{"id":"https://openalex.org/W7140129212","doi":"https://doi.org/10.18653/v1/2026.eacl-long.37","title":"Sparse Adapter Fusion for Continual Learning in NLP","display_name":"Sparse Adapter Fusion for Continual Learning in NLP","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7140129212","doi":"https://doi.org/10.18653/v1/2026.eacl-long.37"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2026.eacl-long.37","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.37","pdf_url":"https://aclanthology.org/2026.eacl-long.37.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 19th Conference of the European Chapter 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/2026.eacl-long.37.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Min Zeng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Min Zeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Xi Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xi Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Haiqin Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Haiqin Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Yike Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yike Guo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"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":0.37233292,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"852","last_page":"863"},"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":0.7832000255584717,"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":0.7832000255584717,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.0340999998152256,"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/T10028","display_name":"Topic Modeling","score":0.0215000007301569,"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/fusion","display_name":"Fusion","score":0.3971000015735626},{"id":"https://openalex.org/keywords/adapter","display_name":"Adapter (computing)","score":0.3781000077724457},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.30300000309944153},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.28049999475479126},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.2775000035762787}],"concepts":[{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.7199000120162964},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.61080002784729},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5392000079154968},{"id":"https://openalex.org/C158525013","wikidata":"https://www.wikidata.org/wiki/Q2593739","display_name":"Fusion","level":2,"score":0.3971000015735626},{"id":"https://openalex.org/C177284502","wikidata":"https://www.wikidata.org/wiki/Q1005390","display_name":"Adapter (computing)","level":2,"score":0.3781000077724457},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.30300000309944153},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.29649999737739563},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.28049999475479126},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2775000035762787},{"id":"https://openalex.org/C33954974","wikidata":"https://www.wikidata.org/wiki/Q486494","display_name":"Sensor fusion","level":2,"score":0.25929999351501465},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2572000026702881}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.18653/v1/2026.eacl-long.37","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.37","pdf_url":"https://aclanthology.org/2026.eacl-long.37.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 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-172332","is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-172332","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Conference paper"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.eacl-long.37","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.37","pdf_url":"https://aclanthology.org/2026.eacl-long.37.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 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7140129212.pdf","grobid_xml":"https://content.openalex.org/works/W7140129212.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Continual":[0],"learning":[1],"in":[2,10,80],"natural":[3],"language":[4],"processing":[5],"plays":[6],"a":[7,62,98,133],"crucial":[8],"role":[9],"adapting":[11],"to":[12,75,96,114,136],"evolving":[13],"data":[14],"and":[15,40,72,86,124],"preventing":[16],"catastrophic":[17,34],"forgetting.Despite":[18],"significant":[19],"progress,":[20],"existing":[21,103],"methods":[22],"still":[23],"face":[24],"challenges,":[25],"such":[26],"as":[27],"inefficient":[28],"parameter":[29,122],"reuse":[30,101],"across":[31],"tasks,":[32],"risking":[33],"forgetting":[35],"when":[36],"tasks":[37],"are":[38],"dissimilar,":[39],"the":[41,83,87,90,127,145,166],"unnecessary":[42],"introduction":[43],"of":[44,165],"new":[45,73,99],"parameters":[46,167],"for":[47],"each":[48],"task,":[49],"which":[50,68],"hampers":[51],"knowledge":[52,143],"sharing":[53],"among":[54],"similar":[55],"tasks.To":[56],"tackle":[57],"these":[58,77],"issues,":[59],"we":[60],"propose":[61],"Sparse":[63],"Adapter":[64],"Fusion":[65],"Method":[66],"(SAFM),":[67],"dynamically":[69],"fuses":[70],"old":[71],"adapters":[74],"address":[76],"challenges.SAFM":[78],"operates":[79],"two":[81],"stages:":[82],"decision":[84,91],"stage":[85],"tuning":[88,128],"stage.In":[89],"stage,":[92,129],"SAFM":[93,130,152],"determines":[94],"whether":[95],"incorporate":[97],"adapter,":[100],"an":[102,107],"one,":[104],"or":[105,117],"add":[106],"empty":[108,119],"adapter.The":[109],"architecture":[110],"search":[111],"procedure,":[112],"designed":[113],"prioritize":[115],"reusing":[116],"adding":[118],"adapters,":[120,140],"minimizes":[121],"consumption":[123],"maximizes":[125],"reuse.In":[126],"especially":[131],"facilitates":[132],"layer-wise":[134],"loss":[135],"encourage":[137],"differentiation":[138],"between":[139],"effectively":[141],"capturing":[142],"within":[144],"same":[146],"task.Experimental":[147],"results":[148],"consistently":[149],"show":[150],"that":[151],"outperforms":[153],"state-of-the-art":[154],"(SOTA)":[155],"methods,":[156],"achieving":[157],"comparable":[158],"performance":[159],"while":[160],"utilizing":[161],"less":[162],"than":[163],"60%":[164],"1":[168],".":[169]},"counts_by_year":[],"updated_date":"2026-06-20T22:02:38.213706","created_date":"2026-02-07T00:00:00"}