论文信息 - Evaluation Methodology and Results for English-to-Arabic MT

Evaluation Methodology and Results for English-to-Arabic MT

This paper describes the evaluation campaign of the MEDAR project for English-to-Arabic (EnAr) MT systems. The campaign aimed at establishing some basic facts about the state of the art for MT on EnAr, collecting enough data to better train and tune systems and assessing the improvements made. The paper details the data used and their formats, the evaluation methodology and the results obtained by the systems. We conclude by giving some recommendations on MT evaluation for EnAr direction in terms of technology and resources.

Khalid Choukri | Olivier Hamon

[1] Andreas Stolcke,et al. SRILM - an extensible language modeling toolkit , 2002, INTERSPEECH.

[3] Philipp Koehn,et al. Findings of the 2010 Joint Workshop on Statistical Machine Translation and Metrics for Machine Translation , 2010, WMT@ACL.

[4] Nizar Habash,et al. Four Techniques for Online Handling of Out-of-Vocabulary Words in Arabic-English Statistical Machine Translation , 2008, ACL.

[5] Mauro Cettolo,et al. IRSTLM: an open source toolkit for handling large scale language models , 2008, INTERSPEECH.

[6] France,et al. Diagnosing Human Judgments in MT Evaluation : an Example based on the Spanish Language , 2008 .

[7] Philipp Koehn,et al. Moses: Open Source Toolkit for Statistical Machine Translation , 2007, ACL.

[8] Nizar Habash,et al. Techniques for Arabic morphological detokenization and orthographic denormalization , 2010 .

[9] Chafic Mokbel,et al. On the use of morphological constraints in n-gram statistical language model , 2005, INTERSPEECH.

[10] Hermann Ney,et al. A Systematic Comparison of Various Statistical Alignment Models , 2003, CL.

[11] Salim Roukos,et al. Bleu: a Method for Automatic Evaluation of Machine Translation , 2002, ACL.