A Discriminative Model for Tree-to-Tree Translation

Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing (EMNLP 2006), pages 232–241, Sydney, July 2006. c⃝2006 Association for Computational Linguistics A Discriminative Model for Tree-to-Tree Translation Brooke Cowan MIT CSAIL brooke@csail.mit.edu Ivona Ku˘cerov´a MIT Linguistics Department kucerova@mit.edu Michael Collins MIT CSAIL mcollins@csail.mit.edu Abstract This paper proposes a statistical, tree- to-tree model for producing translations. Two main contributions are as follows: (1) a method for the extraction of syn- tactic structures with alignment informa- tion from a parallel corpus of translations, and (2) use of a discriminative, feature- based model for prediction of these target- language syntactic structures—which we call aligned extended projections, or AEPs. An evaluation of the method on translation from German to English shows similar performance to the phrase-based model of Koehn et al. (2003). 1 Introduction Phrase-based approaches (Och and Ney, 2004) to statistical machine translation (SMT) have re- cently achieved impressive results, leading to sig- nificant improvements in accuracy over the origi- nal IBM models (Brown et al., 1993). However, phrase-based models lack a direct representation of syntactic information in the source or target lan- guages; this has prompted several researchers to consider various approaches that make use of syn- tactic information. This paper describes a framework for tree-to- tree based statistical translation. Our goal is to learn a model that maps parse trees in the source language to parse trees in the target language. The model is learned from a corpus of transla- tion pairs, where each sentence in the source or target language has an associated parse tree. We see two major benefits of tree-to-tree based trans- lation. First, it is possible to explicitly model the syntax of the target language, thereby improving grammaticality. Second, we can build a detailed model of the correspondence between the source and target parse trees, with the aim of constructing translations that preserve the meaning of source language sentences. Our translation framework involves a process where the target-language parse tree is broken down into a sequence of clauses, and each clause is then translated separately. A central concept we introduce in the translation of clauses is that of an aligned extended projection (AEP). AEPs are de- rived from the concept of an extended projection in lexicalized tree adjoining grammars (LTAG) (Frank, 2002), with the addition of alignment in- formation that is based on work in synchronous LTAG (Shieber and Schabes, 1990). A key con- tribution of this paper is a method for learning to map German clauses to AEPs using a feature- based model with a perceptron learning algorithm. We performed experiments on translation from German to English on the Europarl data set. Eval- uation in terms of both BLEU scores and human judgments shows that our system performs sim- ilarly to the phrase-based model of Koehn et al. (2003). 1.1 A Sketch of the Approach This section provides an overview of the transla- tion process. We will use the German sentence wir wissen daß das haupthemmnis der vorhersehbare widerstand der hersteller war as …

Content truncated to 500 words