{"id":"https://openalex.org/W4412888035","doi":"https://doi.org/10.18653/v1/2025.findings-acl.865","title":"AVG-LLaVA: An Efficient Large Multimodal Model with Adaptive Visual Granularity","display_name":"AVG-LLaVA: An Efficient Large Multimodal Model with Adaptive Visual Granularity","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888035","doi":"https://doi.org/10.18653/v1/2025.findings-acl.865"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.865","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.865","pdf_url":"https://aclanthology.org/2025.findings-acl.865.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-acl.865.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5112974306","display_name":"Zhibin Lan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhibin Lan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005680009","display_name":"Liqiang Niu","orcid":"https://orcid.org/0000-0002-3366-9486"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liqiang Niu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024849044","display_name":"Fandong Meng","orcid":"https://orcid.org/0000-0002-8158-2377"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fandong Meng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100336606","display_name":"Wenbo Li","orcid":"https://orcid.org/0000-0003-4604-778X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wenbo Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102010330","display_name":"Jie Zhou","orcid":"https://orcid.org/0000-0001-6521-1384"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jie Zhou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5066326238","display_name":"Jinsong Su","orcid":"https://orcid.org/0000-0001-5606-7122"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinsong Su","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.9349,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.77173237,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":null,"issue":null,"first_page":"16852","last_page":"16869"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9693999886512756,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9693999886512756,"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/T10824","display_name":"Image Retrieval and Classification Techniques","score":0.9569000005722046,"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/T10531","display_name":"Advanced Vision and Imaging","score":0.9552000164985657,"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/granularity","display_name":"Granularity","score":0.922924280166626},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7869426012039185},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.38700777292251587},{"id":"https://openalex.org/keywords/human\u2013computer-interaction","display_name":"Human\u2013computer interaction","score":0.34624746441841125},{"id":"https://openalex.org/keywords/computer-vision","display_name":"Computer vision","score":0.32687655091285706},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.3217797875404358},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.15716639161109924}],"concepts":[{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.922924280166626},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7869426012039185},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.38700777292251587},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.34624746441841125},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.32687655091285706},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3217797875404358},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.15716639161109924}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.865","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.865","pdf_url":"https://aclanthology.org/2025.findings-acl.865.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.865","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.865","pdf_url":"https://aclanthology.org/2025.findings-acl.865.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320321878","display_name":"Natural Science Foundation of Fujian Province","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412888035.pdf","grobid_xml":"https://content.openalex.org/works/W4412888035.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2772917594","https://openalex.org/W2036807459","https://openalex.org/W2058170566","https://openalex.org/W2755342338","https://openalex.org/W2166024367","https://openalex.org/W3116076068","https://openalex.org/W2229312674","https://openalex.org/W2951359407","https://openalex.org/W2079911747","https://openalex.org/W1969923398"],"abstract_inverted_index":{"Recently,":[0],"large":[1,29],"multimodal":[2],"models":[3],"(LMMs)":[4],"have":[5],"achieved":[6],"significant":[7],"advancements.When":[8],"dealing":[9],"with":[10,119],"high-resolution":[11],"images,":[12],"dominant":[13],"LMMs":[14],"typically":[15],"divide":[16],"them":[17],"into":[18],"multiple":[19,60],"local":[20],"images":[21],"and":[22,54,84,99,134,155,166,178],"a":[23,28,72,78,85,105,167],"global":[24],"image,":[25],"leading":[26],"to":[27,63,89],"number":[30,151],"of":[31,122,152],"visual":[32,47,65,73,93,153,164],"tokens.In":[33],"this":[34],"work,":[35],"we":[36,56,70,101],"introduce":[37],"AVG-LLaVA,":[38],"an":[39,81,160],"LMM":[40],"that":[41,109,137],"can":[42,180],"adaptively":[43],"select":[44,90],"the":[45,51,59,91,97,113,117,120,123,126,150,174],"appropriate":[46,92],"granularity":[48,74,94,114],"based":[49,95],"on":[50,96,173],"input":[52],"image":[53,98],"instruction.Specifically,":[55],"first":[57],"apply":[58],"pooling":[61],"layers":[62],"obtain":[64],"tokens":[66,154,165],"at":[67,111,183],"different":[68],"granularities.Then":[69],"propose":[71],"router,":[75],"which":[76],"includes":[77],"Transformer":[79],"layer,":[80,83,87],"MLP":[82],"voter":[86],"used":[88],"instruction.Furthermore,":[100],"put":[102],"forward":[103],"RGLF,":[104],"novel":[106],"training":[107],"paradigm":[108],"aims":[110],"aligning":[112],"predicted":[115],"by":[116],"router":[118],"preferences":[121],"LMM,":[124],"without":[125],"need":[127],"for":[128],"additional":[129],"manually":[130],"annotated":[131],"data.Extensive":[132],"experiments":[133],"analysis":[135],"show":[136],"AVG-LLaVA":[138],"achieves":[139],"superior":[140],"performance":[141],"across":[142],"11":[143],"benchmarks,":[144],"as":[145,147],"well":[146],"significantly":[148],"reduces":[149],"speeds":[156],"up":[157],"inference":[158,171],"(e.g.,":[159],"85.3%":[161],"reduction":[162],"in":[163,170],"2.53":[168],"increase":[169],"speed":[172],"AI2D":[175],"benchmark).Our":[176],"code":[177],"model":[179],"be":[181],"found":[182],"https:":[184],"//github.com/DeepLearnXMU/AVG-LLaVA.":[185]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}