{"id":"https://openalex.org/W4415428156","doi":"https://doi.org/10.3233/faia251318","title":"Router Upcycling: Leveraging Mixture-of-Routers in Mixture-of-Experts Upcycling","display_name":"Router Upcycling: Leveraging Mixture-of-Routers in Mixture-of-Experts Upcycling","publication_year":2025,"publication_date":"2025-10-21","ids":{"openalex":"https://openalex.org/W4415428156","doi":"https://doi.org/10.3233/faia251318"},"language":null,"primary_location":{"id":"doi:10.3233/faia251318","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia251318","pdf_url":null,"source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},"type":"book-chapter","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.3233/faia251318","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Junfeng Ran","orcid":null},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Junfeng Ran","raw_affiliation_strings":["National Key Laboratory for Multimedia Information Processing, Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Key Laboratory for Multimedia Information Processing, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080990023","display_name":"Gang Zhao","orcid":"https://orcid.org/0000-0001-8584-9510"},"institutions":[{"id":"https://openalex.org/I103261469","display_name":"Guilford Technical Community College","ror":"https://ror.org/03bq13516","country_code":"US","type":"education","lineage":["https://openalex.org/I103261469"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Guangxiang Zhao","raw_affiliation_strings":["Qiyuan Tech"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Qiyuan Tech","institution_ids":["https://openalex.org/I103261469"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031989447","display_name":"Yuhan Wu","orcid":"https://orcid.org/0000-0001-7115-5390"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yuhan Wu","raw_affiliation_strings":["National Key Laboratory for Multimedia Information Processing, Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Key Laboratory for Multimedia Information Processing, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077693662","display_name":"Dawei Zhu","orcid":"https://orcid.org/0000-0003-0167-6603"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dawei Zhu","raw_affiliation_strings":["National Key Laboratory for Multimedia Information Processing, Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Key Laboratory for Multimedia Information Processing, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058133767","display_name":"Lijuan Wu","orcid":"https://orcid.org/0000-0002-5761-5140"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Longyun Wu","raw_affiliation_strings":["National Key Laboratory for Multimedia Information Processing, Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Key Laboratory for Multimedia Information Processing, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048093695","display_name":"Yikai Zhao","orcid":"https://orcid.org/0000-0003-2495-7774"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yikai Zhao","raw_affiliation_strings":["National Key Laboratory for Multimedia Information Processing, Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Key Laboratory for Multimedia Information Processing, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069277955","display_name":"Tong Yang","orcid":"https://orcid.org/0000-0003-2402-5854"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tong Yang","raw_affiliation_strings":["National Key Laboratory for Multimedia Information Processing, Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Key Laboratory for Multimedia Information Processing, Peking University","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037670975","display_name":"Lin Sun","orcid":"https://orcid.org/0000-0002-8781-2057"},"institutions":[{"id":"https://openalex.org/I103261469","display_name":"Guilford Technical Community College","ror":"https://ror.org/03bq13516","country_code":"US","type":"education","lineage":["https://openalex.org/I103261469"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Lin Sun","raw_affiliation_strings":["Qiyuan Tech"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Qiyuan Tech","institution_ids":["https://openalex.org/I103261469"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5047480930","display_name":"X. Brian Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I103261469","display_name":"Guilford Technical Community College","ror":"https://ror.org/03bq13516","country_code":"US","type":"education","lineage":["https://openalex.org/I103261469"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xiangzheng Zhang","raw_affiliation_strings":["Qiyuan Tech"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Qiyuan Tech","institution_ids":["https://openalex.org/I103261469"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5058353424","display_name":"Sujian Li","orcid":"https://orcid.org/0000-0001-7493-0786"},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sujian Li","raw_affiliation_strings":["National Key Laboratory for Multimedia Information Processing, Peking University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National Key Laboratory for Multimedia Information Processing, Peking University","institution_ids":["https://openalex.org/I20231570"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.1812,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.92979832,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":97},"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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.9915000200271606,"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/T10764","display_name":"Privacy-Preserving Technologies in Data","score":0.9915000200271606,"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/T10927","display_name":"Access Control and Trust","score":0.9847000241279602,"subfield":{"id":"https://openalex.org/subfields/3312","display_name":"Sociology and Political Science"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10237","display_name":"Cryptography and Data Security","score":0.9764999747276306,"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/router","display_name":"Router","score":0.7336000204086304},{"id":"https://openalex.org/keywords/reuse","display_name":"Reuse","score":0.597000002861023},{"id":"https://openalex.org/keywords/routing","display_name":"Routing (electronic design automation)","score":0.5896999835968018},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.42829999327659607},{"id":"https://openalex.org/keywords/resource","display_name":"Resource (disambiguation)","score":0.3691999912261963},{"id":"https://openalex.org/keywords/simple","display_name":"Simple (philosophy)","score":0.3646000027656555},{"id":"https://openalex.org/keywords/routing-table","display_name":"Routing table","score":0.3402000069618225},{"id":"https://openalex.org/keywords/resource-allocation","display_name":"Resource allocation","score":0.3292999863624573}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.763700008392334},{"id":"https://openalex.org/C2775896111","wikidata":"https://www.wikidata.org/wiki/Q642560","display_name":"Router","level":2,"score":0.7336000204086304},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.597000002861023},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.5896999835968018},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.4636000096797943},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.42829999327659607},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.4196000099182129},{"id":"https://openalex.org/C206345919","wikidata":"https://www.wikidata.org/wiki/Q20380951","display_name":"Resource (disambiguation)","level":2,"score":0.3691999912261963},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.3646000027656555},{"id":"https://openalex.org/C184896649","wikidata":"https://www.wikidata.org/wiki/Q290066","display_name":"Routing table","level":4,"score":0.3402000069618225},{"id":"https://openalex.org/C29202148","wikidata":"https://www.wikidata.org/wiki/Q287260","display_name":"Resource allocation","level":2,"score":0.3292999863624573},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3273000121116638},{"id":"https://openalex.org/C2983435990","wikidata":"https://www.wikidata.org/wiki/Q22725","display_name":"Network routing","level":3,"score":0.3228999972343445},{"id":"https://openalex.org/C24856439","wikidata":"https://www.wikidata.org/wiki/Q352483","display_name":"Adaptive routing","level":5,"score":0.30230000615119934},{"id":"https://openalex.org/C104954878","wikidata":"https://www.wikidata.org/wiki/Q1648707","display_name":"Routing protocol","level":3,"score":0.301800012588501},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.2962999939918518},{"id":"https://openalex.org/C2780609101","wikidata":"https://www.wikidata.org/wiki/Q17156588","display_name":"Resource management (computing)","level":2,"score":0.2766000032424927},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2597000002861023},{"id":"https://openalex.org/C2984173633","wikidata":"https://www.wikidata.org/wiki/Q22725","display_name":"Routing algorithm","level":4,"score":0.25850000977516174},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.25380000472068787},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.25130000710487366}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.3233/faia251318","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia251318","pdf_url":null,"source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"}],"best_oa_location":{"id":"doi:10.3233/faia251318","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia251318","pdf_url":null,"source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0,30],"Mixture-of-Experts":[1],"(MoE)":[2],"models":[3,27],"have":[4],"gained":[5],"significant":[6],"attention":[7,91,95],"in":[8,107],"deep":[9],"learning":[10],"due":[11],"to":[12,37,76,104],"their":[13],"dynamic":[14],"resource":[15],"allocation":[16],"and":[17,39,118],"superior":[18],"performance":[19,79],"across":[20],"diverse":[21,116],"tasks.":[22],"However,":[23],"efficiently":[24],"training":[25,46],"these":[26],"remains":[28],"challenging.":[29],"MoE":[31,63,81],"upcycling":[32,82,139],"technique":[33,72],"has":[34],"been":[35],"proposed":[36],"reuse":[38],"improve":[40],"existing":[41],"model":[42],"components,":[43],"thereby":[44],"minimizing":[45],"overhead.":[47],"Despite":[48],"this,":[49],"simple":[50],"routers,":[51,55],"such":[52],"as":[53,125],"linear":[54],"often":[56],"struggle":[57],"with":[58,120],"complex":[59],"routing":[60,71],"tasks":[61],"within":[62],"upcycling.":[64,98],"In":[65],"response,":[66],"we":[67],"propose":[68],"a":[69],"novel":[70],"called":[73],"Router":[74],"Upcycling":[75],"enhance":[77],"the":[78,90,121],"of":[80,93],"models.":[83],"Our":[84],"approach":[85],"initializes":[86],"multiple":[87],"routers":[88,100],"from":[89],"heads":[92],"preceding":[94],"layers":[96],"during":[97],"These":[99],"collaboratively":[101],"assign":[102],"tokens":[103],"specialized":[105],"experts":[106],"an":[108],"attention-like":[109],"manner.":[110],"Each":[111],"token":[112],"is":[113],"processed":[114],"into":[115],"queries":[117],"aligned":[119],"experts\u2019":[122],"features":[123],"(serving":[124],"keys).":[126],"Experimental":[127],"results":[128],"demonstrate":[129],"that":[130],"our":[131],"method":[132],"achieves":[133],"state-of-the-art":[134],"(SOTA)":[135],"performance,":[136],"outperforming":[137],"other":[138],"baselines.":[140]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-24T00:00:00"}