{"id":"https://openalex.org/W7204598462","doi":"https://doi.org/10.48550/arxiv.2608.27065","title":"Video-OPSD: Exploiting Privileged Visual Evidence for On-Policy Self-Distillation in Video Large Language Models","display_name":"Video-OPSD: Exploiting Privileged Visual Evidence for On-Policy Self-Distillation in Video Large Language Models","publication_year":2026,"publication_date":"2026-08-27","ids":{"openalex":"https://openalex.org/W7204598462","doi":"https://doi.org/10.48550/arxiv.2608.27065"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.27065","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.27065","pdf_url":"https://arxiv.org/pdf/2608.27065","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.27065","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125487948","display_name":"Ziyue Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Wang, Ziyue","raw_affiliation_strings":["School of Electrical and Electronic Engineering , Nanyang Technological University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electrical and Electronic Engineering , Nanyang Technological University","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123668362","display_name":"Shiqi Huang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Huang, Shiqi","raw_affiliation_strings":["School of Electrical and Electronic Engineering , Nanyang Technological University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electrical and Electronic Engineering , Nanyang Technological University","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148524131","display_name":"Weiwen Xu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xu, Weiwen","raw_affiliation_strings":["The Chinese University of Hong Kong"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109639298","display_name":"Bihan Wen","orcid":"https://orcid.org/0000-0002-6874-6453","observed_orcids":["https://orcid.org/0000-0002-6874-6453"]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Wen, Bihan","raw_affiliation_strings":["School of Electrical and Electronic Engineering , Nanyang Technological University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electrical and Electronic Engineering , Nanyang Technological University","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5148546877","display_name":"Xudong Jiang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Jiang, Xudong","raw_affiliation_strings":["School of Electrical and Electronic Engineering , Nanyang Technological University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electrical and Electronic Engineering , Nanyang Technological University","institution_ids":["https://openalex.org/I172675005"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":5,"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":0.9994999766349792,"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":0.9994999766349792,"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/T10812","display_name":"Human Pose and Action Recognition","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"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":9.999999747378752e-05,"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/on-policy-self-distillation","display_name":"on-policy self-distillation","score":0.9959999918937683},{"id":"https://openalex.org/keywords/video-large-language-models","display_name":"video large language models","score":0.9890000224113464},{"id":"https://openalex.org/keywords/video-reasoning","display_name":"video reasoning","score":0.75},{"id":"https://openalex.org/keywords/video-understanding","display_name":"video understanding","score":0.7670000195503235}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.770799994468689},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.6741999983787537},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.6280999779701233},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.5976999998092651},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4797999858856201},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.44449999928474426},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.42640000581741333},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.42399999499320984},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.39660000801086426},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.39559999108314514},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.37400001287460327},{"id":"https://openalex.org/C2780878386","wikidata":"https://www.wikidata.org/wiki/Q1659648","display_name":"Visual language","level":2,"score":0.3269999921321869},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2896000146865845},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.28690001368522644},{"id":"https://openalex.org/C49774154","wikidata":"https://www.wikidata.org/wiki/Q131765","display_name":"Multimedia","level":1,"score":0.28189998865127563},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.27399998903274536},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.26930001378059387},{"id":"https://openalex.org/C2983448237","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Language understanding","level":2,"score":0.2556999921798706},{"id":"https://openalex.org/C2987052865","wikidata":"https://www.wikidata.org/wiki/Q11033","display_name":"Visual media","level":2,"score":0.25450000166893005},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.2524000108242035},{"id":"https://openalex.org/C2780760462","wikidata":"https://www.wikidata.org/wiki/Q7927952","display_name":"Video modeling","level":4,"score":0.25209999084472656}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.27065","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.27065","pdf_url":"https://arxiv.org/pdf/2608.27065","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.27065","doi":"https://doi.org/10.48550/arxiv.2608.27065","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.27065","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:2608.27065","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.27065","pdf_url":"https://arxiv.org/pdf/2608.27065","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204598462.pdf","grobid_xml":"https://content.openalex.org/works/W7204598462.grobid-xml"},"referenced_works_count":31,"referenced_works":["https://openalex.org/W3206166878","https://openalex.org/W4391631327","https://openalex.org/W4402670320","https://openalex.org/W4402702917","https://openalex.org/W4402727520","https://openalex.org/W4403799688","https://openalex.org/W4405956418","https://openalex.org/W4406796387","https://openalex.org/W4407248386","https://openalex.org/W4407794264","https://openalex.org/W4413146669","https://openalex.org/W4414281281","https://openalex.org/W4414580487","https://openalex.org/W4415037432","https://openalex.org/W4415063999","https://openalex.org/W4416184586","https://openalex.org/W4416780398","https://openalex.org/W4417078362","https://openalex.org/W7125861034","https://openalex.org/W7125915072","https://openalex.org/W7125929755","https://openalex.org/W7126034091","https://openalex.org/W7127613426","https://openalex.org/W7133227718","https://openalex.org/W7150828934","https://openalex.org/W7154505894","https://openalex.org/W7154615764","https://openalex.org/W7155051642","https://openalex.org/W7160655277","https://openalex.org/W7162084357","https://openalex.org/W7169534000"],"related_works":["https://openalex.org/W7160655277","https://openalex.org/W7204449050","https://openalex.org/W7220742813","https://openalex.org/W7215114519","https://openalex.org/W7203705239","https://openalex.org/W7221529141","https://openalex.org/W7212149089","https://openalex.org/W7135206943","https://openalex.org/W7162084357","https://openalex.org/W4416361295"],"abstract_inverted_index":{"On-policy":[0],"self-distillation":[1],"(OPSD)":[2],"has":[3],"recently":[4],"emerged":[5],"as":[6],"an":[7,103,207],"effective":[8,208],"post-training":[9,211],"paradigm":[10],"that":[11,106,184],"improves":[12,187],"policy":[13],"optimization":[14],"through":[15],"dense":[16],"token-level":[17,159],"supervision":[18,69],"from":[19],"a":[20,64,83,94],"privileged":[21,40,68,108,168],"self-teacher.":[22],"Despite":[23],"its":[24],"promise,":[25],"OPSD":[26,104,190],"remains":[27],"largely":[28],"underexplored":[29],"for":[30,55,111,213],"Video":[31,60],"Large":[32],"Language":[33],"Models":[34],"(Video-LLMs).":[35],"Existing":[36],"methods":[37],"typically":[38],"construct":[39],"teachers":[41],"by":[42],"augmenting":[43],"their":[44],"context":[45],"with":[46],"additional":[47],"information":[48],"while":[49,130,200],"keeping":[50],"the":[51,71,89,123,131,137,145],"primary":[52,72],"input":[53,73,143],"unchanged":[54],"both":[56,112],"teacher":[57,124,146],"and":[58,81,115,180,194,209],"student.":[59],"reasoning,":[61],"however,":[62],"offers":[63],"distinct":[65],"source":[66],"of":[67,86],"within":[70],"itself:":[74],"long":[75],"videos":[76],"contain":[77],"substantial":[78],"temporal":[79],"redundancy,":[80],"only":[82],"small":[84],"subset":[85],"frames":[87,129],"provides":[88],"evidence":[90,110,128],"necessary":[91],"to":[92,134,147,162,198],"answer":[93],"question.":[95],"Building":[96],"on":[97,126,167],"this":[98],"observation,":[99],"we":[100],"present":[101],"$\\textbf{Video-OPSD}$,":[102],"framework":[105],"exploits":[107],"visual":[109,142,169],"self-teacher":[113],"construction":[114],"knowledge":[116],"transfer.":[117],"First,":[118],"our":[119,153],"Evidence-Grounded":[120],"Self-Teacher":[121],"conditions":[122],"exclusively":[125],"annotated":[127],"student":[132],"continues":[133],"reason":[135],"over":[136],"complete":[138],"video.":[139],"This":[140],"focused":[141],"enables":[144],"provide":[148],"more":[149],"informative":[150],"supervision.":[151],"Second,":[152],"Evidence-Guided":[154],"Token":[155],"Optimization":[156],"adaptively":[157],"weights":[158],"distillation":[160],"according":[161],"each":[163],"reasoning":[164,181],"token's":[165],"reliance":[166],"evidence,":[170],"thereby":[171],"emphasizing":[172],"perceptually":[173],"grounded":[174],"reasoning.":[175],"Experiments":[176],"across":[177,191],"video":[178],"understanding":[179],"benchmarks":[182],"show":[183],"$\\textbf{Video-OPSD}$":[185],"consistently":[186],"upon":[188],"Standard":[189],"multiple":[192],"backbones":[193],"achieves":[195],"performance":[196],"comparable":[197],"GRPO":[199],"requiring":[201],"substantially":[202],"less":[203],"training":[204],"time,":[205],"establishing":[206],"efficient":[210],"approach":[212],"Video-LLMs.":[214]},"counts_by_year":[{"year":2026,"cited_by_count":5}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-08-29T00:00:00"}