{"id":"https://openalex.org/W7225414705","doi":"https://doi.org/10.1145/3767308.3835704","title":"DraftMind: Efficient Video Understanding in Vision-Language Models via Chain of Draft","display_name":"DraftMind: Efficient Video Understanding in Vision-Language Models via Chain of Draft","publication_year":2026,"publication_date":"2026-11-06","ids":{"openalex":"https://openalex.org/W7225414705","doi":"https://doi.org/10.1145/3767308.3835704"},"language":"en","primary_location":{"id":"doi:10.1145/3767308.3835704","doi":"https://doi.org/10.1145/3767308.3835704","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3767308.3835704","pdf_url":null,"source":{"id":"https://openalex.org/S4363608757","display_name":"ACM International Conference on Multimedia (ACM MM)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 34th ACM International Conference on Multimedia","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.1145/3767308.3835704","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5103189768","display_name":"Zhipeng Wei","orcid":"https://orcid.org/0000-0002-5708-7157","observed_orcids":["https://orcid.org/0000-0002-5708-7157"]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"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/I4210165038","display_name":"University of Chinese Academy of Sciences","ror":"https://ror.org/05qbk4x57","country_code":"CN","type":"education","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhaoyang Wei","raw_affiliation_strings":["Tencent, Beijing, China","University of the Chinese Academy of Sciences, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-5708-7157","affiliations":[{"raw_affiliation_string":"Tencent, Beijing, China","institution_ids":["https://openalex.org/I2250653659"]},{"raw_affiliation_string":"University of the Chinese Academy of Sciences, Beijing, China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126342105","display_name":"Wenchao Ding","orcid":"https://orcid.org/0009-0005-6678-9344","observed_orcids":["https://orcid.org/0009-0005-6678-9344"]},"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":"Wenchao Ding","raw_affiliation_strings":["Tencent, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0005-6678-9344","affiliations":[{"raw_affiliation_string":"Tencent, Beijing, China","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090942417","display_name":"Chenhui Qiang","orcid":"https://orcid.org/0000-0002-5358-8659","observed_orcids":["https://orcid.org/0000-0002-5358-8659"]},"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":"Chenhui Qiang","raw_affiliation_strings":["Tencent, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-5358-8659","affiliations":[{"raw_affiliation_string":"Tencent, Beijing, China","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5161130344","display_name":"Sen Nie","orcid":"https://orcid.org/0009-0004-9015-320X","observed_orcids":["https://orcid.org/0009-0004-9015-320X"]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sen Nie","raw_affiliation_strings":["Chinese Academy of Sciences, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0004-9015-320X","affiliations":[{"raw_affiliation_string":"Chinese Academy of Sciences, Beijing, China","institution_ids":["https://openalex.org/I19820366"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120424829","display_name":"Shiyin Lin","orcid":"https://orcid.org/0009-0006-1644-4214","observed_orcids":["https://orcid.org/0009-0006-1644-4214"]},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Shiyin Lin","raw_affiliation_strings":["Google, San Francisco, USA"],"raw_orcid":"https://orcid.org/0009-0006-1644-4214","affiliations":[{"raw_affiliation_string":"Google, San Francisco, USA","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160992930","display_name":"Bowen Jiang","orcid":"https://orcid.org/0009-0008-5052-3842","observed_orcids":["https://orcid.org/0009-0008-5052-3842"]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210165038","display_name":"University of Chinese Academy of Sciences","ror":"https://ror.org/05qbk4x57","country_code":"CN","type":"education","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bowen Jiang","raw_affiliation_strings":["University of the Chinese Academy of Sciences, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0008-5052-3842","affiliations":[{"raw_affiliation_string":"University of the Chinese Academy of Sciences, Beijing, China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031281125","display_name":"Yanchao Hao","orcid":"https://orcid.org/0000-0002-7345-2425","observed_orcids":["https://orcid.org/0000-0002-7345-2425"]},"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":"Yanchao Hao","raw_affiliation_strings":["Tencent, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-7345-2425","affiliations":[{"raw_affiliation_string":"Tencent, Beijing, China","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5134508626","display_name":"Zheng Wei","orcid":"https://orcid.org/0009-0008-1684-436X","observed_orcids":["https://orcid.org/0009-0008-1684-436X"]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"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/I4210165038","display_name":"University of Chinese Academy of Sciences","ror":"https://ror.org/05qbk4x57","country_code":"CN","type":"education","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zheng Wei","raw_affiliation_strings":["Tencent, Beijing, China","University of the Chinese Academy of Sciences, Beijing, China"],"raw_orcid":"https://orcid.org/0009-0008-1684-436X","affiliations":[{"raw_affiliation_string":"Tencent, Beijing, China","institution_ids":["https://openalex.org/I2250653659"]},{"raw_affiliation_string":"University of the Chinese Academy of Sciences, Beijing, China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.85364699,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"11365","last_page":"11373"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":1.0,"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":1.0,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.0,"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/T10028","display_name":"Topic Modeling","score":0.0,"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/video-understanding","display_name":"video understanding","score":0.9539999961853027},{"id":"https://openalex.org/keywords/vision-language-models","display_name":"vision-language models","score":0.9860000014305115},{"id":"https://openalex.org/keywords/chain-of-thought","display_name":"chain-of-thought","score":0.703000009059906},{"id":"https://openalex.org/keywords/inference-efficiency","display_name":"inference efficiency","score":0.42899999022483826},{"id":"https://openalex.org/keywords/video-captioning","display_name":"video captioning","score":0.4569999873638153}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6119999885559082},{"id":"https://openalex.org/C199185054","wikidata":"https://www.wikidata.org/wiki/Q552299","display_name":"Chain (unit)","level":2,"score":0.4131999909877777},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.3319999873638153},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.31040000915527344},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.2734000086784363},{"id":"https://openalex.org/C108713360","wikidata":"https://www.wikidata.org/wiki/Q1824206","display_name":"Supply chain","level":2,"score":0.266400009393692},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.263700008392334},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2500999867916107},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.23899999260902405},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.23669999837875366}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3767308.3835704","doi":"https://doi.org/10.1145/3767308.3835704","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3767308.3835704","pdf_url":null,"source":{"id":"https://openalex.org/S4363608757","display_name":"ACM International Conference on Multimedia (ACM MM)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 34th ACM International Conference on Multimedia","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3767308.3835704","doi":"https://doi.org/10.1145/3767308.3835704","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3767308.3835704","pdf_url":null,"source":{"id":"https://openalex.org/S4363608757","display_name":"ACM International Conference on Multimedia (ACM MM)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 34th ACM International Conference on Multimedia","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":24,"referenced_works":["https://openalex.org/W4307934016","https://openalex.org/W4366850747","https://openalex.org/W4381586827","https://openalex.org/W4386185600","https://openalex.org/W4392172801","https://openalex.org/W4399655895","https://openalex.org/W4402670037","https://openalex.org/W4402727142","https://openalex.org/W4402753774","https://openalex.org/W4402753874","https://openalex.org/W4402754220","https://openalex.org/W4403853618","https://openalex.org/W4404356490","https://openalex.org/W4404784276","https://openalex.org/W4406755497","https://openalex.org/W4407794264","https://openalex.org/W4413144696","https://openalex.org/W4413146669","https://openalex.org/W4415080454","https://openalex.org/W4415796955","https://openalex.org/W7133193597","https://openalex.org/W7133198646","https://openalex.org/W7133209886","https://openalex.org/W7133224126"],"related_works":["https://openalex.org/W7225383164","https://openalex.org/W7225339504","https://openalex.org/W3176481196","https://openalex.org/W7196924254","https://openalex.org/W4415539763","https://openalex.org/W4412055689","https://openalex.org/W7140077871","https://openalex.org/W7225395907","https://openalex.org/W7225309073","https://openalex.org/W7224484692"],"abstract_inverted_index":{"Large":[0],"vision-language":[1],"models":[2],"(LVLMs)":[3],"such":[4],"as":[5],"Qwen-VL":[6],"series":[7],"have":[8],"demonstrated":[9],"strong":[10],"capabilities":[11],"in":[12,167],"visual":[13],"understanding":[14],"and":[15,34,66,76,101,121,141,143,169],"reasoning,":[16],"particularly":[17,87],"for":[18,50,89],"static":[19],"images.":[20],"However,":[21],"extending":[22],"them":[23],"to":[24,73,98,111,119,126],"video":[25,52,92,99,134,164],"reasoning":[26,53,79,100,135,157,165],"remains":[27],"computationally":[28],"expensive,":[29],"often":[30],"incurring":[31],"substantial":[32],"latency":[33,123],"token":[35,115],"usage":[36],"under":[37],"conventional":[38],"Chain-of-Thought":[39],"(CoT)":[40],"prompting.":[41],"In":[42],"this":[43,96],"paper,":[44],"we":[45],"present":[46],"a":[47],"unified":[48],"framework":[49],"efficient":[51],"that":[54,103,154],"combines":[55],"Chain":[56],"of":[57],"Draft":[58],"(CoD)":[59],"prompting":[60],"with":[61],"adaptive":[62],"sampling,":[63],"event":[64],"compression,":[65],"supervised":[67],"fine-tuning.":[68],"CoD":[69],"encourages":[70],"the":[71],"model":[72],"produce":[74],"concise":[75,155],"informative":[77],"intermediate":[78],"steps":[80],"rather":[81],"than":[82],"verbose":[83],"rationales,":[84],"making":[85],"it":[86],"suitable":[88],"temporally":[90],"complex":[91],"understanding.":[93],"We":[94,128],"adapt":[95],"paradigm":[97],"show":[102,153],"our":[104,130,151],"approach":[105],"achieves":[106],"comparable":[107],"or":[108],"superior":[109],"accuracy":[110],"CoT":[112],"while":[113],"reducing":[114],"consumption":[116],"by":[117,124],"up":[118,125],"78%":[120],"inference":[122],"65%.":[127],"evaluate":[129],"method":[131],"on":[132],"multiple":[133],"benchmarks,":[136],"including":[137],"MVBench,":[138],"EgoSchema,":[139],"PerceptionTest,":[140],"VideoMME,":[142],"demonstrate":[144],"consistent":[145],"gains":[146],"across":[147],"diverse":[148],"tasks.":[149],"Overall,":[150],"results":[152],"draft-style":[156],"offers":[158],"an":[159],"effective":[160],"path":[161],"toward":[162],"deploying":[163],"systems":[166],"resource-constrained":[168],"real-time":[170],"settings.":[171]},"counts_by_year":[],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-11-07T00:00:00"}