Large-scale automatic extraction of an English-Chinese translation lexicon

We report experimental results on automatic extraction of an English-Chinese translation lexicon, by statistical analysis of a large parallel corpus, using limited amounts of linguistic knowledge. To our knowledge, these are the first empirical results of the kind between an Indo-European and non-Indo-European language for any significant vocabulary and corpus size. The learned vocabulary size is about 6,500 English words, achieving translation precision in the 86–96% range, with alignment proceeding at paragraph, sentence, and word levels. Specifically, we report (1) progress on the HKUST English-Chinese Parallel Bilingual Corpus, (2) experiments supporting the usefulness of restricted lexical cues for statistical paragraph and sentence alignment, and (3) experiments that question the role of hand-derived monolingual lexicons for automatic word translation acquisition. Using a hand-derived monolingual lexicon, the learned translation lexicon averages 2.33 Chinese translations per English entry, with a manually-filtered precision of 95.1%, and an automatically-filtered weighted precision of 86.0%. We then introduce a fully automatic two-stage statistical methodology that is able to learn translations for collocations. A statistically-learned monolingual Chinese lexicon is first used to segment the Chinese text, before applying bilingual training to produce 6,429 English entries with 2.25 Chinese translations per entry. This method improves the manually-filtered precision to 96.0% and the automatically-filtered weighted precision to 91.0%, an error rate reduction of 35.7% from using a hand-derived monolingual lexicon.     

Automatic induction of bilingual resources from aligned parallel corpora: application to shallow-transfer machine translation

The availability of machine-readable bilingual linguistic resources is crucial not only for rule-based machine translation but also for other applications such as cross-lingual information retrieval. However, the building of such resources (bilingual single-word and multi-word correspondences, translation rules) demands extensive manual work, and, as a consequence, bilingual resources are usually more difficult to find than “shallow” monolingual resources such as morphological dictionaries or part-of-speech taggers, especially when they involve a less-resourced language. This paper describes a methodology to build automatically both bilingual dictionaries and shallow-transfer rules by extracting knowledge from word-aligned parallel corpora processed with shallow monolingual resources (morphological analysers, and part-of-speech taggers). We present experiments for Brazilian Portuguese–Spanish and Brazilian Portuguese–English parallel texts. The results show that the proposed methodology can enable the rapid creation of valuable computational resources (bilingual dictionaries and shallow-transfer rules) for machine translation and other natural language processing tasks).     

融合单语语言模型的汉越伪平行语料生成

神经机器翻译在资源丰富的语种上取得了良好的翻译效果,但是由于数据稀缺问题在汉语-越南语这类低资源语言对上的性能不佳。目前缓解该问题最有效的方法之一是利用现有资源生成伪平行数据。考虑到单语数据的可利用性,在回译方法的基础上,首先将利用大量单语数据训练的语言模型与神经机器翻译模型进行融合,然后在回译过程中通过语言模型融入语言特性,以此生成更规范质量更优的伪平行数据,最后将生成的语料添加到原始小规模语料中训练最终翻译模型。在汉越翻译任务上的实验结果表明,与普通的回译方法相比,通过融合语言模型生成的伪平行数据使汉越神经机器翻译的BLEU值提升了1.41个百分点。     

Bilingual LSA-based adaptation for statistical machine translation

We propose a novel approach to cross-lingual language model and translation lexicon adaptation for statistical machine translation (SMT) based on bilingual latent semantic analysis. Bilingual LSA enables latent topic distributions to be efficiently transferred across languages by enforcing a one-to-one topic correspondence during training. Using the proposed bilingual LSA framework, model adaptation can be performed by, first, inferring the topic posterior distribution of the source text and then applying the inferred distribution to an n-gram language model of the target language and translation lexicon via marginal adaptation. The background phrase table is enhanced with the additional phrase scores computed using the adapted translation lexicon. The proposed framework also features rapid bootstrapping of LSA models for new languages based on a source LSA model of another language. Our approach is evaluated on the Chinese–English MT06 test set using the medium-scale SMT system and the GALE SMT system measured in BLEU and NIST scores. Improvement in both scores is observed on both systems when the adapted language model and the adapted translation lexicon are applied individually. When the adapted language model and the adapted translation lexicon are applied simultaneously, the gain is additive. At the 95% confidence interval of the unadapted baseline system, the gain in both scores is statistically significant using the medium-scale SMT system, while the gain in the NIST score is statistically significant using the GALE SMT system.     

Tag-less back-translation

An effective method to generate a large number of parallel sentences for training improved neural machine translation (NMT) systems is the use of the back-translations of the target-side monolingual data. The standard back-translation method has been shown to be unable to efficiently utilize huge amounts of existing monolingual data because of the inability of translation models to differentiate between authentic and synthetic parallel data during training. Tagging, or using gates, has been used to enable translation models to distinguish between synthetic and authentic data, improving standard back-translation and also enabling the use of iterative back-translation on language pairs that underperformed using standard back-translation. In this work, we approach back-translation as a domain adaptation problem, eliminating the need for explicit tagging. In our approach—tag-less back-translation—the synthetic and authentic parallel data are treated as out-of-domain and in-domain data, respectively, and through pre-training and fine-tuning, the translation model is shown to be able to learn more efficiently from them during training. Experimental results have shown that the approach outperforms the standard and tagged back-translation approaches on low resource English-Vietnamese and English-German NMT.

     

Soft syntactic constraints for Arabic–English hierarchical phrase-based translation

In adding syntax to statistical machine translation, there is a tradeoff between taking advantage of linguistic analysis and allowing the model to exploit parallel training data with no linguistic analysis: translation quality versus coverage. A number of previous efforts have tackled this tradeoff by starting with a commitment to linguistically motivated analyses and then finding appropriate ways to soften that commitment. We present an approach that explores the tradeoff from the other direction, starting with a translation model learned directly from aligned parallel text, and then adding soft constituent-level constraints based on parses of the source language. We argue that in order for these constraints to improve translation, they must be fine-grained: the constraints should vary by constituent type, and by the type of match or mismatch with the parse. We also use a different feature weight optimization technique, capable of handling large amount of features, thus eliminating the bottleneck of feature selection. We obtain substantial improvements in performance for translation from Arabic to English.     

The UPF learner translation corpus as a resource for translator training

The learner translation corpus developed at the School of Translation and Interpreting of Pompeu Fabra University in Barcelona is a web-searchable resource created for pedagogical and research purposes. It comprises a multiple translation corpus (English–Catalan) featuring automatic linguistic annotation and manual error annotation, complemented with an interface for monolingual or bilingual querying of the data. The corpus can be used to identify common errors in the students’ work and to analyse their patterns of language use. It provides easy access to error samples and to multiple versions of the same source text sequence to be used as learning materials in various courses in the translator-training university curriculum.     

An image-based automatic Arabic translation system

In this paper, we present a system that automatically translates Arabic text embedded in images into English. The system consists of three components: text detection from images, character recognition, and machine translation. We formulate the text detection as a binary classification problem and apply gradient boosting tree (GBT), support vector machine (SVM), and location-based prior knowledge to improve the F1 score of text detection from 78.95% to 87.05%. The detected text images are processed by off-the-shelf optical character recognition (OCR) software. We employ an error correction model to post-process the noisy OCR output, and apply a bigram language model to reduce word segmentation errors. The translation module is tailored with compact data structure for hand-held devices. The experimental results show substantial improvements in both word recognition accuracy and translation quality. For instance, in the experiment of Arabic transparent font, the BLEU score increases from 18.70 to 33.47 with use of the error correction module.     

利用单语数据改进神经机器翻译压缩模型的翻译质量

该文提出利用一个大型且精度高的神经机器翻译模型(教师模型)从单语数据中提取隐性双语知识,从而改进小型且精度低的神经机器翻译模型(学生模型)的翻译质量。该文首先提出了“伪双语数据”的教学方法,利用教师模型翻译单语数据获得的合成双语数据改进学生模型,然后提出了“负对数似然—知识蒸馏联合优化”教学方法,除了利用合成双语数据,还利用教师模型获得的目标语言词语概率分布作为知识,从而在知识蒸馏框架下提高学生模型的翻译质量。实验证明,在中英和德英翻译任务上,使用该方法训练的学生模型不仅在领域内测试集上显著超过了基线学生模型,而且在领域外测试集上的泛化性能也得到了提高。     

Crawl and crowd to bring machine translation to under-resourced languages

We present a widely applicable methodology to bring machine translation (MT) to under-resourced languages in a cost-effective and rapid manner. Our proposal relies on web crawling to automatically acquire parallel data to train statistical MT systems if any such data can be found for the language pair and domain of interest. If that is not the case, we resort to (1) crowdsourcing to translate small amounts of text (hundreds of sentences), which are then used to tune statistical MT models, and (2) web crawling of vast amounts of monolingual data (millions of sentences), which are then used to build language models for MT. We apply these to two respective use-cases for Croatian, an under-resourced language that has gained relevance since it recently attained official status in the European Union. The first use-case regards tourism, given the importance of this sector to Croatia’s economy, while the second has to do with tweets, due to the growing importance of social media. For tourism, we crawl parallel data from 20 web domains using two state-of-the-art crawlers and explore how to combine the crawled data with bigger amounts of general-domain data. Our domain-adapted system is evaluated on a set of three additional tourism web domains and it outperforms the baseline in terms of automatic metrics and/or vocabulary coverage. In the social media use-case, we deal with tweets from the 2014 edition of the soccer World Cup. We build domain-adapted systems by (1) translating small amounts of tweets to be used for tuning by means of crowdsourcing and (2) crawling vast amounts of monolingual tweets. These systems outperform the baseline (Microsoft Bing) by 7.94 BLEU points (5.11 TER) for Croatian-to-English and by 2.17 points (1.94 TER) for English-to-Croatian on a test set translated by means of crowdsourcing. A complementary manual analysis sheds further light on these results.     

利用上下文信息的统计机器翻译领域自适应

统计机器翻译系统用于翻译领域文本时,常常会遇到跨领域的问题 当待翻译文本与训练语料来自同一领域时,通常会得到较好的翻译效果;当领域差别较大时,翻译质量会明显下降。某个特定领域的双语平行语料是有限的,相对来说,领域混杂的平行语料和特定领域的单语文本更容易获得。该文充分利用这一特点,提出了一种包含领域信息的翻译概率计算模型,该模型联合使用混合领域双语和特定领域源语言单语进行机器翻译领域自适应。实验显示,自适应模型在IWSLT机器翻译评测3个测试集上均比Baseline有提高,证明了该文方法的有效性。     

基于自注意力机制预训练跨蒙汉语言模型

针对蒙汉机器翻译中平行语料资源稀缺的问题,提出利用单语语料库对蒙汉机器翻译进行研究。由于利用单语语料库进行机器翻译的效果较差,故将基于自注意力机制预训练跨蒙汉语言模型应用于基于单语语料库训练的蒙汉机器翻译系统中。实验结果表明,基于自注意力机制预训练跨蒙汉语言模型的方法极大改善了蒙汉机器翻译系统的性能。     

Symbolic-to-statistical hybridization: extending generation-heavy machine translation

The last few years have witnessed an increasing interest in hybridizing surface-based statistical approaches and rule-based symbolic approaches to machine translation (MT). Much of that work is focused on extending statistical MT systems with symbolic knowledge and components. In the brand of hybridization discussed here, we go in the opposite direction: adding statistical bilingual components to a symbolic system. Our base system is Generation-heavy machine translation (GHMT), a primarily symbolic asymmetrical approach that addresses the issue of Interlingual MT resource poverty in source-poor/target-rich language pairs by exploiting symbolic and statistical target-language resources. GHMT’s statistical components are limited to target-language models, which arguably makes it a simple form of a hybrid system. We extend the hybrid nature of GHMT by adding statistical bilingual components. We also describe the details of retargeting it to Arabic–English MT. The morphological richness of Arabic brings several challenges to the hybridization task. We conduct an extensive evaluation of multiple system variants. Our evaluation shows that this new variant of GHMT—a primarily symbolic system extended with monolingual and bilingual statistical components—has a higher degree of grammaticality than a phrase-based statistical MT system, where grammaticality is measured in terms of correct verb-argument realization and long-distance dependency translation.     

Statistical Translation of Text and Speech: First Results with the RWTH System

In this paper, we describe a first version of a system for statisticaltranslation and present experimental results. The statistical translationapproach uses two types of information: a translation model and a languagemodel. The language model used is a standard bigram model. The translationmodel is decomposed into lexical and alignment models. After presenting the details of the alignment model, we describe the search problem and present a dynamic programming-based solution for the special case of monotone alignments.So far, the system has been tested on two limited-domain tasks for which abilingual corpus is available: the EuTrans traveller task (Spanish–English,500-word vocabulary) and the Verbmobil task (German–English, 3000-wordvocabulary). We present experimental results on these tasks. In addition to the translation of text input, we also address the problem of speech translation and suitable integration of the acoustic recognition process and the translation process.     

Controlled Translation in an Example-based Environment: What do Automatic Evaluation Metrics Tell Us?

This paper presents an extended, harmonised account of our previous work on integrating controlled language data in an Example-based Machine Translation system. Gough and Way in MT Summit pp. 133–140 (2003) focused on controlling the output text in a novel manner, while Gough and Way (9th Workshop of the EAMT, (2004a), pp. 73–81) sought to constrain the input strings according to controlled language specifications. Our original sub-sentential alignment algorithm could deal only with 1:1 matches, but subsequent refinements enabled n:m alignments to be captured. A direct consequence was that we were able to populate the system’s databases with more than six times as many potentially useful fragments. Together with two simple novel improvements – correcting a small number of mistranslations in the lexicon, and allowing multiple translations in the lexicon – translation quality improves considerably. We provide detailed automatic and human evaluations of a number of experiments carried out to test the quality of the system. We observe that our system outperforms the rule-based on-line system Logomedia on a range of automatic evaluation metrics, and that the ‘best’ translation candidate is consistently highly ranked by our system. Finally, we note in a number of tests that the BLEU metric gives objectively different results than other automatic evaluation metrics and a manual evaluation. Despite these conflicting results, we observe a preference for controlling the source data rather than the target translations.     

利用Markov网络抽取复述增强机器译文自动评价方法

在机器译文自动评价中,匹配具有相同语义、不同表达方式的词或短语是其中一个很大的挑战。许多研究工作提出从双语平行语料或可比语料中抽取复述来增强机器译文和人工译文的匹配。然而双语平行语料或可比语料不仅构建成本高,而且对少数语言对难以大量获取。我们提出通过构建词的Markov网络,从目标语言的单语文本中抽取复述的方法,并利用该复述提高机器译文自动评价方法与人工评价方法的相关性。在WMT14 Metrics task上的实验结果表明,我们从单语文本中提取复述方法的性能与从双语平行语料中提取复述方法的性能具有很强的可比性。因此,该文提出的方法可在保证复述质量的同时,降低复述抽取的成本。
     

基于变分信息瓶颈的半监督神经机器翻译

变分方法是机器翻译领域的有效方法, 其性能较依赖于数据量规模. 然而在低资源环境下, 平行语料资源匮乏, 不能满足变分方法对数据量的需求, 因此导致基于变分的模型翻译效果并不理想. 针对该问题, 本文提出基于变分信息瓶颈的半监督神经机器翻译方法, 所提方法的具体思路为: 首先在小规模平行语料的基础上, 通过引入跨层注意力机制充分利用神经网络各层特征信息, 训练得到基础翻译模型; 随后, 利用基础翻译模型, 使用回译方法从单语语料生成含噪声的大规模伪平行语料, 对两种平行语料进行合并形成组合语料, 使其在规模上能够满足变分方法对数据量的需求; 最后, 为了减少组合语料中的噪声, 利用变分信息瓶颈方法在源与目标之间添加中间表征, 通过训练使该表征具有放行重要信息、阻止非重要信息流过的能力, 从而达到去除噪声的效果. 多个数据集上的实验结果表明, 本文所提方法能够显著地提高译文质量, 是一种适用于低资源场景的半监督神经机器翻译方法.     

基于对抗学习的蒙汉神经机器翻译

在机器翻译模型的构建和训练阶段,为了缓解因端到端机器翻译框架在训练时采用最大似然估计原理导致的翻译模型的质量不高的问题,本文使用对抗学习策略训练生成对抗网络,通过鉴别器协助生成器的方式来提高生成器的翻译质量,通过实验选择出了更适合生成器的机器翻译框架Transformer,更适合鉴别器的卷积神经网络,并且验证了对抗式训练对提高译文的自然度、流利度以及准确性都具有一定的作用.在模型的优化阶段,为了缓解因蒙汉平行数据集匮乏导致的蒙汉机器翻译质量仍然不理想的问题,本文将Dual-GAN (dual-generative adversarial networks,对偶生成对抗网络)算法引入了蒙汉机器翻译中,通过有效的利用大量蒙汉单语数据使用对偶学习策略的方式来进一步提高基于对抗学习的蒙汉机器翻译模型的质量.     

Improving statistical MT by coupling reordering and decoding

In this paper we describe an elegant and efficient approach to coupling reordering and decoding in statistical machine translation, where the n-gram translation model is also employed as distortion model. The reordering search problem is tackled through a set of linguistically motivated rewrite rules, which are used to extend a monotonic search graph with reordering hypotheses. The extended graph is traversed in the global search when a fully informed decision can be taken. Further experiments show that the n-gram translation model can be successfully used as reordering model when estimated with reordered source words. Experiments are reported on the Europarl task (Spanish–English and English–Spanish). Results are presented regarding translation accuracy and computational efficiency, showing significant improvements in translation quality with respect to monotonic search for both translation directions at a very low computational cost.