{"id":"https://openalex.org/W7170191010","doi":"https://doi.org/10.48550/arxiv.2607.20357","title":"Look Less, Think Faster: Joint Token-Compute Adaptation for Multimodal LLMs","display_name":"Look Less, Think Faster: Joint Token-Compute Adaptation for Multimodal LLMs","publication_year":2026,"publication_date":"2026-07-22","ids":{"openalex":"https://openalex.org/W7170191010","doi":"https://doi.org/10.48550/arxiv.2607.20357"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.20357","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.20357","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,"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.2607.20357","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5143471656","display_name":"Pengcheng Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Pengcheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143528533","display_name":"Zhiquan Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Zhiquan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143530698","display_name":"Jayoung Lee","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lee, Jayoung","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143506689","display_name":"Zhuoyan Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Zhuoyan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143473483","display_name":"Ran Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Ran","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047310442","display_name":"Saurabh Bagchi","orcid":"https://orcid.org/0000-0002-4239-5632"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bagchi, Saurabh","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143501279","display_name":"Yin Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Yin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5055585728","display_name":"Somali Chaterji","orcid":"https://orcid.org/0000-0002-3651-6362"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chaterji, Somali","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":0.9132000207901001,"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":0.9132000207901001,"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":0.021400000900030136,"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/T10028","display_name":"Topic Modeling","score":0.00800000037997961,"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/inference","display_name":"Inference","score":0.6276000142097473},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.542900025844574},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.5414999723434448},{"id":"https://openalex.org/keywords/structured-prediction","display_name":"Structured prediction","score":0.5044999718666077},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4657000005245209},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.42239999771118164},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.4034999907016754},{"id":"https://openalex.org/keywords/estimator","display_name":"Estimator","score":0.4034000039100647},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.3921000063419342},{"id":"https://openalex.org/keywords/controller","display_name":"Controller (irrigation)","score":0.3743000030517578}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7807000279426575},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6276000142097473},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.542900025844574},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.5414999723434448},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.5044999718666077},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4657000005245209},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.44690001010894775},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.446399986743927},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.42239999771118164},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.4034999907016754},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.4034000039100647},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3921000063419342},{"id":"https://openalex.org/C203479927","wikidata":"https://www.wikidata.org/wiki/Q5165939","display_name":"Controller (irrigation)","level":2,"score":0.3743000030517578},{"id":"https://openalex.org/C2986314615","wikidata":"https://www.wikidata.org/wiki/Q36829","display_name":"Pareto optimal","level":3,"score":0.36070001125335693},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.3409000039100647},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.3391000032424927},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.33559998869895935},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.31949999928474426},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.31209999322891235},{"id":"https://openalex.org/C6683253","wikidata":"https://www.wikidata.org/wiki/Q7075535","display_name":"Obstacle avoidance","level":4,"score":0.31040000915527344},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.30880001187324524},{"id":"https://openalex.org/C68781425","wikidata":"https://www.wikidata.org/wiki/Q2052203","display_name":"Multi-objective optimization","level":2,"score":0.3086000084877014},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.3005000054836273},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.2840000092983246},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.2718000113964081},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.26739999651908875},{"id":"https://openalex.org/C56461940","wikidata":"https://www.wikidata.org/wiki/Q970687","display_name":"Eye tracking","level":2,"score":0.26440000534057617},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.2639999985694885},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.25110000371932983},{"id":"https://openalex.org/C137635306","wikidata":"https://www.wikidata.org/wiki/Q182667","display_name":"Pareto principle","level":2,"score":0.2500999867916107}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.20357","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.20357","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,"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.2607.20357","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.20357","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,"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/9","score":0.47175246477127075,"display_name":"Industry, innovation and infrastructure"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Multimodal":[0],"Large":[1],"Language":[2],"Models":[3],"(MLLMs)":[4],"have":[5],"recently":[6],"demonstrated":[7],"strong":[8],"performance":[9],"across":[10,89,213],"vision-language":[11],"tasks.":[12],"However,":[13],"their":[14],"high":[15],"inference":[16,107,164],"cost,":[17],"arising":[18],"from":[19],"both":[20,205],"the":[21,29,33,50,81,94,99,162,177],"large":[22,34],"number":[23,114],"of":[24,32],"input":[25,95,123,206],"visual":[26,60,138],"tokens":[27,61,139],"and":[28,66,76,115,125,140,184,208,228],"heavy":[30],"computation":[31],"language":[35],"model":[36,116],"(LLM),":[37],"remains":[38],"a":[39,78,104,130,167,180,186],"key":[40],"barrier":[41],"to":[42,48,121,155,197,204],"practical":[43],"deployment.":[44],"Recent":[45],"work":[46],"attempts":[47],"reduce":[49],"cost":[51,165],"by":[52],"adaptively":[53,146],"optimizing":[54],"individual":[55],"dimensions,":[56],"e.g.,":[57],"pruning":[58],"redundant":[59],"or":[62],"skipping":[63],"LLM":[64,148],"layers":[65],"heads.":[67],"Nonetheless,":[68],"prior":[69,225],"approaches":[70],"typically":[71],"treat":[72],"these":[73,151],"dimensions":[74,91],"independently":[75],"overlook":[77],"fundamental":[79],"coupling:":[80],"available":[82],"compute":[83,117,126,143,210],"resources":[84],"must":[85],"be":[86],"dynamically":[87,135],"allocated":[88],"all":[90],"based":[92],"on":[93],"content.":[96],"To":[97,170],"bridge":[98],"gap,":[100],"we":[101,175],"propose":[102],"SmartVL,":[103],"unified":[105],"adaptive":[106,226],"framework":[108],"that":[109,134,145,161,202],"jointly":[110],"controls":[111],"vision":[112],"token":[113,132],"capability":[118],"in":[119],"response":[120],"varying":[122],"contents":[124],"budgets.":[127],"SmartVL":[128,196,222],"introduces":[129],"vision-side":[131],"controller":[133,144],"selects":[136],"informative":[137],"an":[141],"LLM-side":[142],"adjusts":[147],"computation.":[149],"Importantly,":[150],"controllers":[152,178],"are":[153],"trained":[154],"coordinate":[156],"with":[157,219],"each":[158],"other":[159],"so":[160],"overall":[163],"satisfies":[166],"target":[168],"budget.":[169],"allow":[171],"this":[172],"joint":[173,220],"scheduling,":[174,221],"connect":[176],"using":[179],"shared":[181],"budget":[182],"encoding":[183],"leverage":[185],"differentiable":[187],"latency":[188],"estimator":[189],"for":[190],"end-to-end":[191],"training.":[192],"This":[193],"design":[194],"enables":[195],"learn":[198],"cross-stage":[199],"allocation":[200],"strategies":[201],"adapt":[203],"complexity":[207],"runtime":[209],"constraints.":[211],"Experiments":[212],"multiple":[214],"MLLM":[215],"benchmarks":[216],"demonstrate":[217],"that,":[218],"consistently":[223],"outperforms":[224],"methods":[227],"achieves":[229],"superior":[230],"accuracy-efficiency":[231],"Pareto":[232],"frontiers.":[233],"Project":[234],"page:":[235],"https://www.schaterji.io/publications/2026/jointtokencompute.":[236]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-24T00:00:00"}