LIUM’s SMT Machine Translation Systems for WMT 2011

Proceedings of the 6th Workshop on Statistical Machine Translation, pages 464–469, Edinburgh, Scotland, UK, July 30–31, 2011. c⃝2011 Association for Computational Linguistics LIUM’s SMT Machine Translation Systems for WMT 2011 Holger Schwenk, Patrik Lambert, Lo¨ıc Barrault, Christophe Servan, Haithem Afli, Sadaf Abdul-Rauf and Kashif Shah LIUM, University of Le Mans 72085 Le Mans cedex 9, FRANCE FirstName.LastName@lium.univ-lemans.fr Abstract This paper describes the development of French–English and English–French statisti- cal machine translation systems for the 2011 WMT shared task evaluation. Our main sys- tems were standard phrase-based statistical systems based on the Moses decoder, trained on the provided data only, but we also per- formed initial experiments with hierarchical systems. Additional, new features this year in- clude improved translation model adaptation using monolingual data, a continuous space language model and the treatment of unknown words. 1 Introduction This paper describes the statistical machine trans- lation systems developed by the Computer Science laboratory at the University of Le Mans (LIUM) for the 2011 WMT shared task evaluation. We only considered the translation between French and En- glish (in both directions). The main differences with respect to previous year’s system (Lambert et al., 2010) are as follows: use of more training data as provided by the organizers, improved translation model adaptation by unsupervised training, a con- tinuous space language model for the translation into French, some attempts to automatically induce translations of unknown words and first experiments with hierarchical systems. These different points are described in the rest of the paper, together with a summary of the experimental results showing the impact of each component. 2 Resources Used The following sections describe how the resources provided or allowed in the shared task were used to train the translation and language models of the sys- tem. 2.1 Bilingual data Our system was developed in two stages. First, a baseline system was built to generate automatic translations of some of the monolingual data avail- able. These automatic translations were then used directly with the source texts to create additional bi- texts. In a second stage, these additional bilingual data were incorporated into the system (see Sec- tion 5 and Tables 4 and 5). The latest version of the News-Commentary (NC) corpus and of the Europarl (Eparl) corpus (version 6) were used. We also took as training data a sub- set of the French–English Gigaword (109) corpus. We applied the same filters as last year to select this subset. The first one is a lexical filter based on the IBM model 1 cost (Brown et al., 1993) of each side of a sentence pair given the other side, normalised with respect to both sentence lengths. This filter was trained on a corpus composed of Eparl, NC, and UN data. The other filter is an n-gram language model (LM) cost of the target sentence (see Section 3), nor- malised with respect to its length. This filter was trained with all monolingual resources available ex- cept the 109 data. We generated two subsets, both by selecting sentence pairs …

Content truncated to 500 words