{"id":"https://openalex.org/W4412889930","doi":"https://doi.org/10.18653/v1/2025.acl-long.1290","title":"Advancing SMoE for Continuous Domain Adaptation of MLLMs: Adaptive Router and Domain-Specific Loss","display_name":"Advancing SMoE for Continuous Domain Adaptation of MLLMs: Adaptive Router and Domain-Specific Loss","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412889930","doi":"https://doi.org/10.18653/v1/2025.acl-long.1290"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.1290","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1290","pdf_url":"https://aclanthology.org/2025.acl-long.1290.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-long.1290.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100425268","display_name":"Liang Zhang","orcid":"https://orcid.org/0000-0003-1566-4902"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liang Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101967248","display_name":"Ziyao Lu","orcid":"https://orcid.org/0000-0001-9176-6404"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ziyao Lu","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/A5100423958","display_name":"Hui Li","orcid":"https://orcid.org/0000-0003-1923-0669"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hui Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5059589761","display_name":"Jie Zhou","orcid":"https://orcid.org/0000-0002-2404-875X"},"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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.0871142,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"26584","last_page":"26602"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.7806000113487244,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.7806000113487244,"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/T13999","display_name":"Digital Rights Management and Security","score":0.7501999735832214,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.7376000285148621,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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":0.588183581829071},{"id":"https://openalex.org/keywords/router","display_name":"Router","score":0.587475597858429},{"id":"https://openalex.org/keywords/domain-adaptation","display_name":"Domain adaptation","score":0.5822059512138367},{"id":"https://openalex.org/keywords/domain","display_name":"Domain (mathematical analysis)","score":0.5730046629905701},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.5268241763114929},{"id":"https://openalex.org/keywords/computer-network","display_name":"Computer network","score":0.2963666319847107},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.11653324961662292},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.10471391677856445},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.08595225214958191},{"id":"https://openalex.org/keywords/neuroscience","display_name":"Neuroscience","score":0.06024852395057678}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.588183581829071},{"id":"https://openalex.org/C2775896111","wikidata":"https://www.wikidata.org/wiki/Q642560","display_name":"Router","level":2,"score":0.587475597858429},{"id":"https://openalex.org/C2776434776","wikidata":"https://www.wikidata.org/wiki/Q19246213","display_name":"Domain adaptation","level":3,"score":0.5822059512138367},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.5730046629905701},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.5268241763114929},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.2963666319847107},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.11653324961662292},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.10471391677856445},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.08595225214958191},{"id":"https://openalex.org/C169760540","wikidata":"https://www.wikidata.org/wiki/Q207011","display_name":"Neuroscience","level":1,"score":0.06024852395057678},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.1290","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1290","pdf_url":"https://aclanthology.org/2025.acl-long.1290.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-long.1290","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1290","pdf_url":"https://aclanthology.org/2025.acl-long.1290.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[{"display_name":"Climate action","id":"https://metadata.un.org/sdg/13","score":0.44999998807907104}],"awards":[],"funders":[{"id":"https://openalex.org/F4320321878","display_name":"Natural Science Foundation of Fujian Province","ror":null}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412889930.pdf","grobid_xml":"https://content.openalex.org/works/W4412889930.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2748952813","https://openalex.org/W4394775207","https://openalex.org/W4389474468","https://openalex.org/W4300172004","https://openalex.org/W3203792196","https://openalex.org/W4321649381","https://openalex.org/W2997645659","https://openalex.org/W3180787869","https://openalex.org/W4295929828","https://openalex.org/W3156096827"],"abstract_inverted_index":{"Recent":[0],"studies":[1],"have":[2],"explored":[3],"Continual":[4],"Instruction":[5],"Tuning":[6],"(CIT)":[7],"in":[8,65,78],"Multimodal":[9],"Large":[10],"Language":[11],"Models":[12],"(MLLMs),":[13],"with":[14,96],"a":[15,53,70,97,150],"primary":[16],"focus":[17],"on":[18,38,145],"Task-incremental":[19],"CIT,":[20,36],"where":[21],"MLLMs":[22,43],"are":[23],"required":[24],"to":[25,44,105,125,141,153],"continuously":[26],"acquire":[27],"new":[28,45,54,76,151],"tasks.However,":[29],"the":[30,39,93,107,121,155],"more":[31],"practical":[32],"and":[33,160],"challenging":[34],"Domain-incremental":[35],"focused":[37],"continual":[40],"adaptation":[41],"of":[42,57,82,157],"domains,":[46],"remains":[47],"underexplored.In":[48],"this":[49],"paper,":[50],"we":[51,68,91,129,148],"propose":[52],"Sparse":[55],"Mixture":[56],"Expert":[58],"(SMoE)":[59],"based":[60,144],"method":[61,159,168],"for":[62,74,112],"domain-incremental":[63],"CIT":[64],"MLLMs.During":[66],"training,":[67],"learn":[69,126],"domainspecific":[71],"SMoE":[72,94,110,122],"module":[73,95,111],"each":[75,114],"domain":[77,127],"every":[79],"FFN":[80],"sub-layer":[81],"MLLMs,":[83],"preventing":[84],"catastrophic":[85],"forgetting":[86],"caused":[87],"by":[88],"inter-domain":[89],"conflicts.Moreover,":[90],"equip":[92],"domain-specific":[98],"autoregressive":[99],"loss":[100],"(DSAL),":[101],"which":[102],"is":[103],"used":[104],"identify":[106],"most":[108],"suitable":[109],"processing":[113],"test":[115],"instruction":[116,142],"during":[117],"inference.To":[118],"further":[119],"enhance":[120],"module's":[123],"ability":[124],"knowledge,":[128],"design":[130],"an":[131],"adaptive":[132],"thresholdbased":[133],"router":[134],"(AT-Router)":[135],"that":[136,166],"allocates":[137],"computing":[138],"resources":[139],"(experts)":[140],"tokens":[143],"their":[146],"importance.Finally,":[147],"establish":[149],"benchmark":[152],"evaluate":[154],"efficacy":[156],"our":[158,167],"advance":[161],"future":[162],"research.Extensive":[163],"experiments":[164],"show":[165],"consistently":[169],"outperforms":[170],"all":[171],"competitive":[172],"baselines.":[173]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}