{"id":"https://openalex.org/W4412888556","doi":"https://doi.org/10.18653/v1/2025.findings-acl.387","title":"SEE: Continual Fine-tuning with Sequential Ensemble of Experts","display_name":"SEE: Continual Fine-tuning with Sequential Ensemble of Experts","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888556","doi":"https://doi.org/10.18653/v1/2025.findings-acl.387"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.387","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.387","pdf_url":"https://aclanthology.org/2025.findings-acl.387.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":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-acl.387.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5053747215","display_name":"Zhilin Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhilin Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032649444","display_name":"Yafu Li","orcid":"https://orcid.org/0000-0002-7895-9997"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yafu Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5065871371","display_name":"Xiaoye Qu","orcid":"https://orcid.org/0000-0002-4907-3978"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaoye Qu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5084146010","display_name":"Yu Cheng","orcid":"https://orcid.org/0000-0002-4837-3370"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu Cheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"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":0.14396299,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"7418","last_page":"7432"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.7735999822616577,"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/T10320","display_name":"Neural Networks and Applications","score":0.7735999822616577,"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/T10711","display_name":"Target Tracking and Data Fusion in Sensor Networks","score":0.7354999780654907,"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/T12205","display_name":"Time Series Analysis and Forecasting","score":0.7247999906539917,"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":0.6752641201019287},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4218064844608307}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6752641201019287},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4218064844608307}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.387","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.387","pdf_url":"https://aclanthology.org/2025.findings-acl.387.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":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.387","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.387","pdf_url":"https://aclanthology.org/2025.findings-acl.387.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":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412888556.pdf","grobid_xml":"https://content.openalex.org/works/W4412888556.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Continual":[0],"fine-tuning":[1],"of":[2,20,72,108,163,178],"large":[3],"language":[4],"models":[5],"(LLMs)":[6],"suffers":[7],"from":[8],"catastrophic":[9],"forgetting.Rehearsal-based":[10],"methods":[11],"mitigate":[12],"this":[13],"problem":[14],"by":[15],"retaining":[16],"a":[17,43,90,152],"small":[18],"set":[19],"old":[21],"data.Nevertheless,":[22],"they":[23,57],"still":[24],"suffer":[25],"inevitable":[26],"performance":[27],"loss.Although":[28],"training":[29,107],"separate":[30],"experts":[31,110],"for":[32,61,79,111,156,175],"each":[33,84,170],"task":[34],"can":[35,141,147],"help":[36],"prevent":[37],"forgetting,":[38],"effectively":[39,142],"assembling":[40],"them":[41],"remains":[42],"challenge.Some":[44],"approaches":[45],"use":[46],"routers":[47],"to":[48,51,86,151],"assign":[49],"tasks":[50,113],"experts,":[52],"but":[53],"in":[54,130],"continual":[55,101,131],"learning,":[56,129],"often":[58],"require":[59],"retraining":[60,116],"optimal":[62],"performance.To":[63],"address":[64],"these":[65],"challenges,":[66],"we":[67],"introduce":[68],"the":[69,77,106,117,122,139,160,173,176],"Sequential":[70],"Ensemble":[71],"Experts":[73],"(SEE)":[74],"framework.SEE":[75],"removes":[76],"need":[78],"an":[80],"additional":[81],"router,":[82],"allowing":[83],"expert":[85,140],"independently":[87],"decide":[88],"whether":[89],"query":[91],"should":[92],"be":[93,149],"handled.The":[94],"framework":[95],"employs":[96],"distributed":[97,179],"routing,":[98],"and":[99,166],"during":[100],"fine-tuning,":[102],"SEE":[103,123],"only":[104],"requires":[105],"new":[109],"incoming":[112],"rather":[114],"than":[115],"entire":[118],"system.Experiments":[119],"reveal":[120],"that":[121],"outperforms":[124],"prior":[125],"approaches,":[126],"including":[127],"multi-task":[128],"fine-tuning.It":[132],"also":[133],"demonstrates":[134],"remarkable":[135],"generalization":[136],"ability,":[137],"as":[138],"identify":[143],"out-of-distribution":[144],"queries,":[145],"which":[146],"then":[148],"directed":[150],"more":[153],"generalized":[154],"model":[155,180],"resolution.This":[157],"work":[158],"highlights":[159],"promising":[161],"potential":[162],"integrating":[164],"routing":[165],"response":[167],"mechanisms":[168],"within":[169],"expert,":[171],"paving":[172],"way":[174],"future":[177],"ensembling.The":[181],"code":[182],"is":[183],"available":[184],"at":[185],"url.":[186]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}