{"id":"https://openalex.org/W7203705239","doi":"https://doi.org/10.48550/arxiv.2608.16316","title":"Deep Thought Alignment: Trajectory-Level Latent Distillation for Video Reasoning","display_name":"Deep Thought Alignment: Trajectory-Level Latent Distillation for Video Reasoning","publication_year":2026,"publication_date":"2026-08-17","ids":{"openalex":"https://openalex.org/W7203705239","doi":"https://doi.org/10.48550/arxiv.2608.16316"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.16316","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.16316","pdf_url":"https://arxiv.org/pdf/2608.16316","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-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/2608.16316","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100721606","display_name":"Ao Shen","orcid":"https://orcid.org/0000-0001-7196-2168","observed_orcids":["https://orcid.org/0000-0001-7196-2168"]},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]},{"id":"https://openalex.org/I24943067","display_name":"Fudan University","ror":"https://ror.org/013q1eq08","country_code":"CN","type":"education","lineage":["https://openalex.org/I24943067"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shen, Ao","raw_affiliation_strings":["Fudan University","Tencent Youtu Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Fudan University","institution_ids":["https://openalex.org/I24943067"]},{"raw_affiliation_string":"Tencent Youtu Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5146593270","display_name":"Yongheng Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Yongheng","raw_affiliation_strings":["Tencent Youtu Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tencent Youtu Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121797040","display_name":"Yinghui Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li, Yinghui","raw_affiliation_strings":["Tencent Youtu Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tencent Youtu Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029371371","display_name":"Manning Wang","orcid":"https://orcid.org/0000-0002-9255-3897","observed_orcids":["https://orcid.org/0000-0002-9255-3897"]},"institutions":[{"id":"https://openalex.org/I24943067","display_name":"Fudan University","ror":"https://ror.org/013q1eq08","country_code":"CN","type":"education","lineage":["https://openalex.org/I24943067"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Manning","raw_affiliation_strings":["Fudan University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Fudan University","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123875660","display_name":"Di Yin","orcid":"https://orcid.org/0009-0006-7246-3102","observed_orcids":["https://orcid.org/0009-0006-7246-3102"]},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yin, Di","raw_affiliation_strings":["Tencent Youtu Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tencent Youtu Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5140179585","display_name":"Xing Sun","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250653659","display_name":"Tencent (China)","ror":"https://ror.org/00hhjss72","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250653659"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sun, Xing","raw_affiliation_strings":["Tencent Youtu Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tencent Youtu Lab","institution_ids":["https://openalex.org/I2250653659"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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":0.9994999766349792,"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/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"}},{"id":"https://openalex.org/T10812","display_name":"Human Pose and Action Recognition","score":9.999999747378752e-05,"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/video-reasoning","display_name":"video reasoning","score":0.9300000071525574},{"id":"https://openalex.org/keywords/on-policy-distillation","display_name":"on-policy distillation","score":0.7710000276565552},{"id":"https://openalex.org/keywords/large-multimodal-models","display_name":"large multimodal models","score":0.847000002861023},{"id":"https://openalex.org/keywords/knowledge-distillation","display_name":"knowledge distillation","score":0.9229999780654907},{"id":"https://openalex.org/keywords/evidence-aggregation","display_name":"evidence aggregation","score":0.503000020980835}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7504000067710876},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6427000164985657},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.5435000061988831},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.5429999828338623},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.45019999146461487},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4131999909877777},{"id":"https://openalex.org/C2778496695","wikidata":"https://www.wikidata.org/wiki/Q254128","display_name":"Dilemma","level":2,"score":0.3901999890804291},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.34060001373291016},{"id":"https://openalex.org/C198082294","wikidata":"https://www.wikidata.org/wiki/Q3399648","display_name":"Position (finance)","level":2,"score":0.3246000111103058},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.31630000472068787},{"id":"https://openalex.org/C20162079","wikidata":"https://www.wikidata.org/wiki/Q1151406","display_name":"Case-based reasoning","level":2,"score":0.30979999899864197},{"id":"https://openalex.org/C89288958","wikidata":"https://www.wikidata.org/wiki/Q7301504","display_name":"Reasoning system","level":2,"score":0.2985999882221222},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.262800008058548},{"id":"https://openalex.org/C56461940","wikidata":"https://www.wikidata.org/wiki/Q970687","display_name":"Eye tracking","level":2,"score":0.2581999897956848},{"id":"https://openalex.org/C97364631","wikidata":"https://www.wikidata.org/wiki/Q484284","display_name":"Deductive reasoning","level":2,"score":0.25699999928474426},{"id":"https://openalex.org/C83725634","wikidata":"https://www.wikidata.org/wiki/Q7268699","display_name":"Qualitative reasoning","level":2,"score":0.25540000200271606}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.16316","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.16316","pdf_url":"https://arxiv.org/pdf/2608.16316","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.16316","doi":"https://doi.org/10.48550/arxiv.2608.16316","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.16316","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:2608.16316","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.16316","pdf_url":"https://arxiv.org/pdf/2608.16316","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-nc-sa","license_id":"https://openalex.org/licenses/cc-by-nc-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/W7203705239.pdf","grobid_xml":"https://content.openalex.org/works/W7203705239.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7215114519","https://openalex.org/W7160655277","https://openalex.org/W7165670554","https://openalex.org/W7083823545","https://openalex.org/W7167600618","https://openalex.org/W7162427912","https://openalex.org/W4416322399","https://openalex.org/W4417291863","https://openalex.org/W7204598462","https://openalex.org/W4415961147"],"abstract_inverted_index":{"Large":[0],"Multimodal":[1],"Models":[2],"(LMMs)":[3],"for":[4],"video":[5,55,147,187],"reasoning":[6,30,56,125,148],"have":[7],"long":[8,168],"been":[9],"hindered":[10],"by":[11,47],"the":[12,26,29,82,107,110,120,158],"high":[13],"computational":[14],"cost":[15],"of":[16,20,28,32,112],"processing":[17],"vast":[18],"amounts":[19],"visual":[21,122],"information.":[22],"This":[23],"dilemma":[24],"motivates":[25],"transfer":[27],"capabilities":[31],"large":[33],"models":[34],"to":[35,185],"smaller,":[36],"more":[37],"efficient":[38],"ones.":[39],"On-Policy":[40],"Distillation":[41],"(OPD)":[42],"offers":[43],"a":[44,130,181],"promising":[45],"solution":[46],"matching":[48],"output-token":[49],"distributions":[50],"along":[51],"student-generated":[52],"trajectories.":[53],"However,":[54],"often":[57],"depends":[58],"on":[59,106,145],"evidence":[60,123,174],"accumulated":[61,121],"across":[62],"multiple":[63],"frames.":[64],"In":[65],"this":[66,90],"context,":[67],"output-level":[68],"supervision":[69],"only":[70],"captures":[71],"information":[72],"expressed":[73],"through":[74],"token":[75],"predictions":[76],"and":[77,124],"does":[78],"not":[79],"directly":[80],"constrain":[81],"latent":[83,100],"representations":[84],"formed":[85],"during":[86],"reasoning.":[87,188],"To":[88],"address":[89],"limitation,":[91],"we":[92,128],"propose":[93],"Latent-OPD,":[94],"which":[95,134],"augments":[96],"OPD":[97],"with":[98,139,165],"trajectory-level":[99],"distillation.":[101],"Specifically,":[102],"our":[103],"method":[104],"focuses":[105],"position":[108],"at":[109],"end":[111],"each":[113],"trajectory,":[114],"where":[115],"hidden":[116],"states":[117],"effectively":[118],"summarize":[119],"context.":[126],"Furthermore,":[127],"introduce":[129],"progressive":[131],"teacher-lookahead":[132],"strategy,":[133],"aligns":[135],"middle-to-late":[136],"student":[137],"layers":[138],"increasingly":[140],"deeper":[141],"teacher":[142],"layers.":[143],"Experiments":[144],"six":[146],"benchmarks":[149],"show":[150],"that":[151],"Latent-OPD":[152,179],"consistently":[153],"outperforms":[154],"output-only":[155],"OPD.":[156],"Notably,":[157],"improvements":[159],"are":[160],"particularly":[161],"pronounced":[162],"in":[163],"scenarios":[164],"limited":[166],"frames,":[167],"videos,":[169],"or":[170],"tasks":[171],"requiring":[172],"complex":[173],"aggregation.":[175],"These":[176],"results":[177],"establish":[178],"as":[180],"highly":[182],"effective":[183],"approach":[184],"frame-efficient":[186]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-08-19T00:00:00"}