{"id":"https://openalex.org/W7133342360","doi":"https://doi.org/10.48550/arxiv.2603.01907","title":"Efficient RLVR Training via Weighted Mutual Information Data Selection","display_name":"Efficient RLVR Training via Weighted Mutual Information Data Selection","publication_year":2026,"publication_date":"2026-03-02","ids":{"openalex":"https://openalex.org/W7133342360","doi":"https://doi.org/10.48550/arxiv.2603.01907"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.01907","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.01907","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":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.2603.01907","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5127948830","display_name":"Xinyu Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Xinyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113375392","display_name":"Boyu Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Boyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128029996","display_name":"Haotian Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Haotian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128031258","display_name":"Huiming Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Huiming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5127909908","display_name":"Zhijiang Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Zhijiang","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/T10028","display_name":"Topic Modeling","score":0.3531000018119812,"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"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.3531000018119812,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.23319999873638153,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.05299999937415123,"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/reinforcement-learning","display_name":"Reinforcement learning","score":0.6093999743461609},{"id":"https://openalex.org/keywords/mutual-information","display_name":"Mutual information","score":0.5543000102043152},{"id":"https://openalex.org/keywords/equating","display_name":"Equating","score":0.5457000136375427},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.5446000099182129},{"id":"https://openalex.org/keywords/uncertainty-reduction-theory","display_name":"Uncertainty reduction theory","score":0.49309998750686646},{"id":"https://openalex.org/keywords/bayesian-probability","display_name":"Bayesian probability","score":0.4650999903678894},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.39660000801086426},{"id":"https://openalex.org/keywords/variation","display_name":"Variation (astronomy)","score":0.39660000801086426},{"id":"https://openalex.org/keywords/verifiable-secret-sharing","display_name":"Verifiable secret sharing","score":0.37790000438690186}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7001000046730042},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6401000022888184},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.6324999928474426},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6093999743461609},{"id":"https://openalex.org/C152139883","wikidata":"https://www.wikidata.org/wiki/Q252973","display_name":"Mutual information","level":2,"score":0.5543000102043152},{"id":"https://openalex.org/C106347477","wikidata":"https://www.wikidata.org/wiki/Q5384228","display_name":"Equating","level":3,"score":0.5457000136375427},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.5446000099182129},{"id":"https://openalex.org/C94361409","wikidata":"https://www.wikidata.org/wiki/Q7882500","display_name":"Uncertainty reduction theory","level":2,"score":0.49309998750686646},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.4650999903678894},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.39660000801086426},{"id":"https://openalex.org/C2778334786","wikidata":"https://www.wikidata.org/wiki/Q1586270","display_name":"Variation (astronomy)","level":2,"score":0.39660000801086426},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.37790000438690186},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.3763999938964844},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.37290000915527344},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3544999957084656},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.31779998540878296},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.2906999886035919},{"id":"https://openalex.org/C113336015","wikidata":"https://www.wikidata.org/wiki/Q574010","display_name":"Complete information","level":2,"score":0.2858999967575073},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.2782000005245209},{"id":"https://openalex.org/C3020402766","wikidata":"https://www.wikidata.org/wiki/Q104376712","display_name":"Prior information","level":2,"score":0.274399995803833},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2743000090122223},{"id":"https://openalex.org/C38764148","wikidata":"https://www.wikidata.org/wiki/Q17098245","display_name":"Interaction information","level":2,"score":0.2646999955177307},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.26159998774528503},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.26010000705718994},{"id":"https://openalex.org/C24756922","wikidata":"https://www.wikidata.org/wiki/Q1757694","display_name":"Data quality","level":3,"score":0.25459998846054077},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.2535000145435333}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.01907","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.01907","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":"doi:10.48550/arxiv.2603.01907","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.01907","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":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":{"Reinforcement":[0],"learning":[1,134],"(RL)":[2],"plays":[3],"a":[4,69,97,108,153],"central":[5],"role":[6],"in":[7,68,132],"improving":[8],"the":[9,114],"reasoning":[10],"and":[11,48,93,125,148,167],"alignment":[12],"of":[13,100,117],"large":[14],"language":[15],"models,":[16],"yet":[17],"its":[18],"efficiency":[19],"critically":[20],"depends":[21],"on":[22,34,113,157,164],"how":[23],"training":[24,150],"data":[25,61,76],"are":[26],"selected.":[27],"Existing":[28],"online":[29],"selection":[30],"strategies":[31],"predominantly":[32],"rely":[33],"difficulty-based":[35],"heuristics,":[36],"favouring":[37],"datapoints":[38],"with":[39,46,78,135,172],"intermediate":[40],"success":[41,81,119],"rates,":[42,82],"implicitly":[43],"equating":[44],"difficulty":[45],"informativeness":[47],"neglecting":[49],"epistemic":[50],"uncertainty":[51,87],"arising":[52],"from":[53],"limited":[54],"evidence.":[55],"We":[56],"introduce":[57],"InSight,":[58],"an":[59],"INformation-guided":[60],"SamplInG":[62],"metHod":[63],"for":[64],"RL":[65],"Training,":[66],"grounded":[67],"weighted":[70],"mutual":[71],"information":[72],"objective.":[73],"By":[74],"modeling":[75],"outcomes":[77],"Bayesian":[79],"latent":[80],"we":[83],"show":[84],"that":[85,142],"expected":[86],"reduction":[88],"decomposes":[89],"into":[90],"complementary":[91],"difficulty-":[92],"evidence-dependent":[94],"components,":[95],"revealing":[96],"fundamental":[98],"limitation":[99],"difficulty-only":[101],"selection.":[102],"Leveraging":[103],"this":[104],"observation,":[105],"InSight":[106,143],"constructs":[107],"stable":[109],"acquisition":[110],"score":[111],"based":[112],"mean":[115],"belief":[116],"datapoints'":[118],"rather":[120],"than":[121],"noisy":[122],"sampled":[123],"outcomes,":[124],"naturally":[126],"extends":[127],"to":[128,169],"multi-rollout":[129],"settings":[130],"common":[131],"reinforcement":[133],"verifiable":[136],"rewards":[137],"(RLVR).":[138],"Extensive":[139],"experiments":[140],"demonstrate":[141],"consistently":[144],"achieves":[145],"state-of-the-art":[146],"performance":[147],"improves":[149],"efficiency,":[151],"including":[152],"+1.41":[154],"average":[155],"gain":[156],"Planning":[158],"&":[159],"Mathmatics":[160],"benchmarks,":[161],"+1.01":[162],"improvement":[163],"general":[165],"reasoning,":[166],"up":[168],"~2.2x":[170],"acceleration,":[171],"negligible":[173],"additional":[174],"computational":[175],"overhead.":[176]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-04T00:00:00"}