{"id":"https://openalex.org/W7152513768","doi":"https://doi.org/10.48550/arxiv.2604.06912","title":"Q-Zoom: Query-Aware Adaptive Perception for Efficient Multimodal Large Language Models","display_name":"Q-Zoom: Query-Aware Adaptive Perception for Efficient Multimodal Large Language Models","publication_year":2026,"publication_date":"2026-04-08","ids":{"openalex":"https://openalex.org/W7152513768","doi":"https://doi.org/10.48550/arxiv.2604.06912"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.06912","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06912","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.06912","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5124354341","display_name":"Yuheng Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Yuheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060854400","display_name":"Xiaohuan Pei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pei, Xiaohuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067578488","display_name":"Linfeng Wen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wen, Linfeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133281286","display_name":"Minjing Dong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dong, Minjing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133244113","display_name":"Chang Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Chang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"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":0.6977999806404114,"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":0.6977999806404114,"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":0.045899998396635056,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.04309999942779541,"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/inference","display_name":"Inference","score":0.5903000235557556},{"id":"https://openalex.org/keywords/fuse","display_name":"Fuse (electrical)","score":0.5047000050544739},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.48089998960494995},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.4763000011444092},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.4235000014305115},{"id":"https://openalex.org/keywords/encoding","display_name":"Encoding (memory)","score":0.35019999742507935},{"id":"https://openalex.org/keywords/subnetwork","display_name":"Subnetwork","score":0.3472999930381775},{"id":"https://openalex.org/keywords/margin","display_name":"Margin (machine learning)","score":0.33640000224113464},{"id":"https://openalex.org/keywords/perception","display_name":"Perception","score":0.33570000529289246}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8023999929428101},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5903000235557556},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5170999765396118},{"id":"https://openalex.org/C141353440","wikidata":"https://www.wikidata.org/wiki/Q182221","display_name":"Fuse (electrical)","level":2,"score":0.5047000050544739},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.48089998960494995},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.4763000011444092},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.4235000014305115},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3749000132083893},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.35019999742507935},{"id":"https://openalex.org/C2780186347","wikidata":"https://www.wikidata.org/wiki/Q11414","display_name":"Subnetwork","level":2,"score":0.3472999930381775},{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.33640000224113464},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.33570000529289246},{"id":"https://openalex.org/C194544171","wikidata":"https://www.wikidata.org/wiki/Q21105679","display_name":"Gating","level":2,"score":0.33079999685287476},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.3260999917984009},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.30379998683929443},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.301800012588501},{"id":"https://openalex.org/C137635306","wikidata":"https://www.wikidata.org/wiki/Q182667","display_name":"Pareto principle","level":2,"score":0.2985000014305115},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.29829999804496765},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.289000004529953},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.2849000096321106},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.28029999136924744},{"id":"https://openalex.org/C123138037","wikidata":"https://www.wikidata.org/wiki/Q5570871","display_name":"Global network","level":2,"score":0.2736999988555908},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2728999853134155},{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.27070000767707825},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.26750001311302185},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.2549000084400177},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.250900000333786}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.06912","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06912","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.06912","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06912","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[{"score":0.6277834177017212,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"MLLMs":[0],"require":[1],"high-resolution":[2,51,69],"visual":[3],"inputs":[4],"for":[5,77,188],"fine-grained":[6,80],"tasks":[7],"like":[8],"document":[9],"understanding":[10],"and":[11,39,131,173,200,214],"dense":[12,138],"scene":[13],"perception.":[14],"However,":[15],"current":[16],"global":[17,73,144],"resolution":[18],"scaling":[19],"paradigms":[20],"indiscriminately":[21],"flood":[22],"the":[23,90,104,118,137,142,181,194],"quadratic":[24],"self-attention":[25],"mechanism":[26],"with":[27,141],"visually":[28],"redundant":[29],"tokens,":[30],"severely":[31],"bottlenecking":[32],"inference":[33,164],"throughput":[34],"while":[35,117,179],"ignoring":[36],"spatial":[37],"sparsity":[38],"query":[40],"intent.":[41],"To":[42,99],"overcome":[43],"this,":[44],"we":[45],"propose":[46],"Q-Zoom,":[47],"a":[48,62,82,108,121,152,159],"query-aware":[49],"adaptive":[50],"perception":[52],"framework":[53],"that":[54,149],"operates":[55],"in":[56,176],"an":[57],"efficient":[58],"coarse-to-fine":[59],"manner.":[60],"First,":[61],"lightweight":[63],"Dynamic":[64],"Gating":[65],"Network":[66,86],"safely":[67],"bypasses":[68],"processing":[70],"when":[71,186],"coarse":[72,143],"features":[74],"suffice.":[75],"Second,":[76],"queries":[78],"demanding":[79],"perception,":[81],"Self-Distilled":[83],"Region":[84],"Proposal":[85],"(SD-RPN)":[87],"precisely":[88],"localizes":[89],"task-relevant":[91],"Region-of-Interest":[92],"(RoI)":[93],"directly":[94],"from":[95],"intermediate":[96],"feature":[97],"spaces.":[98],"optimize":[100],"these":[101,203],"modules":[102],"efficiently,":[103],"gating":[105],"network":[106],"uses":[107],"consistency-aware":[109],"generation":[110],"strategy":[111],"to":[112,211],"derive":[113],"deterministic":[114],"routing":[115],"labels,":[116],"SD-RPN":[119],"employs":[120],"fully":[122],"self-supervised":[123],"distillation":[124],"paradigm.":[125],"A":[126],"continuous":[127],"spatio-temporal":[128],"alignment":[129],"scheme":[130],"targeted":[132],"fine-tuning":[133],"then":[134],"seamlessly":[135,210],"fuse":[136],"local":[139],"RoI":[140],"layout.":[145],"Extensive":[146],"experiments":[147],"demonstrate":[148],"Q-Zoom":[150,162,192],"establishes":[151],"dominant":[153],"Pareto":[154],"frontier.":[155],"Using":[156],"Qwen2.5-VL-7B":[157],"as":[158],"primary":[160],"testbed,":[161],"accelerates":[163],"by":[165,198],"2.52":[166],"times":[167,175],"on":[168,202],"Document":[169],"&":[170],"OCR":[171],"benchmarks":[172],"4.39":[174],"High-Resolution":[177],"scenarios":[178],"matching":[180],"baseline's":[182,195],"peak":[183,196],"accuracy.":[184],"Furthermore,":[185],"configured":[187],"maximum":[189],"perceptual":[190],"fidelity,":[191],"surpasses":[193],"performance":[197],"1.1%":[199],"8.1%":[201],"respective":[204],"benchmarks.":[205],"These":[206],"robust":[207],"improvements":[208],"transfer":[209],"Qwen3-VL,":[212],"LLaVA,":[213],"emerging":[215],"RL-based":[216],"thinking-with-image":[217],"models.":[218],"Project":[219],"page":[220],"is":[221],"available":[222],"at":[223],"https://yuhengsss.github.io/Q-Zoom/.":[224]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-04-10T00:00:00"}