Training Machine Translation Quality Estimation Model Based on Pseudo Data

Huanqin Wu; Hongyang Zhang; Jingmei Li; Junguo Zhu; Muyun Yang; Sheng Li

doi:10.13209/j.0479-8023.2017.158

Training Machine Translation Quality Estimation Model Based on Pseudo Data

Huanqin Wu
, Hongyang Zhang
, Jingmei Li
, Junguo Zhu
, Muyun Yang^*
, Sheng Li

^*Corresponding author for this work

School of Computer Science and Technology, Harbin Institute of Technology
Harbin Engineering University

Research output: Contribution to journal › Article › peer-review

1 Scopus citations

Abstract

Aimed at providing efficient training data for neural translation quality estimation model, a pseudo data construction method for target dataset is proposed, the model is trained by two stage model training method: pre training based on pseudo data and fine tuning. The experimental design of different pseudo data scale is carried out. The experiment results show that the machine translation quality estimation model trained by the pseudo data has significantly improved in the correlation between the scores given by human and the artificial scores.

Original language	English
Pages (from-to)	279-285
Number of pages	7
Journal	Beijing Daxue Xuebao (Ziran Kexue Ban)/Acta Scientiarum Naturalium Universitatis Pekinensis
Volume	54
Issue number	2
DOIs	https://doi.org/10.13209/j.0479-8023.2017.158
State	Published - 20 Mar 2018
Externally published	Yes

Keywords

Deep learning
Machine translation quality estimation
Pseudo data

Access to Document

10.13209/j.0479-8023.2017.158

Cite this

@article{7ca81ddbd24344eb8e52ded0f7c1d715,

title = "Training Machine Translation Quality Estimation Model Based on Pseudo Data",

abstract = "Aimed at providing efficient training data for neural translation quality estimation model, a pseudo data construction method for target dataset is proposed, the model is trained by two stage model training method: pre training based on pseudo data and fine tuning. The experimental design of different pseudo data scale is carried out. The experiment results show that the machine translation quality estimation model trained by the pseudo data has significantly improved in the correlation between the scores given by human and the artificial scores.",

keywords = "Deep learning, Machine translation quality estimation, Pseudo data",

author = "Huanqin Wu and Hongyang Zhang and Jingmei Li and Junguo Zhu and Muyun Yang and Sheng Li",

note = "Publisher Copyright: {\textcopyright} 2018 Peking University.",

year = "2018",

month = mar,

day = "20",

doi = "10.13209/j.0479-8023.2017.158",

language = "英语",

volume = "54",

pages = "279--285",

journal = "Beijing Daxue Xuebao (Ziran Kexue Ban)/Acta Scientiarum Naturalium Universitatis Pekinensis",

issn = "0479-8023",

publisher = "Peking University",

number = "2",

}

TY - JOUR

T1 - Training Machine Translation Quality Estimation Model Based on Pseudo Data

AU - Wu, Huanqin

AU - Zhang, Hongyang

AU - Li, Jingmei

AU - Zhu, Junguo

AU - Yang, Muyun

AU - Li, Sheng

PY - 2018/3/20

Y1 - 2018/3/20

N2 - Aimed at providing efficient training data for neural translation quality estimation model, a pseudo data construction method for target dataset is proposed, the model is trained by two stage model training method: pre training based on pseudo data and fine tuning. The experimental design of different pseudo data scale is carried out. The experiment results show that the machine translation quality estimation model trained by the pseudo data has significantly improved in the correlation between the scores given by human and the artificial scores.

AB - Aimed at providing efficient training data for neural translation quality estimation model, a pseudo data construction method for target dataset is proposed, the model is trained by two stage model training method: pre training based on pseudo data and fine tuning. The experimental design of different pseudo data scale is carried out. The experiment results show that the machine translation quality estimation model trained by the pseudo data has significantly improved in the correlation between the scores given by human and the artificial scores.

KW - Deep learning

KW - Machine translation quality estimation

KW - Pseudo data

UR - https://www.scopus.com/pages/publications/85048012958

U2 - 10.13209/j.0479-8023.2017.158

DO - 10.13209/j.0479-8023.2017.158

M3 - 文章

AN - SCOPUS:85048012958

SN - 0479-8023

VL - 54

SP - 279

EP - 285

JO - Beijing Daxue Xuebao (Ziran Kexue Ban)/Acta Scientiarum Naturalium Universitatis Pekinensis

JF - Beijing Daxue Xuebao (Ziran Kexue Ban)/Acta Scientiarum Naturalium Universitatis Pekinensis

IS - 2

ER -

Training Machine Translation Quality Estimation Model Based on Pseudo Data

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this