This paper presents ProphetMT, a monolingual Controlled Language (CL) authoring tool which allows users to easily compose an in-domain sentence with the help of tree-based SMT-driven auto-suggestions. The interface also visualizes target-language sentences as they are built by the SMT system. When the user is finished composing, the final translation(s) are generated by a tree-based SMT system using the text and structural information provided by the user. With this domain-specific controlled language, ProphetMT will produce highly reliable translations. The contributions of this work are: 1) we develop a user-friendly auto-completion-based editor which guarantees that the vocabulary and grammar chosen by a user are compatible with a tree-based SMT model; 2) by applying a shift-reduce-like parsing feature, this editor allows users to write from left-to-right and generates the parsing results on the fly. Accordingly, with this in-domain composing restriction as well as the gold-standard parsing result, a highly reliable translation can be generated.

This paper presents ProphetMT, a tree-based SMT-driven Controlled Language (CL) authoring and post-editing tool. ProphetMT employs the source-side rules in a translation model and provides them as auto-suggestions to users. Accordingly, one might say that users are writing in a 'Controlled Language' that is `understood' by the computer. ProphetMT also allows users to easily attach structural information as they compose content. When a specific rule is selected, a partial translation is promptly generated on-the-fly with the help of the structural information. Our experiments conducted on English-to-Chinese show that our proposed ProphetMT system can not only better regularise an author's writing behaviour, but also significantly improve translation fluency which is vital to reduce the post-editing time. Additionally, when the writing and translation process is over, ProphetMT can provide an effective colour scheme to further improve the productivity of post-editors by explicitly featuring the relations between the source and target rules.

Neural network training has been shown to be advantageous in many natural language processing applications, such as language modelling or machine translation. In this paper, we describe in detail a novel domain adaptation mechanism in neural network training. Instead of learning and adapting the neural network on millions of training sentences – which can be very time-consuming or even infeasible in some cases – we design a domain adaptation gating mechanism which can be used in recurrent neural networks and quickly learn the out-of-domain knowledge directly from the word vector representations with little speed overhead. In our experiments, we use the recurrent neural network language model (LM) as a case study. We show that the neural LM perplexity can be reduced by 7.395 and 12.011 using the proposed domain adaptation mechanism on the Penn Treebank and News data, respectively. Furthermore, we show that using the domain-adapted neural LM to re-rank the statistical machine translation n-best list on the French-to-English language pair can significantly improve translation quality.

Human-designed sub-structures are required by most of the syntax-based machine translation evaluation metrics. In this paper, we propose a novel evaluation metric based on dependency parsing model, which does not need this human involvement. Experimental results show that the new single metric gets better correlation than METEOR on system level and is comparable with it on sentence level. To introduce more information, we combine the new metric with many other metrics. The combined metric obtains state-of-theart performance on both system level evaluation and sentence level evaluation on WMT 2014.

The widely-used automatic evaluation metrics cannot adequately reflect the fluency of the translations. The n-gram-based metrics, like BLEU, limit the maximum length of matched fragments to n and cannot catch the matched fragments longer than n, so they can only reflect the fluency indirectly. METEOR, which is not limited by n-gram, uses the number of matched chunks but it does not consider the length of each chunk. In this paper, we propose an entropy-based method, which can sufficiently reflect the fluency of translations through the distribution of matched words. This method can easily combine with the widely-used automatic evaluation metrics to improve the evaluation of fluency. Experiments show that the correlations of BLEU and METEOR are improved on sentence level after combining with the entropy-based method on WMT 2010 and WMT 2012.

We describe our participation in TweetMT for three language pairs in both directions: Spanish from/to Catalan, Basque and Portuguese. We used a range of techniques: statistical and rule-based MT, morph segmentation, data selection with ParFDA and system combination. As for resources, our focus was on crawling vast amounts of tweets to perform monolingual domain adaptation. Our system was the best of all systems submitted for five out of the six language directions.

The widely-used automatic evaluation metrics cannot adequately reflect the fluency of the translations. The n-gram-based metrics, like BLEU, limit the maximum length of matched fragments to n and cannot catch the matched fragments longer than n, so they can only reflect the fluency indirectly. METEOR, which is not limited by n-gram, uses the number of matched chunks but it does not consider the length of each chunk. In this paper, we propose an entropy-based method, which can sufficiently reflect the fluency of translations through the distribution of matched words. This method can easily combine with the widely-used automatic evaluation metrics to improve the evaluation of fluency. Experiments show that the correlations of BLEU and METEOR are improved on sentence level after combining with the entropy-based method on WMT 2010 and WMT 2012.

Based on the last year’s DCU-CASIST participation on WMT metrics task, we further improve our model in the following ways: 1) parameter tuning 2) support languages other than English. We tuned our system on all the data of WMT 2010, 2012 and 2013. The tuning results as well as the WMT 2014 test results are reported.

Based on the last year’s DCU-CASIST participation on WMT metrics task, we further improve our model in the following ways: 1) parameter tuning 2) support languages other than English. We tuned our system on all the data of WMT 2010, 2012 and 2013. The tuning results as well as the WMT 2014 test results are reported.

This paper describes the DCU-Lingo24 submission to WMT 2014 for the HindiEnglish translation task. We exploit miscellaneous methods in our system, including: Context-Informed PB-SMT, OOV Word Conversion (OWC), MultiAlignment Combination (MAC), Operation Sequence Model (OSM), Stemming Align and Normal Phrase Extraction (SANPE), and Language Model Interpolation (LMI). We also describe various preprocessing steps we tried for Hindi in this task.

This paper describes the DCU submission to WMT 2014 on German-English translation task. Our system uses phrasebased translation model with several popular techniques, including Lexicalized Reordering Model, Operation Sequence Model and Language Model interpolation. Our final submission is the result of system combination on several systems which have different pre-processing and alignments.

Most of the widely-used automatic evaluation metrics consider only the local fragments of the references and translations, and they ignore the evaluation on the syntax level. Current syntaxbased evaluation metrics try to introduce syntax information but suffer from the poor parsing results of the noisy machine translations. To alleviate this problem, we propose a novel dependency-based evaluation metric which only employs the dependency information of the references. We use two kinds of reference dependency structures: headword chain to capture the long distance dependency information, and fixed and floating structures to capture the local continuous ngram. Experiment results show that our metric achieves higher correlations with human judgments than BLEU, TER and HWCM on WMT 2012 and WMT 2013. By introducing extra linguistic resources and tuning parameters, the new metric gets the state-of-the-art performance which is better than METEOR and SEMPOS on system level, and is comparable with METEOR on sentence level on WMT 2012 and WMT 2013.

In this paper, we propose a novel syntactic based MT evaluation metric which only employs the dependency information in the source side. Experimental results show that our method achieves higher correlation with human judgments than BLEU, TER, HWCM and METEOR at both sentence and system level for all of the four language pairs in WMT 2010.

This paper presents the experiments conducted by the Machine Translation group at DCU and Prompsit Language Engineering for the WMT13 translation task. Three language pairs are considered: SpanishEnglish and French-English in both directions and German-English in that direction. For the Spanish-English pair, the use of linguistic information to select parallel data is investigated. For the FrenchEnglish pair, the usefulness of the small indomain parallel corpus is evaluated, compared to an out-of-domain parallel data sub-sampling method. Finally, for the German-English system, we describe our work in addressing the long distance reordering problem and a system combination strategy.

This paper provides the system description of the IHMM team of Dublin City University for our participation in the system combination task in the Second Workshop on Applying Machine Learning Techniques to Optimise the Division of Labour in Hybrid MT (ML4HMT-12). Our work is based on a confusion network-based approach to system combination. We propose a new method to build a confusion network for this: (1) incorporate extra alignment information extracted from given meta data, treating them as sure alignments, into the results from IHMM, and (2) decode together with this information. We also heuristically set one of the system outputs as the default backbone. Our results show that this backbone, which is the RBMT system output, achieves an 0.11% improvement in BLEU over the backbone chosen by TER, while the extra information we added in the decoding part does not improve the results.

In this paper we describe the multi-engine machine translation system we used in the Japanese-Chinese News subtask in the 7 th Chinese Workshop on Machine Translation (CWMT'2011). We employed an MBR decoder and a Confusion Network decoder in our approach to combine different system outputs. We also describe the modular design of the system and report on how we pre-process the corpus.