{"id":"https://openalex.org/W4415971245","doi":"https://doi.org/10.1109/tpami.2025.3629605","title":"Efficient and Effective Weight-Ensembling Mixture of Experts for Multi-Task Model Merging","display_name":"Efficient and Effective Weight-Ensembling Mixture of Experts for Multi-Task Model Merging","publication_year":2025,"publication_date":"2025-11-06","ids":{"openalex":"https://openalex.org/W4415971245","doi":"https://doi.org/10.1109/tpami.2025.3629605","pmid":"https://pubmed.ncbi.nlm.nih.gov/41196794"},"language":"en","primary_location":{"id":"doi:10.1109/tpami.2025.3629605","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpami.2025.3629605","pdf_url":null,"source":{"id":"https://openalex.org/S199944782","display_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","issn_l":"0162-8828","issn":["0162-8828","1939-3539","2160-9292"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","pubmed"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5042802004","display_name":"Li Shen","orcid":"https://orcid.org/0000-0001-5659-3464"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Li Shen","raw_affiliation_strings":["Shenzhen Campus of Sun Yat-sen University, Shenzhen, China","Sun Yat-sen University, China"],"affiliations":[{"raw_affiliation_string":"Shenzhen Campus of Sun Yat-sen University, Shenzhen, China","institution_ids":["https://openalex.org/I157773358"]},{"raw_affiliation_string":"Sun Yat-sen University, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103065439","display_name":"Anke Tang","orcid":"https://orcid.org/0000-0002-0576-8153"},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Anke Tang","raw_affiliation_strings":["Wuhan University, Wuhan, China","Wuhan University, China"],"affiliations":[{"raw_affiliation_string":"Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]},{"raw_affiliation_string":"Wuhan University, China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086494876","display_name":"Enneng Yang","orcid":"https://orcid.org/0000-0001-5419-5286"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]},{"id":"https://openalex.org/I9224756","display_name":"Northeastern University","ror":"https://ror.org/03awzbc87","country_code":"CN","type":"education","lineage":["https://openalex.org/I9224756"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Enneng Yang","raw_affiliation_strings":["Shenzhen Campus of Sun Yat-sen University, Shenzhen, China","Northeastern University, China"],"affiliations":[{"raw_affiliation_string":"Shenzhen Campus of Sun Yat-sen University, Shenzhen, China","institution_ids":["https://openalex.org/I157773358"]},{"raw_affiliation_string":"Northeastern University, China","institution_ids":["https://openalex.org/I9224756"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007061198","display_name":"Guibing Guo","orcid":"https://orcid.org/0000-0002-1709-5056"},"institutions":[{"id":"https://openalex.org/I9224756","display_name":"Northeastern University","ror":"https://ror.org/03awzbc87","country_code":"CN","type":"education","lineage":["https://openalex.org/I9224756"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guibing Guo","raw_affiliation_strings":["Northeastern University, Shenyang, China","Northeastern University, China"],"affiliations":[{"raw_affiliation_string":"Northeastern University, Shenyang, China","institution_ids":["https://openalex.org/I9224756"]},{"raw_affiliation_string":"Northeastern University, China","institution_ids":["https://openalex.org/I9224756"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039158745","display_name":"Yong Luo","orcid":"https://orcid.org/0000-0002-2296-6370"},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yong Luo","raw_affiliation_strings":["Wuhan University, Wuhan, China","Wuhan University, China"],"affiliations":[{"raw_affiliation_string":"Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]},{"raw_affiliation_string":"Wuhan University, China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024278302","display_name":"Lefei Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lefei Zhang","raw_affiliation_strings":["Wuhan University, Wuhan, China","Wuhan University, China"],"affiliations":[{"raw_affiliation_string":"Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]},{"raw_affiliation_string":"Wuhan University, China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068837264","display_name":"Xiaochun Cao","orcid":"https://orcid.org/0000-0001-7141-708X"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiaochun Cao","raw_affiliation_strings":["Shenzhen Campus of Sun Yat-sen University, Shenzhen, China","Sun Yat-sen University, China"],"affiliations":[{"raw_affiliation_string":"Shenzhen Campus of Sun Yat-sen University, Shenzhen, China","institution_ids":["https://openalex.org/I157773358"]},{"raw_affiliation_string":"Sun Yat-sen University, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060042752","display_name":"Bo Du","orcid":"https://orcid.org/0000-0002-0059-8458"},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bo Du","raw_affiliation_strings":["Wuhan University, Wuhan, China","Wuhan University, China"],"affiliations":[{"raw_affiliation_string":"Wuhan University, Wuhan, China","institution_ids":["https://openalex.org/I37461747"]},{"raw_affiliation_string":"Wuhan University, China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101409300","display_name":"Dacheng Tao","orcid":null},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Dacheng Tao","raw_affiliation_strings":["Nanyang Technological University, Singapore"],"affiliations":[{"raw_affiliation_string":"Nanyang Technological University, Singapore","institution_ids":["https://openalex.org/I172675005"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":9,"corresponding_author_ids":["https://openalex.org/A5042802004"],"corresponding_institution_ids":["https://openalex.org/I157773358"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.17038354,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"48","issue":"3","first_page":"2329","last_page":"2341"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.7287999987602234,"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"}},"topics":[{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.7287999987602234,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.03669999912381172,"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/T11704","display_name":"Mobile Crowdsensing and Crowdsourcing","score":0.02979999966919422,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/task","display_name":"Task (project management)","score":0.6272000074386597},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.602400004863739},{"id":"https://openalex.org/keywords/core","display_name":"Core (optical fiber)","score":0.41620001196861267},{"id":"https://openalex.org/keywords/task-analysis","display_name":"Task analysis","score":0.41519999504089355},{"id":"https://openalex.org/keywords/routing","display_name":"Routing (electronic design automation)","score":0.412200003862381},{"id":"https://openalex.org/keywords/data-modeling","display_name":"Data modeling","score":0.3467000126838684}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.83160001039505},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6272000074386597},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.602400004863739},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5270000100135803},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4555000066757202},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.41620001196861267},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.41519999504089355},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.412200003862381},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3467000126838684},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.3176000118255615},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.31209999322891235},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.30149999260902405},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.3003000020980835},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2621000111103058},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.25609999895095825},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.2517000138759613},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.25060001015663147}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/tpami.2025.3629605","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tpami.2025.3629605","pdf_url":null,"source":{"id":"https://openalex.org/S199944782","display_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","issn_l":"0162-8828","issn":["0162-8828","1939-3539","2160-9292"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","raw_type":"journal-article"},{"id":"pmid:41196794","is_oa":false,"landing_page_url":"https://pubmed.ncbi.nlm.nih.gov/41196794","pdf_url":null,"source":{"id":"https://openalex.org/S4306525036","display_name":"PubMed","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE transactions on pattern analysis and machine intelligence","raw_type":null}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1163755948","display_name":null,"funder_award_id":"62576364","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G1284973545","display_name":null,"funder_award_id":"62276195","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G1463586485","display_name":null,"funder_award_id":"2024AFA017","funder_id":"https://openalex.org/F4320327654","funder_display_name":"Foundation for Innovative Research Groups of Hubei Province of China"},{"id":"https://openalex.org/G1836220556","display_name":null,"funder_award_id":"62225113","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G4414281452","display_name":null,"funder_award_id":"62025604","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G7474108740","display_name":null,"funder_award_id":"62576083","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8276535520","display_name":null,"funder_award_id":"U23A20318","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320327654","display_name":"Foundation for Innovative Research Groups of Hubei Province of China","ror":null},{"id":"https://openalex.org/F4320329791","display_name":"Shenzhen Fundamental Research Program","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":26,"referenced_works":["https://openalex.org/W2001610032","https://openalex.org/W2017814585","https://openalex.org/W2047643928","https://openalex.org/W2067713319","https://openalex.org/W2112796928","https://openalex.org/W2138011018","https://openalex.org/W2150884987","https://openalex.org/W2592962403","https://openalex.org/W2901458284","https://openalex.org/W2923014074","https://openalex.org/W4313156423","https://openalex.org/W4379013917","https://openalex.org/W4379929801","https://openalex.org/W4385245566","https://openalex.org/W4390814657","https://openalex.org/W4394625633","https://openalex.org/W4400767051","https://openalex.org/W4402671950","https://openalex.org/W4403940973","https://openalex.org/W4404782795","https://openalex.org/W4406472599","https://openalex.org/W4406844855","https://openalex.org/W4415795533","https://openalex.org/W4415796747","https://openalex.org/W4416650086","https://openalex.org/W7121445788"],"related_works":[],"abstract_inverted_index":{"Multi-task":[0],"learning":[1],"(MTL)":[2],"leverages":[3],"a":[4,40,75,120,139],"shared":[5,172],"model":[6,47,84,195,214],"to":[7,57],"accomplish":[8],"multiple":[9,175],"tasks":[10,62,204],"and":[11,63,105,142,170,189,203,209,223],"facilitate":[12],"knowledge":[13],"transfer.":[14],"Recent":[15],"research":[16],"on":[17,135,147],"task":[18],"arithmetic-based":[19],"MTL":[20,220],"demonstrates":[21],"that":[22,206],"merging":[23,36,144,215],"the":[24,45,58,129,165,182,185,193],"parameters":[25],"of":[26,78,101,168,192,219],"independently":[27],"fine-tuned":[28],"models":[29,103],"can":[30],"effectively":[31],"achieve":[32],"MTL.":[33],"However,":[34],"existing":[35],"methods":[37,216],"primarily":[38],"seek":[39],"static":[41],"optimal":[42],"solution":[43],"within":[44],"original":[46],"parameter":[48,96,187],"space,":[49],"which":[50],"often":[51],"results":[52,199],"in":[53,70,98,128,164,217],"performance":[54],"degradation":[55],"due":[56],"inherent":[59],"diversity":[60],"among":[61],"potential":[64],"interferences.":[65],"To":[66],"address":[67],"this":[68,71],"challenge,":[69],"paper,":[72],"we":[73,87,149],"propose":[74],"Weight-Ensembling":[76],"Mixture":[77],"Experts":[79],"(WEMoE)":[80],"method":[81],"for":[82],"multi-task":[83],"merging.":[85],"Specifically,":[86],"first":[88],"identify":[89],"critical":[90,117,166],"(or":[91],"sensitive)":[92],"modules":[93,100,114,118,127,167],"by":[94,196],"analyzing":[95],"variations":[97],"core":[99,158],"Transformer-based":[102],"before":[104],"after":[106],"fine-tuning.":[107],"Then,":[108],"our":[109],"WEMoE":[110,154,169,208],"statically":[111],"merges":[112],"non-critical":[113],"while":[115],"transforming":[116],"into":[119],"mixture-of-experts":[121],"(MoE)":[122],"structure.":[123],"During":[124],"inference,":[125],"expert":[126],"MoE":[130,176],"are":[131],"dynamically":[132],"merged":[133,194],"based":[134],"input":[136],"samples,":[137],"enabling":[138],"more":[140],"flexible":[141],"adaptive":[143],"approach.":[145],"Building":[146],"WEMoE,":[148],"further":[150],"introduce":[151],"an":[152],"efficient-and-effective":[153],"(E-WEMoE)":[155],"method,":[156],"whose":[157],"mechanism":[159],"involves":[160],"eliminating":[161],"non-essential":[162],"elements":[163],"implementing":[171],"routing":[173],"across":[174,200],"modules,":[177],"thereby":[178],"significantly":[179],"reducing":[180],"both":[181,207],"trainable":[183],"parameters,":[184],"overall":[186],"count,":[188],"computational":[190],"overhead":[191],"WEMoE.":[197],"Experimental":[198],"various":[201],"architectures":[202],"demonstrate":[205],"E-WEMoE":[210],"outperform":[211],"state-of-the-art":[212],"(SOTA)":[213],"terms":[218],"performance,":[221],"generalization,":[222],"robustness.":[224]},"counts_by_year":[],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-11-06T00:00:00"}