{"id":"https://openalex.org/W4404400385","doi":"https://doi.org/10.1162/tacl.a.646","title":"Fine-Grained Reward Optimization for Machine Translation using Error Severity Mappings","display_name":"Fine-Grained Reward Optimization for Machine Translation using Error Severity Mappings","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W4404400385","doi":"https://doi.org/10.1162/tacl.a.646"},"language":"en","primary_location":{"id":"doi:10.1162/tacl.a.646","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl.a.646","pdf_url":null,"source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},"type":"article","indexed_in":["arxiv","crossref","datacite","doaj"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://doi.org/10.1162/tacl.a.646","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5113048516","display_name":"Miguel Moura Ramos","orcid":null},"institutions":[{"id":"https://openalex.org/I203847022","display_name":"Instituto Polit\u00e9cnico de Lisboa","ror":"https://ror.org/04ea70f07","country_code":"PT","type":"education","lineage":["https://openalex.org/I203847022"]},{"id":"https://openalex.org/I4210120471","display_name":"Instituto de Telecomunica\u00e7\u00f5es","ror":"https://ror.org/02ht4fk33","country_code":"PT","type":"nonprofit","lineage":["https://openalex.org/I4210120471"]}],"countries":["PT"],"is_corresponding":false,"raw_author_name":"Miguel Moura Ramos","raw_affiliation_strings":["Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal. miguel.moura.ramos@tecnico.ulisboa.pt","Instituto de Telecomunica\u00e7\u00f5es, Portugal"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal. miguel.moura.ramos@tecnico.ulisboa.pt","institution_ids":["https://openalex.org/I203847022"]},{"raw_affiliation_string":"Instituto de Telecomunica\u00e7\u00f5es, Portugal","institution_ids":["https://openalex.org/I4210120471"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110177880","display_name":"Tom\u00e1s Almeida","orcid":null},"institutions":[{"id":"https://openalex.org/I203847022","display_name":"Instituto Polit\u00e9cnico de Lisboa","ror":"https://ror.org/04ea70f07","country_code":"PT","type":"education","lineage":["https://openalex.org/I203847022"]}],"countries":["PT"],"is_corresponding":false,"raw_author_name":"Tom\u00e1s Almeida","raw_affiliation_strings":["Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal","institution_ids":["https://openalex.org/I203847022"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114656596","display_name":"Daniel Vareta","orcid":null},"institutions":[{"id":"https://openalex.org/I203847022","display_name":"Instituto Polit\u00e9cnico de Lisboa","ror":"https://ror.org/04ea70f07","country_code":"PT","type":"education","lineage":["https://openalex.org/I203847022"]}],"countries":["PT"],"is_corresponding":false,"raw_author_name":"Daniel Vareta","raw_affiliation_strings":["Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal","institution_ids":["https://openalex.org/I203847022"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006592427","display_name":"Filipe Azevedo","orcid":"https://orcid.org/0000-0003-0841-362X"},"institutions":[{"id":"https://openalex.org/I203847022","display_name":"Instituto Polit\u00e9cnico de Lisboa","ror":"https://ror.org/04ea70f07","country_code":"PT","type":"education","lineage":["https://openalex.org/I203847022"]},{"id":"https://openalex.org/I4210120471","display_name":"Instituto de Telecomunica\u00e7\u00f5es","ror":"https://ror.org/02ht4fk33","country_code":"PT","type":"nonprofit","lineage":["https://openalex.org/I4210120471"]}],"countries":["PT"],"is_corresponding":false,"raw_author_name":"Filipe Azevedo","raw_affiliation_strings":["Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal","Instituto de Telecomunica\u00e7\u00f5es, Portugal"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal","institution_ids":["https://openalex.org/I203847022"]},{"raw_affiliation_string":"Instituto de Telecomunica\u00e7\u00f5es, Portugal","institution_ids":["https://openalex.org/I4210120471"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103140610","display_name":"Sweta Agrawal","orcid":"https://orcid.org/0000-0001-5692-0816"},"institutions":[{"id":"https://openalex.org/I4210120471","display_name":"Instituto de Telecomunica\u00e7\u00f5es","ror":"https://ror.org/02ht4fk33","country_code":"PT","type":"nonprofit","lineage":["https://openalex.org/I4210120471"]}],"countries":["PT"],"is_corresponding":false,"raw_author_name":"Sweta Agrawal","raw_affiliation_strings":["Instituto de Telecomunica\u00e7\u00f5es, Portugal"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Instituto de Telecomunica\u00e7\u00f5es, Portugal","institution_ids":["https://openalex.org/I4210120471"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5061762931","display_name":"Patrick Fernandes","orcid":null},"institutions":[{"id":"https://openalex.org/I203847022","display_name":"Instituto Polit\u00e9cnico de Lisboa","ror":"https://ror.org/04ea70f07","country_code":"PT","type":"education","lineage":["https://openalex.org/I203847022"]},{"id":"https://openalex.org/I4210120471","display_name":"Instituto de Telecomunica\u00e7\u00f5es","ror":"https://ror.org/02ht4fk33","country_code":"PT","type":"nonprofit","lineage":["https://openalex.org/I4210120471"]},{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["PT","US"],"is_corresponding":false,"raw_author_name":"Patrick Fernandes","raw_affiliation_strings":["Carnegie Mellon University, USA","Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal","Instituto de Telecomunica\u00e7\u00f5es, Portugal"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, USA","institution_ids":["https://openalex.org/I74973139"]},{"raw_affiliation_string":"Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal","institution_ids":["https://openalex.org/I203847022"]},{"raw_affiliation_string":"Instituto de Telecomunica\u00e7\u00f5es, Portugal","institution_ids":["https://openalex.org/I4210120471"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101751817","display_name":"Andr\u00e9 F. T. Martins","orcid":"https://orcid.org/0000-0001-6863-4681"},"institutions":[{"id":"https://openalex.org/I1304373396","display_name":"Transparency International","ror":"https://ror.org/048cscx90","country_code":"CZ","type":"nonprofit","lineage":["https://openalex.org/I1304373396"]},{"id":"https://openalex.org/I203847022","display_name":"Instituto Polit\u00e9cnico de Lisboa","ror":"https://ror.org/04ea70f07","country_code":"PT","type":"education","lineage":["https://openalex.org/I203847022"]},{"id":"https://openalex.org/I4210120471","display_name":"Instituto de Telecomunica\u00e7\u00f5es","ror":"https://ror.org/02ht4fk33","country_code":"PT","type":"nonprofit","lineage":["https://openalex.org/I4210120471"]}],"countries":["CZ","PT"],"is_corresponding":false,"raw_author_name":"Andr\u00e9 F. T. Martins","raw_affiliation_strings":["Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal","Instituto de Telecomunica\u00e7\u00f5es, Portugal","TransPerfect, Portugal"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Instituto Superior T\u00e9cnico, Universidade de Lisboa (ELLIS Unit Lisbon), Portugal","institution_ids":["https://openalex.org/I203847022"]},{"raw_affiliation_string":"Instituto de Telecomunica\u00e7\u00f5es, Portugal","institution_ids":["https://openalex.org/I4210120471"]},{"raw_affiliation_string":"TransPerfect, Portugal","institution_ids":["https://openalex.org/I1304373396"]}]}],"institutions":[],"countries_distinct_count":3,"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":0.00524734,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"14","issue":null,"first_page":"733","last_page":"754"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.7958999872207642,"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":0.7958999872207642,"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/translation","display_name":"Translation (biology)","score":0.6562078595161438},{"id":"https://openalex.org/keywords/machine-translation","display_name":"Machine translation","score":0.6341170072555542},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5926094651222229},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4535329341888428},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3328963816165924},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.3214406669139862},{"id":"https://openalex.org/keywords/biology","display_name":"Biology","score":0.11717802286148071}],"concepts":[{"id":"https://openalex.org/C149364088","wikidata":"https://www.wikidata.org/wiki/Q185917","display_name":"Translation (biology)","level":4,"score":0.6562078595161438},{"id":"https://openalex.org/C203005215","wikidata":"https://www.wikidata.org/wiki/Q79798","display_name":"Machine translation","level":2,"score":0.6341170072555542},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5926094651222229},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4535329341888428},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3328963816165924},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3214406669139862},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.11717802286148071},{"id":"https://openalex.org/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0},{"id":"https://openalex.org/C105580179","wikidata":"https://www.wikidata.org/wiki/Q188928","display_name":"Messenger RNA","level":3,"score":0.0},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1162/tacl.a.646","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl.a.646","pdf_url":null,"source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},{"id":"pmh:oai:arXiv.org:2411.05986","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.05986","pdf_url":"https://arxiv.org/pdf/2411.05986","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2411.05986","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2411.05986","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.1162/tacl.a.646","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl.a.646","pdf_url":null,"source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320334322","display_name":"HORIZON EUROPE Framework Programme","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2961085424","https://openalex.org/W4306674287","https://openalex.org/W3046775127","https://openalex.org/W2775554247","https://openalex.org/W4394896187","https://openalex.org/W3170094116","https://openalex.org/W4386462264","https://openalex.org/W2883671469","https://openalex.org/W3107602296","https://openalex.org/W2728761353"],"abstract_inverted_index":{"Abstract":[0],"Reinforcement":[1],"learning":[2,47],"(RL)":[3],"has":[4,35],"been":[5],"proven":[6],"to":[7,45,50,148],"be":[8],"an":[9],"effective":[10],"and":[11,105,112,151],"robust":[12],"method":[13],"for":[14,60],"training":[15,135,159,170],"neural":[16],"machine":[17,116],"translation":[18,30,107,117,129,140],"systems,":[19,118],"especially":[20],"when":[21],"paired":[22],"with":[23,79,109,136],"powerful":[24],"reward":[25,52,98,126,156],"models":[26],"that":[27,40,72,134],"accurately":[28],"assess":[29],"quality.":[31,130],"However,":[32],"most":[33],"research":[34],"focused":[36],"on":[37,103,128],"RL":[38,84],"methods":[39],"use":[41,88],"sentence-level":[42,123],"feedback,":[43],"leading":[44],"inefficient":[46],"signals":[48,127],"due":[49],"the":[51,61,120],"sparsity":[53],"problem\u2014the":[54],"model":[55],"receives":[56],"a":[57,69,90,163],"single":[58],"score":[59],"entire":[62],"sentence.":[63],"To":[64],"address":[65],"this,":[66],"we":[67,87],"propose":[68],"novel":[70],"approach":[71],"leverages":[73],"fine-grained,":[74],"token-level":[75,97,137,155],"quality":[76,92,141],"assessments":[77],"along":[78],"error":[80],"severity":[81],"levels":[82],"using":[83],"methods.":[85],"Specifically,":[86],"xCOMET,":[89],"state-of-the-art":[91],"estimation":[93],"system,":[94],"as":[95],"our":[96],"model.":[99],"We":[100],"conduct":[101],"experiments":[102],"small":[104],"large":[106,113],"datasets":[108],"standard":[110],"encoder-decoder":[111],"language":[114,143],"models-based":[115],"comparing":[119],"impact":[121],"of":[122],"versus":[124],"fine-grained":[125],"Our":[131],"results":[132],"show":[133],"rewards":[138,168],"improves":[139,158],"across":[142],"pairs":[144],"over":[145,169],"baselines":[146],"according":[147],"both":[149],"automatic":[150],"human":[152],"evaluation.":[153],"Furthermore,":[154],"optimization":[157],"stability,":[160],"evidenced":[161],"by":[162],"steady":[164],"increase":[165],"in":[166],"mean":[167],"epochs.":[171]},"counts_by_year":[],"updated_date":"2026-07-15T18:14:33.161393","created_date":"2025-10-10T00:00:00"}