{"id":"https://openalex.org/W7117251697","doi":"https://doi.org/10.1109/access.2025.3648480","title":"Enhancing Low-Resource Indian Language Machine Translation Using Large Language Models With Preference Optimization and Hypergeometric-Gamma Reward","display_name":"Enhancing Low-Resource Indian Language Machine Translation Using Large Language Models With Preference Optimization and Hypergeometric-Gamma Reward","publication_year":2025,"publication_date":"2025-12-25","ids":{"openalex":"https://openalex.org/W7117251697","doi":"https://doi.org/10.1109/access.2025.3648480"},"language":null,"primary_location":{"id":"doi:10.1109/access.2025.3648480","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2025.3648480","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1109/access.2025.3648480","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5113234656","display_name":"Aarathi Rajagopalan Nair","orcid":null},"institutions":[{"id":"https://openalex.org/I81556334","display_name":"Amrita Vishwa Vidyapeetham","ror":"https://ror.org/03am10p12","country_code":"IN","type":"education","lineage":["https://openalex.org/I81556334"]}],"countries":["IN"],"is_corresponding":true,"raw_author_name":"Aarathi Rajagopalan Nair","raw_affiliation_strings":["Department of Computer Science and Engineering, Amrita School of Computing, Amrita Vishwa Vidyapeetham, Bengaluru, India"],"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, Amrita School of Computing, Amrita Vishwa Vidyapeetham, Bengaluru, India","institution_ids":["https://openalex.org/I81556334"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022692187","display_name":"Deepa Gupta","orcid":null},"institutions":[{"id":"https://openalex.org/I81556334","display_name":"Amrita Vishwa Vidyapeetham","ror":"https://ror.org/03am10p12","country_code":"IN","type":"education","lineage":["https://openalex.org/I81556334"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Deepa Gupta","raw_affiliation_strings":["Department of Computer Science and Engineering, Amrita School of Computing, Amrita Vishwa Vidyapeetham, Bengaluru, India"],"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, Amrita School of Computing, Amrita Vishwa Vidyapeetham, Bengaluru, India","institution_ids":["https://openalex.org/I81556334"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072503037","display_name":"Biswajit Paul","orcid":"https://orcid.org/0000-0002-4077-5789"},"institutions":[{"id":"https://openalex.org/I4210121405","display_name":"Centre for Artificial Intelligence and Robotics","ror":"https://ror.org/01xnbq218","country_code":"IN","type":"facility","lineage":["https://openalex.org/I1340206300","https://openalex.org/I4210121405","https://openalex.org/I4210150591"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Biswajit Paul","raw_affiliation_strings":["Centre for Artificial Intelligence and Robotics, Defence Research and Development Organisation, Bengaluru, India"],"affiliations":[{"raw_affiliation_string":"Centre for Artificial Intelligence and Robotics, Defence Research and Development Organisation, Bengaluru, India","institution_ids":["https://openalex.org/I4210121405"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5121269778","display_name":"J Sivabhavani","orcid":null},"institutions":[{"id":"https://openalex.org/I4210121405","display_name":"Centre for Artificial Intelligence and Robotics","ror":"https://ror.org/01xnbq218","country_code":"IN","type":"facility","lineage":["https://openalex.org/I1340206300","https://openalex.org/I4210121405","https://openalex.org/I4210150591"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"J. Siva Bhavani","raw_affiliation_strings":["Centre for Artificial Intelligence and Robotics, Defence Research and Development Organisation, Bengaluru, India"],"affiliations":[{"raw_affiliation_string":"Centre for Artificial Intelligence and Robotics, Defence Research and Development Organisation, Bengaluru, India","institution_ids":["https://openalex.org/I4210121405"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5113234656"],"corresponding_institution_ids":["https://openalex.org/I81556334"],"apc_list":{"value":1850,"currency":"USD","value_usd":1850},"apc_paid":{"value":1850,"currency":"USD","value_usd":1850},"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.83968432,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"14","issue":null,"first_page":"1641","last_page":"1665"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.22360000014305115,"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":0.22360000014305115,"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":0.0478999987244606,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.04520000144839287,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/machine-translation","display_name":"Machine translation","score":0.7208999991416931},{"id":"https://openalex.org/keywords/pairwise-comparison","display_name":"Pairwise comparison","score":0.5803999900817871},{"id":"https://openalex.org/keywords/similarity","display_name":"Similarity (geometry)","score":0.451200008392334},{"id":"https://openalex.org/keywords/preference","display_name":"Preference","score":0.4262000024318695},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.40049999952316284},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.38589999079704285},{"id":"https://openalex.org/keywords/cosine-similarity","display_name":"Cosine similarity","score":0.34380000829696655},{"id":"https://openalex.org/keywords/translation","display_name":"Translation (biology)","score":0.3377000093460083}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8046000003814697},{"id":"https://openalex.org/C203005215","wikidata":"https://www.wikidata.org/wiki/Q79798","display_name":"Machine translation","level":2,"score":0.7208999991416931},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6909000277519226},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5889000296592712},{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.5803999900817871},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5027999877929688},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.451200008392334},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.4262000024318695},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.40049999952316284},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.38589999079704285},{"id":"https://openalex.org/C2780762811","wikidata":"https://www.wikidata.org/wiki/Q1784941","display_name":"Cosine similarity","level":3,"score":0.34380000829696655},{"id":"https://openalex.org/C149364088","wikidata":"https://www.wikidata.org/wiki/Q185917","display_name":"Translation (biology)","level":4,"score":0.3377000093460083},{"id":"https://openalex.org/C2986862884","wikidata":"https://www.wikidata.org/wiki/Q7553","display_name":"Language translation","level":3,"score":0.3353999853134155},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.33059999346733093},{"id":"https://openalex.org/C135784402","wikidata":"https://www.wikidata.org/wiki/Q6958279","display_name":"Evaluation of machine translation","level":5,"score":0.32670000195503235},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.3124000132083893},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.30469998717308044},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.2946000099182129},{"id":"https://openalex.org/C16910744","wikidata":"https://www.wikidata.org/wiki/Q7705759","display_name":"Test data","level":2,"score":0.2939999997615814},{"id":"https://openalex.org/C130318100","wikidata":"https://www.wikidata.org/wiki/Q2268914","display_name":"Semantic similarity","level":2,"score":0.2924000024795532},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.29089999198913574},{"id":"https://openalex.org/C8642999","wikidata":"https://www.wikidata.org/wiki/Q4171168","display_name":"Hyperparameter","level":2,"score":0.288100004196167},{"id":"https://openalex.org/C176671685","wikidata":"https://www.wikidata.org/wiki/Q730600","display_name":"Hypergeometric distribution","level":2,"score":0.27570000290870667},{"id":"https://openalex.org/C2779439875","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Natural language understanding","level":3,"score":0.26660001277923584}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/access.2025.3648480","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2025.3648480","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1109/access.2025.3648480","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2025.3648480","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":49,"referenced_works":["https://openalex.org/W2101105183","https://openalex.org/W2118593414","https://openalex.org/W2250342921","https://openalex.org/W2788256404","https://openalex.org/W2798935874","https://openalex.org/W2970641574","https://openalex.org/W3035922389","https://openalex.org/W3090350559","https://openalex.org/W3124827513","https://openalex.org/W3169483174","https://openalex.org/W3170253133","https://openalex.org/W3174716116","https://openalex.org/W4210590928","https://openalex.org/W4220806038","https://openalex.org/W4224902833","https://openalex.org/W4285803230","https://openalex.org/W4312505683","https://openalex.org/W4323023270","https://openalex.org/W4323520828","https://openalex.org/W4382202554","https://openalex.org/W4382513180","https://openalex.org/W4385565879","https://openalex.org/W4388691863","https://openalex.org/W4389117869","https://openalex.org/W4389518624","https://openalex.org/W4389519421","https://openalex.org/W4389520741","https://openalex.org/W4389524378","https://openalex.org/W4392623252","https://openalex.org/W4401042957","https://openalex.org/W4401042977","https://openalex.org/W4401042993","https://openalex.org/W4401043452","https://openalex.org/W4401467830","https://openalex.org/W4402264578","https://openalex.org/W4402670451","https://openalex.org/W4402683150","https://openalex.org/W4402684005","https://openalex.org/W4403422562","https://openalex.org/W4403826725","https://openalex.org/W4404033281","https://openalex.org/W4404782829","https://openalex.org/W4404783577","https://openalex.org/W4405180297","https://openalex.org/W4406911809","https://openalex.org/W4409965723","https://openalex.org/W4411113160","https://openalex.org/W4411119805","https://openalex.org/W4413471766"],"related_works":[],"abstract_inverted_index":{"Machine":[0],"translation":[1,51,112,281],"has":[2],"advanced":[3],"considerably":[4],"in":[5,34,45,91,218],"recent":[6],"years":[7],"yet":[8],"producing":[9],"accurate":[10],"and":[11,83,127,137,147,154,160,162,170,183,208,222,250,273,285],"natural":[12],"translations":[13,123],"remains":[14],"challenging":[15],"for":[16,149,156,164,283],"languages":[17],"with":[18,77],"limited":[19],"digital":[20],"resources.":[21],"The":[22,118,139,197],"scarcity":[23],"of":[24,32,60,145,186,260,268],"high-quality":[25],"parallel":[26],"data":[27,233],"particularly":[28],"hampers":[29],"the":[30,58,61,111,179,191,203,215,235,239,258,265],"performance":[31,246],"models":[33],"low-resource":[35,284],"languages.":[36,287],"This":[37],"study":[38],"introduces":[39],"a":[40,87],"preference-based":[41],"learning":[42,47,272],"approach":[43,119],"grounded":[44],"reinforcement":[46,269],"principles":[48],"to":[49,151,158,166,247,279],"improve":[50],"quality":[52,282],"by":[53],"incorporating":[54],"human":[55,78,262],"feedback":[56,263],"during":[57],"fine-tuning":[59],"multilingual":[62],"mT5-Large":[63],"model.":[64],"Two":[65],"optimization":[66],"techniques":[67],"are":[68,108,194],"employed:":[69],"Direct":[70,205],"Preference":[71,206],"Optimization,":[72],"which":[73],"aligns":[74],"model":[75,113,141,212],"outputs":[76],"preferences":[79],"through":[80,114,178,264],"pairwise":[81],"comparison,":[82],"Hypergeometric":[84,209],"Gamma":[85,210],"Reward,":[86],"novel":[88],"method":[89,241],"proposed":[90,240],"this":[92],"work":[93],"that":[94,190,202],"generates":[95],"reward":[96],"signals":[97],"based":[98],"on":[99,122,168],"semantic":[100,223],"similarity":[101,224],"scores":[102,144],"computed":[103],"using":[104,131,229],"Sentence-BERT.":[105],"These":[106,255],"rewards":[107],"integrated":[109],"into":[110,129],"reward-weighted":[115],"loss":[116],"optimization.":[117],"is":[120],"evaluated":[121],"from":[124],"Dogri,":[125],"Kashmiri,":[126],"Konkani":[128,165],"English":[130,167],"two":[132],"standard":[133],"test":[134],"sets,":[135],"IN22-Gen":[136,169],"IN22-Conv.":[138],"enhanced":[140],"achieves":[142,214],"BLEU":[143],"49.81":[146],"57.21":[148],"Dogri":[150],"English,":[152,159],"43.57":[153],"46.57":[155],"Kashmiri":[157],"38.75":[161],"34.56":[163],"IN22-Conv,":[171],"respectively.":[172],"All":[173],"reported":[174],"improvements":[175,217],"were":[176],"validated":[177],"Approximate":[180],"Randomization":[181],"Test":[182],"repeated-measures":[184],"Analysis":[185],"Variance,":[187],"both":[188,219],"confirming":[189],"observed":[192],"gains":[193],"statistically":[195],"significant.":[196],"statistical":[198],"analysis":[199],"further":[200],"demonstrates":[201],"combined":[204,266],"Optimization":[207],"Reward":[211],"consistently":[213],"strongest":[216],"syntactic":[220],"accuracy":[221],"across":[225],"all":[226],"datasets.":[227],"Despite":[228],"significantly":[230],"less":[231],"training":[232],"than":[234],"state-of-the-art":[236],"IndicTrans2":[237],"model,":[238],"delivers":[242],"superior":[243],"or":[244],"comparable":[245],"transformer-based":[248],"architectures":[249],"other":[251],"large":[252],"language":[253],"models.":[254],"results":[256],"demonstrate":[257],"effectiveness":[259],"integrating":[261],"use":[267],"learning-inspired":[270],"preference":[271],"reward-based":[274],"optimization,":[275],"highlighting":[276],"their":[277],"potential":[278],"enhance":[280],"underrepresented":[286]},"counts_by_year":[],"updated_date":"2026-01-08T20:05:33.558190","created_date":"2025-12-25T00:00:00"}