-method","display_name":"Monte Carlo method","score":0.3971000015735626}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7152000069618225},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.6549000144004822},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.6126000285148621},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5787000060081482},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5475000143051147},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.5303999781608582},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5270000100135803},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.5248000025749207},{"id":"https://openalex.org/C19499675","wikidata":"https://www.wikidata.org/wiki/Q232207","display_name":"Monte Carlo method","level":2,"score":0.3971000015735626},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3231000006198883},{"id":"https://openalex.org/C7545210","wikidata":"https://www.wikidata.org/wiki/Q838123","display_name":"Data redundancy","level":2,"score":0.29179999232292175},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.26510000228881836},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.26170000433921814}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2602.04145","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2602.04145","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.04145","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":"pmh:doi:10.48550/arxiv.2602.04145","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"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":{"Multimodal":[0],"Process":[1],"Reward":[2],"Models":[3],"(MPRMs)":[4],"are":[5],"central":[6],"to":[7],"step-level":[8],"supervision":[9],"for":[10,34],"visual":[11],"reasoning":[12],"in":[13],"MLLMs.":[14],"Training":[15],"MPRMs":[16],"typically":[17],"requires":[18],"large-scale":[19],"Monte":[20],"Carlo":[21],"(MC)-annotated":[22],"corpora,":[23],"incurring":[24,118],"substantial":[25,54],"training":[26,43,51,155],"cost.":[27,121],"This":[28],"paper":[29],"studies":[30],"the":[31,50,98,114,137,144,154],"data":[32],"efficiency":[33],"MPRM":[35,42],"training.":[36],"Our":[37],"preliminary":[38],"experiments":[39],"reveal":[40,69],"that":[41,70],"quickly":[44],"saturates":[45],"under":[46],"random":[47,159],"subsampling":[48,160],"of":[49,80,89,153],"data,":[52,156],"indicating":[53],"redundancy":[55],"within":[56],"existing":[57,110],"MC-annotated":[58],"corpora.":[59],"To":[60],"explain":[61],"this,":[62],"we":[63,96],"formalize":[64],"a":[65,162],"theoretical":[66],"framework":[67],"and":[68,83,106,126,134],"informative":[71],"gradient":[72],"updates":[73],"depend":[74],"on":[75,109,128],"two":[76,123],"factors:":[77],"label":[78,84],"mixtures":[79],"positive/negative":[81],"steps":[82],"reliability":[85,107],"(average":[86],"MC":[87,111],"scores":[88],"positive":[90],"steps).":[91],"Guided":[92],"by":[93,161],"these":[94],"insights,":[95],"propose":[97],"Balanced-Information":[99],"Score":[100],"(BIS),":[101],"which":[102],"prioritizes":[103],"both":[104],"mixture":[105],"based":[108],"signals":[112],"at":[113,140],"rollout":[115],"level,":[116],"without":[117],"any":[119],"additional":[120],"Across":[122],"backbones":[124],"(InternVL2.5-8B":[125],"Qwen2.5-VL-7B)":[127],"VisualProcessBench,":[129],"BIS-selected":[130],"subsets":[131],"consistently":[132],"match":[133],"even":[135],"surpass":[136],"full-data":[138,148],"performance":[139,149],"small":[141],"fractions.":[142],"Notably,":[143],"BIS":[145],"subset":[146],"reaches":[147],"using":[150],"only":[151],"10%":[152],"improving":[157],"over":[158],"relative":[163],"4.1%.":[164]},"counts_by_year":[],"updated_date":"2026-04-04T16:13:02.066488","created_date":"2026-02-07T00:00:00"}