Browse DORAS
Browse Theses
Search
Latest Additions
Creative Commons License
Except where otherwise noted, content on this site is licensed for use under a:

Improving the objective function in minimum error rate training

He, Yifan and Way, Andy (2009) Improving the objective function in minimum error rate training. In: MT Summit XII - The twelfth Machine Translation Summit, 26-30 August 2009, Ottawa, Canada.

Full text available as:

[img]PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
81Kb

Abstract

In Minimum Error Rate Training (MERT), the parameters of an SMT system are tuned on a certain evaluation metric to improve translation quality. In this paper, we present empirical results in which parameters tuned on one metric (e.g. BLEU) may not lead to optimal scores on the same metric. The score can be improved significantly by tuning on an entirely different metric (e.g. METEOR, by 0.82 BLEU points or 3.38% relative improvement on WMT08 English–French dataset). We analyse the impact of choice of objective function in MERT and further propose three combination strategies of different metrics to reduce the bias of a single metric, and obtain parameters that receive better scores (0.99 BLEU points or 4.08% relative improvement) on evaluation metrics than those tuned on the standalone metric itself.

Item Type:Conference or Workshop Item (Paper)
Event Type:Conference
Refereed:Yes
Uncontrolled Keywords:statistical machine translation; minimum error rate training;
Subjects:Computer Science > Machine translating
DCU Faculties and Centres:Research Initiatives and Centres > Centre for Next Generation Localisation (CNGL)
Research Initiatives and Centres > National Centre for Language Technology (NCLT)
DCU Faculties and Schools > Faculty of Engineering and Computing > School of Computing
Official URL:http://summitxii.amtaweb.org/
Use License:This item is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 3.0 License. View License
Funders:Science Foundation Ireland, SFI 07/CE/I1142
ID Code:15162
Deposited On:15 Feb 2010 12:04 by DORAS Administrator. Last Modified 27 Apr 2010 11:31

Download statistics

Archive Staff Only: edit this record