{"id":"https://openalex.org/W4391376671","doi":"https://doi.org/10.48550/arxiv.2401.15947","title":"MoE-LLaVA: Mixture of Experts for Large Vision-Language Models","display_name":"MoE-LLaVA: Mixture of Experts for Large Vision-Language Models","publication_year":2024,"publication_date":"2024-01-29","ids":{"openalex":"https://openalex.org/W4391376671","doi":"https://doi.org/10.48550/arxiv.2401.15947"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2401.15947","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2401.15947","pdf_url":"https://arxiv.org/pdf/2401.15947","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"text"},"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2401.15947","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101138559","display_name":"Bin Lin","orcid":"https://orcid.org/0009-0008-1384-0152"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Lin, Bin","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101553435","display_name":"Zhenyu Tang","orcid":"https://orcid.org/0000-0002-6998-2669"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tang, Zhenyu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101635291","display_name":"Ye Yang","orcid":"https://orcid.org/0000-0001-9344-9349"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ye, Yang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Huang, Jinfa","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Jinfa","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100371684","display_name":"Junwu Zhang","orcid":"https://orcid.org/0000-0002-6310-8807"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Junwu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Pang, Yatian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pang, Yatian","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100731314","display_name":"Peng Jin","orcid":"https://orcid.org/0000-0003-2228-131X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jin, Peng","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058644424","display_name":"Munan Ning","orcid":"https://orcid.org/0009-0005-3418-085X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ning, Munan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Luo, Jiebo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Jiebo","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Yuan, Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan, Li","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":10,"corresponding_author_ids":["https://openalex.org/A5101138559"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":33,"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":0.9987999796867371,"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":0.9987999796867371,"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":0.9578999876976013,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9528999924659729,"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/computer-science","display_name":"Computer science","score":0.44849443435668945},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.40443164110183716},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.40128540992736816},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.3606405556201935}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.44849443435668945},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.40443164110183716},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.40128540992736816},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.3606405556201935}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2401.15947","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2401.15947","pdf_url":"https://arxiv.org/pdf/2401.15947","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"text"},{"id":"doi:10.48550/arxiv.2401.15947","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2401.15947","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":"pmh:oai:arXiv.org:2401.15947","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2401.15947","pdf_url":"https://arxiv.org/pdf/2401.15947","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"text"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4391376671.pdf","grobid_xml":"https://content.openalex.org/works/W4391376671.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W2358668433","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W2382290278","https://openalex.org/W2350741829","https://openalex.org/W2530322880","https://openalex.org/W1596801655","https://openalex.org/W3204019825"],"abstract_inverted_index":{"Recent":[0],"advances":[1],"demonstrate":[2],"that":[3],"scaling":[4,16],"Large":[5],"Vision-Language":[6],"Models":[7],"(LVLMs)":[8],"effectively":[9],"improves":[10],"downstream":[11],"task":[12],"performances.":[13],"However,":[14],"existing":[15],"methods":[17],"enable":[18],"all":[19],"model":[20,70],"parameters":[21,76],"to":[22,138,159],"be":[23],"active":[24],"for":[25,50,163,170],"each":[26],"token":[27],"in":[28,62,116,151,173],"the":[29,56,85,96,104,111,139,149],"calculation,":[30],"which":[31,92],"brings":[32],"massive":[33],"training":[34,47],"and":[35,122,146,166,177],"inferring":[36],"costs.":[37],"In":[38],"this":[39],"work,":[40],"we":[41,83,157],"propose":[42],"a":[43,68,78,87,117,161],"simple":[44],"yet":[45],"effective":[46,178],"strategy":[48,53],"MoE-Tuning":[49],"LVLMs.":[51],"This":[52],"innovatively":[54],"addresses":[55],"common":[57],"issue":[58],"of":[59,75,114,119],"performance":[60,113,136],"degradation":[61],"multi-modal":[63,179],"sparsity":[64],"learning,":[65],"consequently":[66],"constructing":[67],"sparse":[69,89,164],"with":[71,127],"an":[72],"outrageous":[73],"number":[74],"but":[77],"constant":[79],"computational":[80],"cost.":[81],"Furthermore,":[82],"present":[84],"MoE-LLaVA,":[86,156],"MoE-based":[88],"LVLM":[90],"architecture,":[91],"uniquely":[93],"activates":[94],"only":[95,128],"top-k":[97],"experts":[98,106],"through":[99],"routers":[100],"during":[101],"deployment,":[102],"keeping":[103],"remaining":[105],"inactive.":[107],"Extensive":[108],"experiments":[109],"show":[110],"significant":[112],"MoE-LLaVA":[115,134],"variety":[118],"visual":[120,143],"understanding":[121,144],"object":[123,152],"hallucination":[124,153],"benchmarks.":[125],"Remarkably,":[126],"approximately":[129],"3B":[130],"sparsely":[131],"activated":[132],"parameters,":[133],"demonstrates":[135],"comparable":[137],"LLaVA-1.5-7B":[140],"on":[141],"various":[142],"datasets":[145],"even":[147],"surpasses":[148],"LLaVA-1.5-13B":[150],"benchmark.":[154],"Through":[155],"aim":[158],"establish":[160],"baseline":[162],"LVLMs":[165],"provide":[167],"valuable":[168],"insights":[169],"future":[171],"research":[172],"developing":[174],"more":[175],"efficient":[176],"learning":[180],"systems.":[181],"Code":[182],"is":[183],"released":[184],"at":[185],"https://github.com/PKU-YuanGroup/MoE-LLaVA.":[186]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":16},{"year":2024,"cited_by_count":16}],"updated_date":"2026-04-16T08:26:57.006410","created_date":"2025-10-10T00:00:00"}