Integration of Multiple Bilingually-Trained Segmentation Schemes into Statistical Machine Translation

This paper proposes an unsupervised word segmentation algorithm that identifies word boundaries in continuous source language text in order to improve the translation quality of statistical machine translation (SMT) approaches for the translation of local dialects by exploiting linguistic information of the standard language. The method iteratively learns multiple segmentation schemes that are consistent with (1) the standard dialect segmentations and (2) the phrasal segmentations of an SMT system trained on the resegmented bitext of the local dialect. In a second step multiple segmentation schemes are integrated into a single SMT system by characterizing the source language side and merging identical translation pairs of differently segmented SMT models. Experimental results translating three Japanese local dialects (Kumamoto, Kyoto, Osaka) into three Indo-European languages (English, German, Russian) revealed that the proposed system outperforms SMT engines trained on character-based as well as standard dialect segmentation schemes for the majority of the investigated translation tasks and automatic evaluation metrics.KeywordsMachine TranslationTarget LanguageSource LanguageStatistical Machine TranslationWord SegmentationThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

In this paper, we present our work on a case study between statistical machine translation (SMT) and rule-based machine translation (RBMT) systems on English-Indian language and Indian to Indian language perspective. Main objective of our study is to make a five-way performance comparison: such as, (a) SMT and RBMT; (b) SMT on English–Indian language; (c) RBMT on English–Indian language; (d) SMT on Indian to Indian language perspective; (e) RBMT on Indian to Indian language perspective. Through a detailed analysis, we describe the rule-based and the statistical machine translation system developments and its evaluations. Further, with a detailed error analysis, we point out the relative strengths and weaknesses of both the systems. The observations based on our study are: (a) SMT systems outperform RBMT; (b) In the case of SMT: English to Indian language MT systems perform better than Indian to English language MT systems; (c) In the case of RBMT: English to Indian language MT systems perform better than Indian to English language MT systems; (d) SMT systems perform better for Indian to Indian language MT systems compared to RBMT. Effectively, we shall see that even with a small amount of training corpus SMT system has many advantages for high-quality domain-specific machine translation over that of a rule-based counterpart.

In recent years, statistical machine translation has gained much attention. The phrase‐based statistical machine translation model has made significant advancement in translation quality over the word‐based model. In this paper, we attempt to apply the technique of proportional analogy to statistical machine translation systems. We propose a novel approach to apply proportional analogy to generate translations of unseen n‐grams from the phrase table for phrase‐based statistical machine translation. Experiments are conducted with two datasets of different sizes. We also investigate two methods to integrate n‐grams translations produced by proportional analogy into the state‐of‐the‐art statistical machine translation system, Moses.1The experimental results show that unseen n‐grams translations generated using the technique of proportional analogy are rewarding for statistical machine translation systems with small datasets. © 2016 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

We report insights from translating Spanish conversational telephone speech into English text by cascading an automatic speech recognition (ASR) system with a statistical machine translation (SMT) system. The key new insight is that the informal register of conversational speech is a greater challenge for ASR than for SMT: the BLEU score for translating the reference transcript is 64%, but drops to 32% for translating automatic transcripts, whose word error rate (WER) is 40%. Several strategies are examined to mitigate the impact of ASR errors on the SMT output: (i) providing the ASR lattice, instead of the 1-best output, as input to the SMT system, (ii) training the SMT system on Spanish ASR output paired with English text, instead of Spanish reference transcripts, and (iii) improving the core ASR system. Each leads to consistent and complementary improvements in the SMT output. Compared to translating the 1-best output of an ASR system with 40% WER using an SMT system trained on Spanish reference transcripts, translating the output lattice of a better ASR system with 35% WER using an SMT system trained on ASR output improves BLEU from 32% to 38%.

A simplification–translation–restoration framework for domain adaptation in statistical machine translation: A case study in medical record translation

Statistical Machine Translation (SMT) has been the dominant corpus-based machine translation (MT) approach in the last twenty years. While SMT has been studied in detail among European languages, it has not been studied sufficiently in language pairs including Turkish as source or target language, and its study has been limited mostly to English ↔ Turkish language pair. This study aims to broaden the perspective on Turkish corpus-based MT studies by training MT engines between Turkish and a wide variety of languages with different features. It surveys customized SMT between Turkish and 11 different languages. Twenty-two SMT engines have been trained in KantanMT with open parallel corpora using Turkish as both source and target language. Three automatic evaluation metrics F-Measure, BLEU, and TER have been used for evaluating MT quality. Due to the variations in the corpus quality and size, highly varying results have been achieved. While Turkish ↔ Catalan engines have had the highest automatic evaluation scores, Turkish ↔ Arabic engines have had the lowest automatic scores. While the quality results are highly varying across languages, we obtain baseline scores for a wide variety of languages coupled with Turkish. These results may provide a reference point for evaluating future MT systems including Turkish.

There is no doubt that in the last couple of years corpus-based machine translation (CBMT) approaches have been in focus. Each of the CBMT approaches (statistical MT and example-based MT) has its advantages and disadvantages. Therefore, hybrid approaches have been developed. This paper presents a comparative study of CBMT approaches, using three types of systems: a statistical MT (SMT) system, an example-based MT (EBMT) system and a hybrid (EBMT-SMT) system. For our experiments we considered German-Romanian as language-pair, in this direction of translation, and a domain-restricted corpus, manually created and aligned (RoGER). The translation results have been both automatically and manually evaluated. For both evaluation types, the SMT system is ranked best in our experiments.

In the modern era, the field of Machine Translation (MT) has seen a significant shift towards Neural Machine Translation (NMT) techniques, which have surpassed traditional Statistical Machine Translation (SMT) models in terms of the quality of translation. Despite this, the efficacy of these techniques may differ based on the language combination in consideration. While SMT is somewhat more flexible in this regard, NMT often needs sizable parallel corpora to attain high translation accuracy. As a result, a benchmark system capable of offering sufficient translation for languages with limited resources, like Nepali, remains a pipe dream. This paper focuses on translating text using statistical and neural MT techniques for the under-resourced English-Nepali language pair. As a part of this system development, we built a parallel corpus of English-Nepali in the tourism domain. We explore the impact of different tokenization techniques on translation outcomes. A substantial analysis is also done for the performance of both approaches using automatic evaluation metrics, BLEU and TER. This paper aims to provide insights into the applicability of SMT and NMT for the under-resourced English-Nepali language pair in light of two popular epitomes of tokenization and to determine the most effective approach for achieving accurate translations.

In terms of language technology, Maithili is a resource-poor language. It is spoken in India and Nepal and is one of the 22 scheduled languages in India. Maithili has almost no language technology resource. English in India happens to be a dominant language in terms of content and usage. However, since more than 90% of Indians do not use English, a translation from English to Maithili becomes critical. An absence of basic tools in this language has affected resource creation of machine translation (MT). The present work discusses efforts for language technology resource (LTR) creation and divergence study for a statistical English-Maithili MT (EMMT) system. Creating any statistical MT (SMT) system requires sizeable parallel, aligned corpora for training and testing. Creating general-purpose source corpora for English language and creating translation equivalents with possible alignments require time and effort. The paper focuses on the data collection methods, cleaning, the size and structure of the text corpora, alignment and parallelization strategies, training, testing, and a study of divergence between the language pair.

Although Machine Translation has historically trusted on huge amounts of parallel corpora, the latest analysis has accomplished to prepare each Neural and Statistical Machine Translation system using monolingual corpora only. In spite of the prospective of this methodology for low-resource settings, obtainable structures square measure way outstanding their supervised counterparts, restraining their concrete interest. In this paper, Sect. 1 contains numerous deficiencies of existing unsupervised SMT approaches by exploiting subword information. Section 2 consists of another methodology established on phrase-based statistical machine translation that significantly cessations the gap with supervised structures. Principled Unsupervised Statistical Machine Translation in Sect. 3. Results and discussions in Sect. 4 and conclusion in Sect. 5.

Morphologically complex languages such as Arabic pose several challenges in Natural Language Processing (NLP) due to their complexity and token sparsity. Most techniques approach the problem by transforming the words of the language from their sparse surface form representation to a less sparse form before processing. The transformation usually takes the form of a morphological analysis or a morphological segmentation. This dissertation addresses two tasks in Arabic NLP: Statistical Machine Translation(SMT) and Sentiment Analysis. To improve English-Arabic SMT, we apply segmentation on Arabic to decrease token sparsity and enhance the correspondence between tokens of the English and Arabic language. However, due to this segmentation, the translation system is limited to extracting features based on morphemes (partial words) and only outputting morphemes during decoding. Such a system lacks knowledge of the original form of the words. We further improve translation from English to Arabic by combining both segmented and desegmented views of the target language. The system can benefit from segmentation's sparsity reduction and verifies its generation of correct words. We present a language-independent technique to desegmentation that approaches the problem as a string transduction task. We propose a new algorithm that desegments the decoder's search space encoded as a lattice, thus allowing the system to use features from the desegmented view of the search space. We extend the phrase-based statistical machine translation system to allow desegmentation during the decoding process on the fly. In addition, we conduct an experimental study to verify what matters most in morphologically segmented SMT models. Our second task is sentiment analysis, where we resort to Arabic lemmatization to improve sentiment analysis of Arabic tweets and blog posts. We explore translation in the opposite direction, from Arabic into English in order to evaluate the loss of sentiment predictability when Arabic social media posts are translated to English, manually or using an SMT system. We use state-of-the-art Arabic and English sentiment Analysis systems and develop automatically generated Arabic lexicons from lemmatized tweets to improve this task.

We revisit the one sense per discourse hypothesis of Gale et al. in the context of machine translation. Since a given sense can be lexicalized differently in translation, do we observe one translation per discourse? Analysis of manual translations reveals that the hypothesis still holds when using translations in parallel text as sense annotation, thus confirming that translational differences represent useful sense distinctions. Analysis of Statistical Machine Translation (SMT) output showed that despite ignoring document structure, the one translation per discourse hypothesis is strongly supported in part because of the low variability in SMT lexical choice. More interestingly, cases where the hypothesis does not hold can reveal lexical choice errors. A preliminary study showed that enforcing the one translation per discourse constraint in SMT can potentially improve translation quality, and that SMT systems might benefit from translating sentences within their entire document context.

Deep neural networks (DNNs) have provably enhanced the state-of-the-art natural language process (NLP) with their capability of feature learning and representation. As one of the more challenging NLP tasks, neural machine translation (NMT) becomes a new approach to machine translation and generates much more fluent results compared to statistical machine translation (SMT). However, SMT is usually better than NMT in translation adequacy and word coverage. It is therefore a promising direction to combine the advantages of both NMT and SMT. In this article, we propose a deep neural network--based system combination framework leveraging both minimum Bayes-risk decoding and multi-source NMT, which take as input the N-best outputs of NMT and SMT systems and produce the final translation. In particular, we apply the proposed model to both RNN and self-attention networks with different segmentation granularity. We verify our approach empirically through a series of experiments on resource-rich Chinese⇒English and low-resource English⇒Vietnamese translation tasks. Experimental results demonstrate the effectiveness and universality of our proposed approach, which significantly outperforms the conventional system combination methods and the best individual system output.

We conduct the first experiment in the literature in which a novel is translated automatically and then post-edited by professional literary translators. Our case study is Warbreaker, a popular fantasy novel originally written in English, which we translate into Catalan. We translated one chapter of the novel (over 3,700 words, 330 sentences) with two data-driven approaches to Machine Translation (MT): phrase-based statistical MT (PBMT) and neural MT (NMT). Both systems are tailored to novels; they are trained on over 100 million words of fiction. In the post-editing experiment, six professional translators with previous experience in literary translation translate subsets of this chapter under three alternating conditions: from scratch (the norm in the novel translation industry), post-editing PBMT, and post-editing NMT. We record all the keystrokes, the time taken to translate each sentence, as well as the number of pauses and their duration. Based on these measurements, and using mixed-effects models, we study post-editing effort across its three commonly studied dimensions: temporal, technical and cognitive. We observe that both MT approaches result in increases in translation productivity: PBMT by 18%, and NMT by 36%. Post-editing also leads to reductions in the number of keystrokes: by 9% with PBMT, and by 23% with NMT. Finally, regarding cognitive effort, post-editing results in fewer (29 and 42% less with PBMT and NMT, respectively) but longer pauses (14 and 25%).

The inability to deal with words in foreign languages imposes difficulties to both Vietnamese speech recognition and text-to-speech systems. A common solution is to look up a dictionary, but the number of available entries is finite and therefore not flexible because speech recognition and text-to-speech systems are expected to handle arbitrary words. Alternatively, data-driven approaches can be employed to transliterate a foreign word into its Vietnamese pronunciation by learning samples and predicting unseen words. This paper presents a comparative analysis between two data-driven approaches based on statistical machine translation and joint-sequence model. Two systems based on these approaches are developed and tested using the same experimental protocol and a dataset consisting of 8050 English words. Results show that joint-sequence model outperforms statistical machine translation in English-to-Vietnamese transliteration.