String-to-Tree Multi Bottom-up Tree Transducers

Abstract

We achieve significant improvements in several syntax-based machine translation experiments using a string-to-tree variant of multi bottom-up tree transducers. Our new parameterized rule extraction algorithm extracts string-to-tree rules that can be discontiguous and non-minimal in contrast to existing algorithms for the tree-to-tree setting. The obtained models significantly outperform the string-to-tree component of the Moses framework in a large-scale empirical evaluation on several known translation tasks. Our linguistic analysis reveals the remarkable benefits of discontiguous and non-minimal rules.

Discontinuous Statistical Machine Translation with Target-Side Dependency Syntax

“…These fragments are applied synchronously, which allows the model to synchronously develop discontinuous parts in the output (e.g., to realize agreement). Overall, this translation model already proved to be useful when translating from English into German, Chinese, and Arabic as demonstrated by Seemann et al (2015). The goal of the current contribution is to adjust the approach and the system to Eastern European languages, for which we expect discontinuities to occur.…”

“…We use the string-to-tree variant (Seemann et al, 2015) of the multi bottom-up tree transducer (Maletti, 2010) as translation model. For simplicity, we call the variant 'MBOT'.…”

“…The goal of the current contribution is to adjust the approach and the system to Eastern European languages, for which we expect discontinuities to occur. The existing system (Seemann et al, 2015) cannot readily be applied since it requires constituent-like parses for the target side in our string-to-tree setting. However, for the target languages discussed here (Polish and Russian), only dependency parses are readily available.…”

Efficient techniques for parsing with tree automata

“…Classical grammar formalisms such as TAG (Joshi and Schabes, 1997) and CCG (Steedman, 2001) have been equipped with expressive statistical models, and high-performance parsers have become available (Clark and Curran, 2007;Lewis and Steedman, 2014;Kallmeyer and Maier, 2013). Synchronous grammar formalisms such as synchronous context-free grammars (Chiang, 2007) and tree-to-string transducers (Galley et al, 2004;Graehl et al, 2008;Seemann et al, 2015) are being used as models that incorporate syntactic information in statistical machine translation. Synchronous string-to-tree (Wong and Mooney, 2006) and string-to-graph grammars (Chiang et al, 2013) have been applied to semantic parsing; and so forth.…”

Query Rewriting for Nondeterministic Tree Transducers

“…In machine translation, each sentence in a source natural language can possibly be translated into more than one sentence in a target language (see, e.g., [12]- [14]). Thus we need nondeterminism in tree transducers that model syntax-based machine translations (see, e.g., [15], [16]). Bilingual documents are essentially required to construct statistical syntax-based translators.…”