{"id":"https://openalex.org/W4412889455","doi":"https://doi.org/10.18653/v1/2025.acl-srw.5","title":"Rethinking Full Finetuning from Pretraining Checkpoints in Active Learning for African Languages","display_name":"Rethinking Full Finetuning from Pretraining Checkpoints in Active Learning for African Languages","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412889455","doi":"https://doi.org/10.18653/v1/2025.acl-srw.5"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-srw.5","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-srw.5","pdf_url":"https://aclanthology.org/2025.acl-srw.5.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 4: Student Research Workshop)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-srw.5.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5058219704","display_name":"Bonaventure F. P. Dossou","orcid":"https://orcid.org/0000-0002-0519-1761"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Bonaventure F. P. Dossou","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5021242144","display_name":"In\u00e8s Arous","orcid":"https://orcid.org/0000-0001-7513-6197"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ines Arous","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5079519850","display_name":"Jackie CK Cheung","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jackie CK Cheung","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5058219704"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.09370892,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"64","last_page":"78"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11902","display_name":"Intelligent Tutoring Systems and Adaptive Learning","score":0.9739999771118164,"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/T11902","display_name":"Intelligent Tutoring Systems and Adaptive Learning","score":0.9739999771118164,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.935699999332428,"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/T12884","display_name":"Educational Assessment and Pedagogy","score":0.9355000257492065,"subfield":{"id":"https://openalex.org/subfields/3304","display_name":"Education"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7518535256385803},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.36908188462257385},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.34090656042099},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.3239586055278778}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7518535256385803},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.36908188462257385},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.34090656042099},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.3239586055278778}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.acl-srw.5","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-srw.5","pdf_url":"https://aclanthology.org/2025.acl-srw.5.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 4: Student Research Workshop)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-srw.5","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-srw.5","pdf_url":"https://aclanthology.org/2025.acl-srw.5.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 4: Student Research Workshop)","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education","score":0.8500000238418579}],"awards":[],"funders":[{"id":"https://openalex.org/F4320309480","display_name":"Nvidia","ror":"https://ror.org/03jdj4y14"},{"id":"https://openalex.org/F4320309949","display_name":"Canadian Institute for Advanced Research","ror":"https://ror.org/01sdtdd95"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412889455.pdf","grobid_xml":"https://content.openalex.org/works/W4412889455.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/W3204019825"],"abstract_inverted_index":{"Active":[0],"learning":[1,157],"(AL)":[2],"aims":[3],"to":[4,97,122],"reduce":[5],"annotation":[6],"effort":[7],"by":[8],"iteratively":[9],"selecting":[10],"the":[11,24,52,92,145,163],"most":[12],"informative":[13],"samples":[14,60],"for":[15,88,111,141,155],"labeling.The":[16],"dominant":[17],"strategy":[18,50],"in":[19,37,91,100,158],"AL":[20,167],"involves":[21],"fully":[22],"finetuning":[23,45,68],"model":[25,53],"on":[26,57,130,174],"all":[27],"acquired":[28,59],"data":[29],"after":[30],"each":[31,62],"round,":[32],"which":[33],"is":[34,54,183],"computationally":[35],"expensive":[36],"multilingual":[38,159],"and":[39,76,104,179],"low-resource":[40,143],"settings.This":[41],"paper":[42],"investigates":[43],"continual":[44],"(CF),":[46],"an":[47],"alternative":[48,154],"update":[49],"where":[51],"updated":[55],"only":[56],"newly":[58],"at":[61],"round.We":[63],"evaluate":[64],"CF":[65,83,107,149],"against":[66],"full":[67],"(FA)":[69],"across":[70],"28":[71],"African":[72],"languages":[73,89,112],"using":[74],"MasakhaNEWS":[75],"SIB-200.Our":[77],"analysis":[78],"reveals":[79],"three":[80],"key":[81],"findings.First,":[82],"matches":[84],"or":[85],"outperforms":[86],"FA":[87,138],"included":[90],"model's":[93],"pretraining,":[94],"achieving":[95],"up":[96],"35%":[98],"reductions":[99],"GPU":[101],"memory,":[102],"FLOPs,":[103],"training":[105],"time.Second,":[106],"performs":[108],"comparably":[109],"even":[110],"not":[113],"seen":[114],"during":[115],"pretraining":[116,175],"when":[117],"they":[118],"are":[119],"typologically":[120],"similar":[121],"those":[123],"that":[124,169],"were.Third,":[125],"CF's":[126],"effectiveness":[127],"depends":[128],"critically":[129],"uncertainty-based":[131],"acquisition;":[132],"without":[133],"it,":[134],"performance":[135],"deteriorates":[136],"significantly.While":[137],"remains":[139],"preferable":[140],"some":[142],"languages,":[144],"overall":[146],"results":[147],"establish":[148],"as":[150],"a":[151],"robust,":[152],"cost-efficient":[153],"active":[156],"NLP.These":[160],"findings":[161],"motivate":[162],"development":[164],"of":[165],"hybrid":[166],"strategies":[168],"adapt":[170],"fine-tuning":[171],"behavior":[172],"based":[173],"coverage,":[176],"language":[177],"typology,":[178],"acquisition":[180],"dynamics.Our":[181],"code":[182],"available":[184],"here.":[185]},"counts_by_year":[],"updated_date":"2026-04-21T08:09:41.155169","created_date":"2025-10-10T00:00:00"}