{"id":"https://openalex.org/W4396844175","doi":"https://doi.org/10.1145/3589335.3648321","title":"Large Multimodal Model Compression via Iterative Efficient Pruning and Distillation","display_name":"Large Multimodal Model Compression via Iterative Efficient Pruning and Distillation","publication_year":2024,"publication_date":"2024-05-12","ids":{"openalex":"https://openalex.org/W4396844175","doi":"https://doi.org/10.1145/3589335.3648321"},"language":"en","primary_location":{"id":"doi:10.1145/3589335.3648321","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3589335.3648321","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3589335.3648321","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Companion Proceedings of the ACM Web Conference 2024","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3589335.3648321","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5021037797","display_name":"Maolin Wang","orcid":"https://orcid.org/0000-0002-0073-0172"},"institutions":[{"id":"https://openalex.org/I168719708","display_name":"City University of Hong Kong","ror":"https://ror.org/03q8dnn23","country_code":"HK","type":"education","lineage":["https://openalex.org/I168719708"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Maolin Wang","raw_affiliation_strings":["City University of Hong Kong, Hong Kong, China"],"affiliations":[{"raw_affiliation_string":"City University of Hong Kong, Hong Kong, China","institution_ids":["https://openalex.org/I168719708"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019108113","display_name":"Yao Zhao","orcid":"https://orcid.org/0000-0002-0807-929X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yao Zhao","raw_affiliation_strings":["AntGroup, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"AntGroup, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060310537","display_name":"Jiajia Liu","orcid":"https://orcid.org/0000-0003-3020-2677"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiajia Liu","raw_affiliation_strings":["AntGroup, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"AntGroup, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014922345","display_name":"Jingdong Chen","orcid":"https://orcid.org/0000-0002-1872-2592"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jingdong Chen","raw_affiliation_strings":["AntGroup, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"AntGroup, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5061810614","display_name":"Chenyi Zhuang","orcid":"https://orcid.org/0000-0003-1891-5666"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chenyi Zhuang","raw_affiliation_strings":["AntGroup, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"AntGroup, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5053242349","display_name":"Jinjie Gu","orcid":"https://orcid.org/0000-0001-7596-4945"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinjie Gu","raw_affiliation_strings":["AntGroup, Hangzhou, China"],"affiliations":[{"raw_affiliation_string":"AntGroup, Hangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054719216","display_name":"Ruocheng Guo","orcid":"https://orcid.org/0000-0002-8522-6142"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ruocheng Guo","raw_affiliation_strings":["ByteDance Research, London, United Kingdom"],"affiliations":[{"raw_affiliation_string":"ByteDance Research, London, United Kingdom","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100645854","display_name":"Xiangyu Zhao","orcid":"https://orcid.org/0000-0003-2926-4416"},"institutions":[{"id":"https://openalex.org/I168719708","display_name":"City University of Hong Kong","ror":"https://ror.org/03q8dnn23","country_code":"HK","type":"education","lineage":["https://openalex.org/I168719708"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xiangyu Zhao","raw_affiliation_strings":["City University of Hong Kong, Hong Kong, China"],"affiliations":[{"raw_affiliation_string":"City University of Hong Kong, Hong Kong, China","institution_ids":["https://openalex.org/I168719708"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5021037797"],"corresponding_institution_ids":["https://openalex.org/I168719708"],"apc_list":null,"apc_paid":null,"fwci":1.9599,"has_fulltext":false,"cited_by_count":8,"citation_normalized_percentile":{"value":0.87050579,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":94,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"235","last_page":"244"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10901","display_name":"Advanced Data Compression Techniques","score":0.9965000152587891,"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/T10901","display_name":"Advanced Data Compression Techniques","score":0.9965000152587891,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.9807000160217285,"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/T10860","display_name":"Speech and Audio Processing","score":0.9621000289916992,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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":0.7255657911300659},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.6904696226119995},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.6413254141807556},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.598303496837616},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.5105815529823303},{"id":"https://openalex.org/keywords/iterative-method","display_name":"Iterative method","score":0.41666197776794434},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.38467487692832947},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.2371421456336975}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7255657911300659},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.6904696226119995},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.6413254141807556},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.598303496837616},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.5105815529823303},{"id":"https://openalex.org/C159694833","wikidata":"https://www.wikidata.org/wiki/Q2321565","display_name":"Iterative method","level":2,"score":0.41666197776794434},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.38467487692832947},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2371421456336975},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C6557445","wikidata":"https://www.wikidata.org/wiki/Q173113","display_name":"Agronomy","level":1,"score":0.0},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0},{"id":"https://openalex.org/C178790620","wikidata":"https://www.wikidata.org/wiki/Q11351","display_name":"Organic chemistry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3589335.3648321","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3589335.3648321","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3589335.3648321","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Companion Proceedings of the ACM Web Conference 2024","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3589335.3648321","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3589335.3648321","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3589335.3648321","source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Companion Proceedings of the ACM Web Conference 2024","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320323817","display_name":"Universitas Brawijaya","ror":"https://ror.org/01wk3d929"}],"has_content":{"pdf":true,"grobid_xml":false},"content_urls":{"pdf":"https://content.openalex.org/works/W4396844175.pdf"},"referenced_works_count":12,"referenced_works":["https://openalex.org/W2276892413","https://openalex.org/W2965862774","https://openalex.org/W2990714382","https://openalex.org/W2996913633","https://openalex.org/W2998183051","https://openalex.org/W3015609966","https://openalex.org/W3034288893","https://openalex.org/W3174657338","https://openalex.org/W4382463788","https://openalex.org/W4385571831","https://openalex.org/W4386076668","https://openalex.org/W6775706467"],"related_works":["https://openalex.org/W2373300491","https://openalex.org/W2378744544","https://openalex.org/W2594301978","https://openalex.org/W2379704676","https://openalex.org/W1998810860","https://openalex.org/W4206442282","https://openalex.org/W2384505857","https://openalex.org/W2355171581","https://openalex.org/W2612632602","https://openalex.org/W2321805087"],"abstract_inverted_index":{"The":[0],"deployment":[1,29,188],"of":[2,30,49,115,122,189],"Large":[3],"Multimodal":[4,98],"Models":[5],"(LMMs)":[6],"within":[7,106],"Ant":[8],"Group":[9],"has":[10],"significantly":[11],"advanced":[12,86],"multimodal":[13,134],"tasks":[14,24],"in":[15,25,37,127,131,150],"payment,":[16],"security,":[17],"and":[18,40,83,108],"advertising,":[19],"notably":[20],"enhancing":[21],"advertisement":[22,135],"audition":[23,136],"Alipay.":[26],"However,":[27],"the":[28,47,97,113,120,186],"such":[31],"sizable":[32],"models":[33],"introduces":[34,54],"challenges,":[35],"particularly":[36],"increased":[38],"latency":[39],"carbon":[41],"emissions,":[42],"which":[43],"are":[44],"antithetical":[45],"to":[46,111,156,174,185,194],"ideals":[48],"Green":[50],"AI.":[51],"This":[52],"paper":[53],"a":[55,95,147,164],"novel":[56],"multi-stage":[57,81],"compression":[58],"strategy":[59,124],"for":[60,137],"our":[61,91,116,123,144,169,192],"proprietary":[62],"LLM,":[63],"AntGMM.":[64],"Our":[65],"methodology":[66],"pivots":[67],"on":[68],"three":[69,138],"main":[70],"aspects:":[71],"employing":[72],"small":[73],"training":[74],"sample":[75],"sizes,":[76],"addressing":[77],"multi-level":[78],"redundancy":[79],"through":[80],"pruning,":[82],"introducing":[84],"an":[85],"distillation":[87],"loss":[88],"design.":[89],"In":[90],"research,":[92],"we":[93],"constructed":[94],"dataset,":[96],"Advertisement":[99],"Audition":[100],"Dataset":[101],"(MAAD),":[102],"from":[103,140,154],"real-world":[104,133],"scenarios":[105],"Alipay,":[107],"conducted":[109],"experiments":[110],"validate":[112],"reliability":[114],"proposed":[117],"strategy.":[118],"Furthermore,":[119],"effectiveness":[121],"is":[125,172],"evident":[126],"its":[128],"operational":[129],"success":[130],"Alipay's":[132],"months":[139],"September":[141],"2023.":[142],"Notably,":[143],"approach":[145],"achieved":[146],"substantial":[148],"reduction":[149],"latency,":[151],"decreasing":[152],"it":[153],"700ms":[155],"90ms,":[157],"while":[158],"maintaining":[159],"online":[160],"performance":[161,166],"with":[162],"only":[163],"slight":[165],"decrease.":[167],"Moreover,":[168],"compressed":[170],"model":[171],"estimated":[173],"reduce":[175],"electricity":[176],"consumption":[177],"by":[178],"approximately":[179],"75":[180],"million":[181],"kWh":[182],"annually":[183],"compared":[184],"direct":[187],"AntGMM,":[190],"demonstrating":[191],"commitment":[193],"green":[195],"AI":[196],"initiatives.":[197]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":5},{"year":2024,"cited_by_count":2}],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-10-10T00:00:00"}