{"id":"https://openalex.org/W7138090585","doi":"https://doi.org/10.1609/aaai.v40i15.38234","title":"D\u00b2Pruner: Debiased Importance and Structural Diversity for MLLM Token Pruning","display_name":"D\u00b2Pruner: Debiased Importance and Structural Diversity for MLLM Token Pruning","publication_year":2026,"publication_date":"2026-03-14","ids":{"openalex":"https://openalex.org/W7138090585","doi":"https://doi.org/10.1609/aaai.v40i15.38234"},"language":null,"primary_location":{"id":"doi:10.1609/aaai.v40i15.38234","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i15.38234","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://doi.org/10.1609/aaai.v40i15.38234","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129651924","display_name":"Evelyn Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]},{"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":true,"raw_author_name":"Evelyn Zhang","raw_affiliation_strings":["Shanghai Jiaotong University\nTencent Youtu Lab"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiaotong University\nTencent Youtu Lab","institution_ids":["https://openalex.org/I2250653659","https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005929117","display_name":"Fufu Yu","orcid":"https://orcid.org/0000-0002-4062-696X"},"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":"Fufu Yu","raw_affiliation_strings":["Tencent YouTu Lab"],"affiliations":[{"raw_affiliation_string":"Tencent YouTu Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038991203","display_name":"Aoqi Wu","orcid":null},"institutions":[{"id":"https://openalex.org/I116953780","display_name":"Tongji University","ror":"https://ror.org/03rc6as71","country_code":"CN","type":"education","lineage":["https://openalex.org/I116953780"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Aoqi Wu","raw_affiliation_strings":["Tongji University"],"affiliations":[{"raw_affiliation_string":"Tongji University","institution_ids":["https://openalex.org/I116953780"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102635197","display_name":"Zichen Wen","orcid":"https://orcid.org/0009-0002-6157-5898"},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zichen Wen","raw_affiliation_strings":["Shanghai Jiaotong University"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiaotong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129704294","display_name":"Ke Yan","orcid":null},"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":"Ke Yan","raw_affiliation_strings":["Tencent YouTu Lab"],"affiliations":[{"raw_affiliation_string":"Tencent YouTu Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129749619","display_name":"Shouhong Ding","orcid":null},"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":"Shouhong Ding","raw_affiliation_strings":["Tencent YouTu Lab"],"affiliations":[{"raw_affiliation_string":"Tencent YouTu Lab","institution_ids":["https://openalex.org/I2250653659"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129684848","display_name":"Biqing Qi","orcid":null},"institutions":[{"id":"https://openalex.org/I4210122302","display_name":"ShangHai JiAi Genetics & IVF Institute","ror":"https://ror.org/02rgbry52","country_code":"CN","type":"healthcare","lineage":["https://openalex.org/I4210122302"]},{"id":"https://openalex.org/I4391012619","display_name":"Shanghai Artificial Intelligence Laboratory","ror":"https://ror.org/03wkvpx79","country_code":null,"type":"facility","lineage":["https://openalex.org/I4391012619"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Biqing Qi","raw_affiliation_strings":["Shanghai AI Laboratory"],"affiliations":[{"raw_affiliation_string":"Shanghai AI Laboratory","institution_ids":["https://openalex.org/I4210122302","https://openalex.org/I4391012619"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129642313","display_name":"Linfeng Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Linfeng Zhang","raw_affiliation_strings":["Shanghai Jiaotong University"],"affiliations":[{"raw_affiliation_string":"Shanghai Jiaotong University","institution_ids":["https://openalex.org/I183067930"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5129651924"],"corresponding_institution_ids":["https://openalex.org/I183067930","https://openalex.org/I2250653659"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.39552239,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"40","issue":"15","first_page":"12412","last_page":"12420"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.8675000071525574,"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":0.8675000071525574,"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":0.029400000348687172,"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":0.013799999840557575,"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/security-token","display_name":"Security token","score":0.8361999988555908},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.7121999859809875},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.6008999943733215},{"id":"https://openalex.org/keywords/path","display_name":"Path (computing)","score":0.5424000024795532},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.4997999966144562},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.492000013589859},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.4438000023365021},{"id":"https://openalex.org/keywords/token-passing","display_name":"Token passing","score":0.41769999265670776}],"concepts":[{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.8361999988555908},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.73580002784729},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.7121999859809875},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.6008999943733215},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5615000128746033},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.5424000024795532},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4997999966144562},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.492000013589859},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.4438000023365021},{"id":"https://openalex.org/C115067241","wikidata":"https://www.wikidata.org/wiki/Q1639854","display_name":"Token passing","level":3,"score":0.41769999265670776},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.35580000281333923},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3296000063419342},{"id":"https://openalex.org/C2775955345","wikidata":"https://www.wikidata.org/wiki/Q7449071","display_name":"Semantic mapping","level":2,"score":0.29789999127388},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.2849999964237213},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.28299999237060547},{"id":"https://openalex.org/C2779010991","wikidata":"https://www.wikidata.org/wiki/Q2720909","display_name":"Artifact (error)","level":2,"score":0.26759999990463257},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2651999890804291},{"id":"https://openalex.org/C148483581","wikidata":"https://www.wikidata.org/wiki/Q446488","display_name":"Feature selection","level":2,"score":0.26460000872612},{"id":"https://openalex.org/C130318100","wikidata":"https://www.wikidata.org/wiki/Q2268914","display_name":"Semantic similarity","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2515999972820282}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1609/aaai.v40i15.38234","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i15.38234","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1609/aaai.v40i15.38234","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i15.38234","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"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":{"Processing":[0],"long":[1],"visual":[2],"token":[3,17,162],"sequences":[4],"poses":[5],"a":[6,20,57,89,101,109,121,128,142],"significant":[7],"computational":[8],"burden":[9],"on":[10,36,120,134,141],"Multimodal":[11],"Large":[12],"Language":[13],"Models":[14],"(MLLMs).":[15],"While":[16],"pruning":[18,103],"offers":[19],"path":[21],"to":[22,44,174],"acceleration,":[23],"we":[24,86],"find":[25],"that":[26,65,91,168,182],"current":[27],"methods,":[28],"while":[29,70,163],"adequate":[30],"for":[31],"general":[32],"understanding,":[33],"catastrophically":[34],"fail":[35],"fine-grained":[37],"localization":[38],"tasks.":[39],"We":[40],"attribute":[41],"this":[42],"failure":[43],"the":[45,49,77,113,135,157,169],"inherent":[46,62],"flaws":[47],"of":[48,112],"two":[50],"prevailing":[51],"strategies:":[52],"importance-based":[53],"methods":[54,72],"suffer":[55],"from":[56,67],"strong":[58],"positional":[59],"bias,":[60],"an":[61],"model":[63],"artifact":[64],"distracts":[66],"semantic":[68,151],"content,":[69],"diversity-based":[71],"exhibit":[73],"structural":[74,102],"blindness,":[75],"disregarding":[76],"user's":[78],"prompt":[79],"and":[80,150,160,177,187],"spatial":[81,148],"redundancy.":[82],"To":[83],"address":[84],"this,":[85],"introduce":[87],"D\u00b2Pruner,":[88],"framework":[90],"rectifies":[92],"these":[93],"issues":[94],"by":[95],"uniquely":[96],"combining":[97],"debiased":[98,122],"importance":[99,176],"with":[100],"mechanism.":[104],"Our":[105],"method":[106],"first":[107],"secures":[108],"core":[110],"set":[111],"most":[114,158],"critical":[115],"tokens":[116,171],"as":[117],"pivots":[118],"based":[119],"attention":[123],"score.":[124],"It":[125],"then":[126],"performs":[127],"Maximal":[129],"Independent":[130],"Set":[131],"(MIS)":[132],"selection":[133],"remaining":[136],"tokens,":[137],"which":[138],"are":[139,172],"modeled":[140],"hybrid":[143],"graph":[144],"where":[145],"edges":[146],"signify":[147],"proximity":[149],"similarity.":[152],"This":[153],"process":[154],"iteratively":[155],"preserves":[156],"important":[159],"available":[161],"removing":[164],"its":[165],"neighbors,":[166],"ensuring":[167],"supplementary":[170],"chosen":[173],"maximize":[175],"diversity.":[178],"Extensive":[179],"experiments":[180],"demonstrate":[181],"D\u00b2Pruner":[183],"achieves":[184],"exceptional":[185],"efficiency":[186],"fidelity.":[188]},"counts_by_year":[],"updated_date":"2026-03-20T20:47:17.329874","created_date":"2026-03-18T00:00:00"}