Error analysis in automatic speech recognition and machine translation
Autor(a) principal: | |
---|---|
Data de Publicação: | 2021 |
Tipo de documento: | Dissertação |
Idioma: | eng |
Título da fonte: | Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
Texto Completo: | http://hdl.handle.net/10451/57155 |
Resumo: | Automatic speech recognition and machine translation are well-known terms in the translation world nowadays. Systems that carry out these processes are taking over the work of humans more and more. Reasons for this are the speed at which the tasks are performed and their costs. However, the quality of these systems is debatable. They are not yet capable of delivering the same performance as human transcribers or translators. The lack of creativity, the ability to interpret texts and the sense of language is often cited as the reason why the performance of machines is not yet at the level of human translation or transcribing work. Despite this, there are companies that use these machines in their production pipelines. Unbabel, an online translation platform powered by artificial intelligence, is one of these companies. Through a combination of human translators and machines, Unbabel tries to provide its customers with a translation of good quality. This internship report was written with the aim of gaining an overview of the performance of these systems and the errors they produce. Based on this work, we try to get a picture of possible error patterns produced by both systems. The present work consists of an extensive analysis of errors produced by automatic speech recognition and machine translation systems after automatically transcribing and translating 10 English videos into Dutch. Different videos were deliberately chosen to see if there were significant differences in the error patterns between videos. The generated data and results from this work, aims at providing possible ways to improve the quality of the services already mentioned. |
id |
RCAP_b6ed198a29955576c4c6f7e1beebfaa2 |
---|---|
oai_identifier_str |
oai:repositorio.ul.pt:10451/57155 |
network_acronym_str |
RCAP |
network_name_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository_id_str |
7160 |
spelling |
Error analysis in automatic speech recognition and machine translationDomínio/Área Científica::Humanidades::Línguas e LiteraturasAutomatic speech recognition and machine translation are well-known terms in the translation world nowadays. Systems that carry out these processes are taking over the work of humans more and more. Reasons for this are the speed at which the tasks are performed and their costs. However, the quality of these systems is debatable. They are not yet capable of delivering the same performance as human transcribers or translators. The lack of creativity, the ability to interpret texts and the sense of language is often cited as the reason why the performance of machines is not yet at the level of human translation or transcribing work. Despite this, there are companies that use these machines in their production pipelines. Unbabel, an online translation platform powered by artificial intelligence, is one of these companies. Through a combination of human translators and machines, Unbabel tries to provide its customers with a translation of good quality. This internship report was written with the aim of gaining an overview of the performance of these systems and the errors they produce. Based on this work, we try to get a picture of possible error patterns produced by both systems. The present work consists of an extensive analysis of errors produced by automatic speech recognition and machine translation systems after automatically transcribing and translating 10 English videos into Dutch. Different videos were deliberately chosen to see if there were significant differences in the error patterns between videos. The generated data and results from this work, aims at providing possible ways to improve the quality of the services already mentioned.O reconhecimento automático de fala e a tradução automática são termos conhecidos no mundo da tradução, hoje em dia. Os sistemas que realizam esses processos estão a assumir cada vez mais o trabalho dos humanos. As razões para isso são a velocidade com que as tarefas são realizadas e os seus custos. No entanto, a qualidade desses sistemas é discutível. As máquinas ainda não são capazes de ter o mesmo desempenho dos transcritores ou tradutores humanos. A falta de criatividade, de capacidade de interpretar textos e de sensibilidade linguística são motivos frequentemente usados para justificar o facto de as máquinas ainda não estarem suficientemente desenvolvidas para terem um desempenho comparável com o trabalho de tradução ou transcrição humano. Mesmo assim, existem empresas que fazem uso dessas máquinas. A Unbabel, uma plataforma de tradução online baseada em inteligência artificial, é uma dessas empresas. Através de uma combinação de tradutores humanos e de máquinas, a Unbabel procura oferecer aos seus clientes traduções de boa qualidade. O presente relatório de estágio foi feito com o intuito de obter uma visão geral do desempenho desses sistemas e das falhas que cometem, propondo delinear uma imagem dos possíveis padrões de erro existentes nos mesmos. Para tal, fez-se uma análise extensa das falhas que os sistemas de reconhecimento automático de fala e de tradução automática cometeram, após a transcrição e a tradução automática de 10 vídeos. Foram deliberadamente escolhidos registos videográficos diversos, de modo a verificar possíveis diferenças nos padrões de erro. Através dos dados gerados e dos resultados obtidos, propõe-se encontrar uma forma de melhorar a qualidade dos serviços já mencionados.Mendes, Sara Gonçalves Pedro ParenteSanchez, MarinaRepositório da Universidade de LisboaLoomans, Nicolaas Dirk Petrus2023-04-18T09:19:25Z2022-02-252021-09-132022-02-25T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfapplication/pdfhttp://hdl.handle.net/10451/57155TID:203098390enginfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2023-11-08T17:05:11Zoai:repositorio.ul.pt:10451/57155Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-19T22:07:33.871970Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse |
dc.title.none.fl_str_mv |
Error analysis in automatic speech recognition and machine translation |
title |
Error analysis in automatic speech recognition and machine translation |
spellingShingle |
Error analysis in automatic speech recognition and machine translation Loomans, Nicolaas Dirk Petrus Domínio/Área Científica::Humanidades::Línguas e Literaturas |
title_short |
Error analysis in automatic speech recognition and machine translation |
title_full |
Error analysis in automatic speech recognition and machine translation |
title_fullStr |
Error analysis in automatic speech recognition and machine translation |
title_full_unstemmed |
Error analysis in automatic speech recognition and machine translation |
title_sort |
Error analysis in automatic speech recognition and machine translation |
author |
Loomans, Nicolaas Dirk Petrus |
author_facet |
Loomans, Nicolaas Dirk Petrus |
author_role |
author |
dc.contributor.none.fl_str_mv |
Mendes, Sara Gonçalves Pedro Parente Sanchez, Marina Repositório da Universidade de Lisboa |
dc.contributor.author.fl_str_mv |
Loomans, Nicolaas Dirk Petrus |
dc.subject.por.fl_str_mv |
Domínio/Área Científica::Humanidades::Línguas e Literaturas |
topic |
Domínio/Área Científica::Humanidades::Línguas e Literaturas |
description |
Automatic speech recognition and machine translation are well-known terms in the translation world nowadays. Systems that carry out these processes are taking over the work of humans more and more. Reasons for this are the speed at which the tasks are performed and their costs. However, the quality of these systems is debatable. They are not yet capable of delivering the same performance as human transcribers or translators. The lack of creativity, the ability to interpret texts and the sense of language is often cited as the reason why the performance of machines is not yet at the level of human translation or transcribing work. Despite this, there are companies that use these machines in their production pipelines. Unbabel, an online translation platform powered by artificial intelligence, is one of these companies. Through a combination of human translators and machines, Unbabel tries to provide its customers with a translation of good quality. This internship report was written with the aim of gaining an overview of the performance of these systems and the errors they produce. Based on this work, we try to get a picture of possible error patterns produced by both systems. The present work consists of an extensive analysis of errors produced by automatic speech recognition and machine translation systems after automatically transcribing and translating 10 English videos into Dutch. Different videos were deliberately chosen to see if there were significant differences in the error patterns between videos. The generated data and results from this work, aims at providing possible ways to improve the quality of the services already mentioned. |
publishDate |
2021 |
dc.date.none.fl_str_mv |
2021-09-13 2022-02-25 2022-02-25T00:00:00Z 2023-04-18T09:19:25Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/masterThesis |
format |
masterThesis |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/10451/57155 TID:203098390 |
url |
http://hdl.handle.net/10451/57155 |
identifier_str_mv |
TID:203098390 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf application/pdf |
dc.source.none.fl_str_mv |
reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação instacron:RCAAP |
instname_str |
Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
instacron_str |
RCAAP |
institution |
RCAAP |
reponame_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
collection |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository.name.fl_str_mv |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
repository.mail.fl_str_mv |
|
_version_ |
1799134629147443200 |