Hierarchical Machine Translation With Discontinuous Phrases

Proceedings of the Tenth Workshop on Statistical Machine Translation, pages 228–238, Lisboa, Portugal, 17-18 September 2015. c⃝2015 Association for Computational Linguistics. Hierarchical Machine Translation With Discontinuous Phrases Miriam Kaeshammer University of D¨usseldorf Universit¨atsstraße 1 40225 D¨usseldorf, Germany kaeshammer@phil.uni-duesseldorf.de Abstract We present a hierarchical statistical ma- chine translation system which supports discontinuous constituents. It is based on synchronous linear context-free rewriting systems (SLCFRS), an extension to syn- chronous context-free grammars in which synchronized non-terminals span k ≥ 1 continuous blocks on either side of the bitext. This extension beyond context- freeness is motivated by certain complex alignment configurations that are beyond the alignment capacity of current transla- tion models and their relatively frequent occurrence in hand-aligned data. Our experiments for translating from German to English demonstrate the feasibility of training and decoding with more expres- sive translation models such as SLCFRS and show a modest improvement over a context-free baseline. 1 Introduction In statistical machine translation, phrase-based translation models with a beam search decoder (Koehn et al., 2003) and tree-based models with a CYK decoder represent two prominent types of approaches. The latter usually employ some form of synchronous context-free grammar (SCFG). They can be grouped into so-called hierarchi- cal phrase-based models that are formally syntax- based, such as in Chiang (2007), and models where hierarchical units are somehow linguisti- cally motivated, e.g. in Zollmann and Venugopal (2006) and Hoang and Koehn (2010). The adequacy of all of these models has been questioned, as the space of alignments that they generate is limited. Inside-out alignments are be- yond the alignment capacity of SCFG of rank 2 (henceforth 2-SCFG) and inversion transduc- (i) a b c d b d a c (ii) a b a1 b1 a2 b2 (iii) a1 b a2 b1 a b2 Figure 1: Complex alignment configurations: (i) inside-out alignment; (ii) CDTU; (iii) bonbon. The configurations can also occur upside down. tion grammar (Wu, 1997), but they can be gener- ated with phrase-based translation models thanks to the reordering component of standard de- coders. Cross-serial discontinuous translation units (CDTU) (Søgaard and Kuhn, 2009) and bon- bon configurations (Simard et al., 2005) in con- trast can neither be generated by a phrase-based translation system nor by an SCFG-based one. It is thereby assumed that a translation unit, the tran- sitive closure of a set of nodes of the bipartite alignment graph, represents minimal translational equivalence, and therefore that an adequate trans- lation grammar formalism should be able to gen- erate each translation unit separately. The aforementioned problematic alignment configurations are schematically depicted in Fig- ure 1. Alignment (i) is an inside-out alignment; it is formed by four translation units (a, b, c and d). CDTUs (ii) and bonbons (iii) each consist of two intertwined discontinuous translation units. Several studies have investigated the alignment capacity of SCFG-based and phrase-based trans- lation models in different setups (Wellington et al., 2006; Søgaard and Kuhn, 2009; Søgaard and Wu, 2009; Søgaard, 2010; Kaeshammer, 2013). For example, Wellington et al. (2006) find that inside-out alignments …

Content truncated to 500 words