{"id":"https://openalex.org/W7202212165","doi":"https://doi.org/10.48550/arxiv.2608.07629","title":"Embedding Initialization for Unseen Low-resource Languages in Multilingual NMT: A Case Study on Limbum-English Translation","display_name":"Embedding Initialization for Unseen Low-resource Languages in Multilingual NMT: A Case Study on Limbum-English Translation","publication_year":2026,"publication_date":"2026-08-07","ids":{"openalex":"https://openalex.org/W7202212165","doi":"https://doi.org/10.48550/arxiv.2608.07629"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.07629","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.07629","pdf_url":"https://arxiv.org/pdf/2608.07629","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.07629","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5118595744","display_name":"Samiratu Ntohsi","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405266581","display_name":"African Leadership University","ror":"https://ror.org/040n04m36","country_code":"RW","type":"education","lineage":["https://openalex.org/I4405266581"]}],"countries":["RW"],"is_corresponding":false,"raw_author_name":"Ntohsi, Samiratu","raw_affiliation_strings":["Ruzivo Research Lab , African Leadership University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Ruzivo Research Lab , African Leadership University","institution_ids":["https://openalex.org/I4405266581"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5146551995","display_name":"Neza David Tuyishimire","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tuyishimire, Neza David","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5146442199","display_name":"Anesu Kafesu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kafesu, Anesu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057313166","display_name":"Marvin Muyonga Ogore","orcid":"https://orcid.org/0000-0002-1644-9625","observed_orcids":["https://orcid.org/0000-0002-1644-9625"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ogore, Marvin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134707161","display_name":"Samuel Babalola","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Babalola, Samuel Oluwajunwonlo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130351995","display_name":"Oche David Ankeli","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ankeli, Oche","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9854000210762024,"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":0.9854000210762024,"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":0.009600000455975533,"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":0.0017999999690800905,"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/low-resource-languages","display_name":"low-resource languages","score":0.8420000076293945},{"id":"https://openalex.org/keywords/embedding-initialization","display_name":"embedding initialization","score":0.878000020980835},{"id":"https://openalex.org/keywords/multilingual-neural-machine-translation","display_name":"multilingual neural machine translation","score":0.7710000276565552},{"id":"https://openalex.org/keywords/nllb-200","display_name":"NLLB-200","score":0.9769999980926514},{"id":"https://openalex.org/keywords/bantu-languages","display_name":"Bantu languages","score":0.9139999747276306},{"id":"https://openalex.org/keywords/zero-shot-translation","display_name":"zero-shot translation","score":0.6489999890327454},{"id":"https://openalex.org/keywords/multilingual-transfer","display_name":"multilingual transfer","score":0.7319999933242798},{"id":"https://openalex.org/keywords/new-testament-translation","display_name":"New Testament translation","score":0.4729999899864197}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8065000176429749},{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.7218000292778015},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.6383000016212463},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6212999820709229},{"id":"https://openalex.org/C203005215","wikidata":"https://www.wikidata.org/wiki/Q79798","display_name":"Machine translation","level":2,"score":0.6148999929428101},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.49059998989105225},{"id":"https://openalex.org/C99878080","wikidata":"https://www.wikidata.org/wiki/Q33146","display_name":"Bantu languages","level":2,"score":0.48840001225471497},{"id":"https://openalex.org/C2779913364","wikidata":"https://www.wikidata.org/wiki/Q7838","display_name":"Swahili","level":2,"score":0.4691999852657318},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.42309999465942383},{"id":"https://openalex.org/C2777530160","wikidata":"https://www.wikidata.org/wiki/Q41796","display_name":"Sentence","level":2,"score":0.37720000743865967},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3659999966621399},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.35760000348091125},{"id":"https://openalex.org/C140745168","wikidata":"https://www.wikidata.org/wiki/Q1210082","display_name":"Tree traversal","level":2,"score":0.3427000045776367},{"id":"https://openalex.org/C2985367798","wikidata":"https://www.wikidata.org/wiki/Q1346592","display_name":"Parallel corpora","level":3,"score":0.3312000036239624},{"id":"https://openalex.org/C48105269","wikidata":"https://www.wikidata.org/wiki/Q1141160","display_name":"Header","level":2,"score":0.3061000108718872},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.29980000853538513},{"id":"https://openalex.org/C2778121359","wikidata":"https://www.wikidata.org/wiki/Q8096","display_name":"Lexicon","level":2,"score":0.29980000853538513},{"id":"https://openalex.org/C2986862884","wikidata":"https://www.wikidata.org/wiki/Q7553","display_name":"Language translation","level":3,"score":0.29580000042915344},{"id":"https://openalex.org/C149364088","wikidata":"https://www.wikidata.org/wiki/Q185917","display_name":"Translation (biology)","level":4,"score":0.29030001163482666},{"id":"https://openalex.org/C548217200","wikidata":"https://www.wikidata.org/wiki/Q251","display_name":"Java","level":2,"score":0.2851000130176544},{"id":"https://openalex.org/C155092808","wikidata":"https://www.wikidata.org/wiki/Q182557","display_name":"Computational linguistics","level":2,"score":0.2840000092983246},{"id":"https://openalex.org/C2776321320","wikidata":"https://www.wikidata.org/wiki/Q857525","display_name":"Annotation","level":2,"score":0.2808000147342682},{"id":"https://openalex.org/C2777200299","wikidata":"https://www.wikidata.org/wiki/Q52943","display_name":"Conversation","level":2,"score":0.27410000562667847},{"id":"https://openalex.org/C2780148112","wikidata":"https://www.wikidata.org/wiki/Q1432581","display_name":"Proxy (statistics)","level":2,"score":0.26030001044273376},{"id":"https://openalex.org/C622187","wikidata":"https://www.wikidata.org/wiki/Q3500773","display_name":"BLEU","level":3,"score":0.26019999384880066},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.25769999623298645},{"id":"https://openalex.org/C159254197","wikidata":"https://www.wikidata.org/wiki/Q1144915","display_name":"Lexicographical order","level":2,"score":0.2522999942302704}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.07629","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.07629","pdf_url":"https://arxiv.org/pdf/2608.07629","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.07629","doi":"https://doi.org/10.48550/arxiv.2608.07629","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.07629","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.07629","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.07629","pdf_url":"https://arxiv.org/pdf/2608.07629","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7202212165.pdf","grobid_xml":"https://content.openalex.org/works/W7202212165.grobid-xml"},"referenced_works_count":1,"referenced_works":["https://openalex.org/W4285077564"],"related_works":["https://openalex.org/W4390659578","https://openalex.org/W3141231691","https://openalex.org/W7134042189","https://openalex.org/W3014432386","https://openalex.org/W3093816530","https://openalex.org/W3087621941","https://openalex.org/W2952614664","https://openalex.org/W2772120246","https://openalex.org/W2972840215","https://openalex.org/W2985301125"],"abstract_inverted_index":{"Multilingual":[0],"neural":[1],"machine":[2],"translation":[3,76,170],"models":[4,25],"such":[5],"as":[6],"NLLB-200":[7,96,109,120,144],"cover":[8],"200":[9],"languages":[10,19,64],"but":[11],"leave":[12],"thousands":[13],"unsupported,":[14],"including":[15],"most":[16],"Grassfields":[17],"Bantu":[18,169],"of":[20,58,81],"Cameroon.":[21],"When":[22],"fine-tuning":[23],"these":[24],"for":[26,42,175],"an":[27,47,188],"unseen":[28],"language,":[29],"practitioners":[30],"must":[31],"choose":[32],"a":[33,52,78,90,101,112],"proxy":[34,114,177],"language":[35,53],"token,":[36],"yet":[37],"no":[38],"principled":[39],"method":[40],"exists":[41],"this":[43,72],"selection.":[44,178],"We":[45,70,93,129,191],"implemented":[46],"embedding":[48,124],"initialization":[49,125,134],"strategy":[50],"where":[51],"token":[54,115],"is":[55,162],"the":[56,67,132,139,148,163,173],"average":[57],"embeddings":[59],"from":[60,85,104],"multiple":[61],"typologically":[62],"related":[63],"already":[65],"in":[66,166],"mod":[68],"el.":[69],"evaluate":[71],"approach":[73],"on":[74],"Limbum-to-English":[75],"using":[77],"parallel":[79],"corpus":[80],"8,837":[82],"sentence":[83],"pairs":[84],"New":[86],"Testament":[87],"text":[88],"and":[89,119,195],"bilingual":[91],"dictionary.":[92],"compare":[94],"models:":[95],"zero-shot":[97],"(chrF2++":[98,106,116,126],"=":[99,107,117,127],"12.5),":[100],"Transformer":[102],"trained":[103],"scratch":[105],"14.5),":[108],"fine-tuned":[110],"with":[111,121],"Swahili":[113],"47.3),":[118],"our":[122,193],"averaged":[123],"46.7).":[128],"find":[130],"that":[131,159],"multi-language":[133],"achieves":[135],"performance":[136],"comparable":[137],"to":[138,183,198],"best":[140],"single-language":[141],"proxy.":[142],"Both":[143],"variants":[145],"improve":[146],"over":[147,152],"from-scratch":[149],"baseline":[150],"by":[151],"32":[153],"chrF2++":[154],"points.":[155],"These":[156],"results":[157],"show":[158],"multilingual":[160],"transfer":[161],"dominant":[164],"factor":[165],"extremely":[167],"low-resource":[168],"while":[171],"eliminating":[172],"need":[174],"heuristic":[176],"However,":[179],"all":[180],"systems":[181],"fail":[182],"preserve":[184],"tonal":[185],"diacritics,":[186],"highlighting":[187],"open":[189],"challenge.":[190],"make":[192],"dataset":[194],"code":[196],"available":[197],"support":[199],"further":[200],"research.":[201]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-08-12T00:00:00"}