Syntax-based Multi-system Machine Translation

Syntax-based Multi-system Machine Translation Matīss Rikters1, Inguna Skadiņa2 1,2University of Latvia, 19 Raina Blvd., Riga, Latvia 2Institute of Mathematics and Computer Science, 29 Raina Blvd., Riga, Latvia E-mail: matiss@lielakeda.lv, inguna.skadina@lumii.lv Abstract This paper describes a hybrid machine translation system that explores a parser to acquire syntactic chunks of a source sentence, translates the chunks with multiple online machine translation (MT) system application program interfaces (APIs) and creates output by combining translated chunks to obtain the best possible translation. The selection of the best translation hypothesis is performed by calculating the perplexity for each translated chunk. The goal of this approach is to enhance the baseline multi-system hybrid translation (MHyT) system that uses only a language model to select best translation from translations obtained with different APIs and to improve overall English – Latvian machine translation quality over each of the individual MT APIs. The presented syntax-based multi-system translation (SyMHyT) system demonstrates an improvement in terms of BLEU and NIST scores compared to the baseline system. Improvements reach from 1.74 up to 2.54 BLEU points. Keywords: hybrid machine translation, multi-system translation, syntactic parsing, under-resourced languages 1. Introduction Multi-system machine translation (MMT) is a type of hybrid machine translation (HMT) where multiple MT systems are combined in a single system in order to boost the accuracy and fluency of the translations. It is also referred to as multi-engine MT (Mellebeek et al., 2006), coupling MT (Ahsan and Kolachina, 2010), or just MT system combination (Barrault, 2010). Different approaches have been proposed for MMT. Traditional MSMT (Hildebrand, and Vogel, 2009) selects the best translation from a list of possible candidate translations generated by different MT engines using n-gram approach. Improvement has been reported when translated from French (+1.6 BLEU), German (+1.95 BLEU) or Hungarian (+1 BLEU) into English. However, application of similar approach for English-Latvian MT has resulted in insignificant improvement by only +0.12 BLEU points (Rikters, 2015). Recently Freitag et al. (2015) presented a novel system combination approach that enhances the traditional confusion network system combination approach (Heafield et al., .2009) with an additional model trained by a neural network. The proposed approach yielded in translation improvement from up to +0.9 points in BLEU and -0.5 points in TER for Chinese-English and Arabic-English. This paper presents a method that allows improving the MMT approach by incorporating syntactic information. These experiments were inspired by analysis of typical errors produced by statistical MT engines when translation is performed into a morphologically rich language with rather free order – Latvian (Skadiņa et al., 2012). This error analysis showed that the main type of errors is wrong inflection, which is usually caused by ignoring syntax rules. Our hypothesis is that translation of smaller, linguistically motivated chunks can improve this situation. We aim to enhance the simple baseline multi-system hybrid translation (MHyT) system1 (Rikters, 2015) and to improve MT quality for English-Latvian texts over each of the individual MT APIs. The experiments described in this paper use multiple combinations of outputs from two or three English-Latvian MT …

Content truncated to 500 words