What do post-editors correct? A fine-grained analysis of SMT and NMT errors

The recent improvements in neural MT (NMT) have driven a shift from statistical MT (SMT) to NMT. However, to assess the usefulness of MT models for post-editing (PE) and have a detailed insight of the output they produce, we need to analyse the most frequent errors and how they affect the task. We p...

Descripción completa

Detalles Bibliográficos
Autores: Alvarez Vidal, Sergi, Oliver, Antoni, Badia, Toni
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2021
País:España
Institución:Universitat Oberta de Catalunya (UOC)
Repositorio:O2, repositorio institucional de la UOC
OAI Identifier:oai:openaccess.uoc.edu:10609/150203
Acceso en línea:http://hdl.handle.net/10609/150203
https://doi.org/10.5565/rev/tradumatica.286
Access Level:acceso abierto
Palabra clave:machine translation
MT
NMT
post-editing
neuram machine translation
error taxonomy
traducció automàtica
taxonomia d'errors
traducció automàtica neuronal
postedició
TAN
TA
traducción automática
taxonomía de errores
posedición
traducción automática neuronal
Descripción
Sumario:The recent improvements in neural MT (NMT) have driven a shift from statistical MT (SMT) to NMT. However, to assess the usefulness of MT models for post-editing (PE) and have a detailed insight of the output they produce, we need to analyse the most frequent errors and how they affect the task. We present a pilot study of a fine-grained analysis of MT errors based on post-editors corrections for an English to Spanish medical text translated with SMT and NMT. We use the MQM taxonomy to compare the two MT models and have a categorized classification of the errors produced. Even though results show a great variation among posteditors’ corrections, for this language combination fewer errors are corrected by post-editors in the NMT output. NMT also produces fewer accuracy errors and errors that are less critical.