{"id":"https://openalex.org/W7135206943","doi":"https://doi.org/10.48550/arxiv.2603.11423","title":"Beyond Single-Sample: Reliable Multi-Sample Distillation for Video Understanding","display_name":"Beyond Single-Sample: Reliable Multi-Sample Distillation for Video Understanding","publication_year":2026,"publication_date":"2026-03-12","ids":{"openalex":"https://openalex.org/W7135206943","doi":"https://doi.org/10.48550/arxiv.2603.11423"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.11423","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11423","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.2603.11423","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128973071","display_name":"Songlin Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Songlin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129007430","display_name":"Xin Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Xin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128974897","display_name":"Zechao Guan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guan, Zechao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031383306","display_name":"Peipeng Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Peipeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128931740","display_name":"Jian Yao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yao, Jian","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9139999747276306,"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":0.9139999747276306,"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":0.050999999046325684,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.00559999980032444,"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/distillation","display_name":"Distillation","score":0.8177000284194946},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.5914999842643738},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.5388000011444092},{"id":"https://openalex.org/keywords/adversarial-system","display_name":"Adversarial system","score":0.5382000207901001},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.5291000008583069},{"id":"https://openalex.org/keywords/sample","display_name":"Sample (material)","score":0.5235999822616577},{"id":"https://openalex.org/keywords/variance","display_name":"Variance (accounting)","score":0.4878000020980835}],"concepts":[{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.8177000284194946},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7268999814987183},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.5914999842643738},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.5388000011444092},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.5382000207901001},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5310999751091003},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.5291000008583069},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.5235999822616577},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.4878000020980835},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4641000032424927},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.4551999866962433},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4359000027179718},{"id":"https://openalex.org/C154030694","wikidata":"https://www.wikidata.org/wiki/Q1436074","display_name":"Fractionating column","level":3,"score":0.29159998893737793},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.28290000557899475},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.27559998631477356},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2597000002861023},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.258899986743927},{"id":"https://openalex.org/C13944312","wikidata":"https://www.wikidata.org/wiki/Q7512748","display_name":"Signal-to-noise ratio (imaging)","level":2,"score":0.2524999976158142},{"id":"https://openalex.org/C150921843","wikidata":"https://www.wikidata.org/wiki/Q1170431","display_name":"Resampling","level":2,"score":0.2513999938964844}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.11423","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11423","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.2603.11423","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11423","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":[{"id":"https://metadata.un.org/sdg/4","score":0.7738659381866455,"display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Traditional":[0],"black-box":[1],"distillation":[2,51,87,113],"for":[3],"Large":[4],"Vision-Language":[5],"Models":[6],"(LVLMs)":[7],"typically":[8],"relies":[9],"on":[10,56,148],"a":[11,41,57,64,140],"single":[12,58,111],"teacher":[13,46,59,66,93],"response":[14],"per":[15],"input,":[16],"which":[17,128],"often":[18],"yields":[19],"high-variance":[20],"responses":[21],"and":[22,76,153],"format":[23],"inconsistencies":[24],"in":[25],"multimodal":[26],"or":[27],"temporal":[28],"scenarios.":[29],"To":[30],"mitigate":[31],"this":[32],"unreliable":[33],"supervision,":[34],"we":[35],"propose":[36],"R-MSD":[37,108],"(Reliable":[38],"Multi-Sample":[39],"Distillation),":[40],"framework":[42],"that":[43,107],"explicitly":[44],"models":[45],"sampling":[47],"variance":[48],"to":[49,68,73],"enhance":[50],"stability.":[52],"Rather":[53],"than":[54],"relying":[55],"response,":[60],"our":[61,89,134,144],"approach":[62,90,145],"leverages":[63],"task-adaptive":[65],"pool":[67],"provide":[69],"robust":[70],"supervision":[71],"tailored":[72],"both":[74],"closed-ended":[75],"open-ended":[77],"reasoning.":[78],"By":[79],"integrating":[80],"quality-aware":[81],"signal":[82],"matching":[83],"with":[84],"an":[85,118],"adversarial":[86],"objective,":[88],"effectively":[91],"filters":[92],"noise":[94],"while":[95,133],"maximizing":[96],"knowledge":[97],"transfer.":[98],"Extensive":[99],"evaluations":[100],"across":[101],"comprehensive":[102],"video":[103],"understanding":[104],"benchmarks":[105],"demonstrate":[106],"consistently":[109],"outperforms":[110],"sample":[112],"methods.":[114],"We":[115],"additionally":[116],"include":[117],"original":[119],"SFT+RL":[120],"4B":[121,141],"baseline":[122],"under":[123],"the":[124],"same":[125],"training":[126],"budget,":[127],"shows":[129],"only":[130],"marginal":[131],"gains,":[132],"method":[135],"achieves":[136],"significant":[137],"improvements.":[138],"With":[139],"student":[142],"model,":[143],"delivers":[146],"gains":[147],"VideoMME":[149],"(+1.5%),":[150],"Video-MMMU":[151],"(+3.2%),":[152],"MathVerse":[154],"(+3.6%).":[155]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-14T00:00:00"}