In practice, even a well-trained neural machine translation (NMT) model can still make biased inferences on the training set due to distribution shifts. For the human learning process, if we can not reproduce something correctly after learning it multiple times, we consider it to be more difficult. Likewise, a training example causing a large discrepancy between inference and reference implies higher learning difficulty for the MT model. Therefore, we propose to adopt the inference discrepancy of each training example as the difficulty criterion, and according to which rank training examples from easy to hard. In this way, a trained model can guide the curriculum learning process of an initial model identical to itself. We put forward an analogy to this training scheme as guiding the learning process of a curriculum NMT model by a pretrained vanilla model. In this paper, we assess the effectiveness of the proposed training scheme and take an insight into the influence of translation direction, evaluation metrics and different curriculum schedules. Experimental results on translation benchmarks WMT14 English ⇒ German, WMT17 Chinese ⇒ English and Multitarget TED Talks Task (MTTT) English ⇔ German, English ⇔ Chinese, English ⇔ Russian demonstrate that our proposed method consistently improves the translation performance against the advanced Transformer baseline.

As the mainstream approach in the field of machine translation, neural machine translation (NMT) has achieved great improvements on many rich-source languages, but performance of NMT for low-resource languages ae not very good yet. This paper uses data enhancement technology to construct Mongolian-Chinese pseudo parallel corpus, so as to improve the translation ability of Mongolian-Chinese translation model. Experiments show that the above methods can improve the translation ability of the translation model. Finally, a translation model trained with large-scale pseudo parallel corpus and integrated with soft context data enhancement technology is obtained, and its BLEU value is 39.3.

Translation of long and complex sentence has always been a challenge for machine translation. In recent years, neural machine translation (NMT) has achieved substantial progress in modeling the semantic connection between words in a sentence, but it is still insufficient in capturing discourse structure information between clauses within complex sentences, which often leads to poor discourse coherence when translating long and complex sentences. On the other hand, the hypotactic structure, a main component of the discourse structure, plays an important role in the coherence of discourse translation, but it is not specifically studied. To tackle this problem, we propose a novel Chinese-English NMT approach that incorporates the hypotactic structure knowledge of complex sentences. Specifically, we first annotate and build a hypotactic structure aligned parallel corpus to provide explicit hypotactic structure knowledge of complex sentences for NMT. Then we propose three hypotactic structure-aware NMT models with three different fusion strategies, including source-side fusion, target-side fusion, and both-side fusion, to integrate the annotated structure knowledge into NMT. Experimental results on WMT17, WMT18 and WMT19 Chinese-English translation tasks demonstrate that the proposed method can significantly improve the translation performance and enhance the discourse coherence of machine translation.

Standard neural machine translation (NMT) is on the assumption that the document-level context is independent. Most existing document-level NMT approaches are satisfied with a smattering sense of global document-level information, while this work focuses on exploiting detailed document-level context in terms of a memory network. The capacity of the memory network that detecting the most relevant part of the current sentence from memory renders a natural solution to model the rich document-level context. In this work, the proposed document-aware memory network is implemented to enhance the Transformer NMT baseline. Experiments on several tasks show that the proposed method significantly improves the NMT performance over strong Transformer baselines and other related studies.

Since deep learning was introduced, a series of achievements has been published in the field of automatic machine translation (MT). However, Korean-Vietnamese MT systems face many challenges because of a lack of data, multiple meanings of individual words, and grammatical diversity that depends on context. Therefore, the quality of Korean-Vietnamese MT systems is still sub-optimal. This paper discusses a method for applying Named Entity Recognition (NER) and Part-of-Speech (POS) tagging to Vietnamese sentences to improve the performance of Korean-Vietnamese MT systems. In terms of implementation, we used a tool to tag NER and POS in Vietnamese sentences. In addition, we had access to a Korean-Vietnamese parallel corpus with more than 450K paired sentences from our previous research paper. The experimental results indicate that tagging NER and POS in Vietnamese sentences can improve the quality of Korean-Vietnamese Neural MT (NMT) in terms of the Bi-Lingual Evaluation Understudy (BLEU) and Translation Error Rate (TER) score. On average, our MT system improved by 1.21 BLEU points or 2.33 TER scores after applying both NER and POS tagging to the Vietnamese corpus. Due to the structural features of language, the MT systems in the Korean to Vietnamese direction always give better BLEU and TER results than translation machines in the reverse direction.

Since a concept can be represented by different vocabularies, styles, and levels of detail, a translation task resembles a many-to-many mapping task from a distribution of sentences in the source language into a distribution of sentences in the target language. This viewpoint, however, is not fully implemented in current neural machine translation (NMT), which is one-to-one sentence mapping. In this study, we represent the distribution itself as multiple paraphrase sentences, which will enrich the model context understanding and trigger it to produce numerous hypotheses. We use a visually grounded paraphrase (VGP), which uses images as a constraint of the concept in paraphrasing, to guarantee that the created paraphrases are within the intended distribution. In this way, our method can also be considered as incorporating image information into NMT without using the image itself. We implement this idea by crowdsourcing a paraphrasing corpus that realizes VGP and construct neural paraphrasing that behaves as expert models in a NMT. Our experimental results reveal that our proposed VGP augmentation strategies showed improvement against a vanilla NMT baseline.

Attention mechanism, which selectively focuses on source-side information to learn a context vector for generating target words, has been shown to be an effective method for neural machine translation (NMT). In fact, generating target words depends on not only the source-side information but also the target-side information. Although the vanilla NMT can acquire target-side information implicitly by recurrent neural networks (RNN), RNN cannot adequately capture the global relationship between target-side words. To solve this problem, this paper proposes a novel target-attention approach to capture this information, thus enhancing target word predictions in NMT. Specifically, we propose three variants of target-attention model to directly obtain the global relationship among target words: 1) a forward target-attention model that uses a target attention mechanism to incorporate previous historical target words into the prediction of the current target word; 2) a reverse target-attention model that adopts a reverse RNN model to obtain the entire reverse target words information, and then to combine with source context information to generate target sequence; 3) a bidirectional target-attention model that combines the forward target-attention model and reverse target-attention model together, which can make full use of target words to further improve the performance of NMT. Our methods can be integrated into both RNN based NMT and self-attention based NMT, and help NMT get global target-side information to improve translation performance. Experiments on the NIST Chinese-to-English and the WMT English-to-German translation tasks show that the proposed models achieve significant improvements over state-of-the-art baselines.

Word order is one of the most significant differences between the Chinese and Vietnamese. In the phrase-based statistical machine translation, the reordering model will learn reordering rules from bilingual corpora. If the bilingual corpora are large and good enough, the reordering rules are exact and coverable. However, Chinese-Vietnamese is a low-resource language pair, the extraction of reordering rules is limited. This leads to the quality of reordering in Chinese-Vietnamese machine translation is not high. In this paper, we have combined Chinese dependency relation and Chinese-Vietnamese word alignment results in order to pre-order Chinese word order to be suitable to Vietnamese one. The experimental results show that our methodology has improved the machine translation performance compared to the translation system using only the reordering models of phrase-based statistical machine translation.

Learning semantic representation for translation context is beneficial to statistical machine translation (SMT). Previous efforts have focused on implicitly encoding syntactic and semantic knowledge in translation context by neural networks, which are weak in capturing explicit structural syntax information. In this paper, we propose a new neural network with a tree-based convolutional architecture to explicitly learn structural syntax information in translation context, thus improving translation prediction. Specifically, we first convert parallel sentences with source parse trees into syntax-based linear sequences based on a minimum syntax subtree algorithm, and then define a tree-based convolutional network over the linear sequences to learn syntax-based context representation and translation prediction jointly. To verify the effectiveness, the proposed model is integrated into phrase-based SMT. Experiments on large-scale Chinese-to-English and German-to-English translation tasks show that the proposed approach can achieve a substantial and significant improvement over several baseline systems.

The-state-of-the-art neural quality estimation (QE) of machine translation model consists of two sub-networks that are tuned separately, a bidirectional recurrent neural network (RNN) encoder-decoder trained for neural machine translation, called the predictor, and an RNN trained for sentence-level QE tasks, called the estimator. We propose to combine the two sub-networks into a whole neural network, called the unified neural network. When training, the bidirectional RNN encoder-decoder are initialized and pre-trained with the bilingual parallel corpus, and then, the networks are trained jointly to minimize the mean absolute error over the QE training samples. Compared with the predictor and estimator approach, the use of a unified neural network helps to train the parameters of the neural networks that are more suitable for the QE task. Experimental results on the benchmark data set of the WMT17 sentence-level QE shared task show that the proposed unified neural network approach consistently outperforms the predictor and estimator approach and significantly outperforms the other baseline QE approaches.

This study introduces large-scale field experiments of VoiceTra, which is the world's first speech-to-speech multilingual translation application for smart phones. In the study, approximately 10 million input utterances were collected since the experiments commenced. The usage of collected data was analyzed and discussed. The study has several important contributions. First, it explains system configuration, communication protocol between clients and servers, and details of multilingual automatic speech recognition, multilingual machine translation, and multilingual speech synthesis subsystems. Second, it demonstrates the effects of mid-term system updates using collected data to improve an acoustic model, a language model, and a dictionary. Third, it analyzes system usage.

In this paper, a morpheme-based weighting and its integration method are proposed as a smoothing method to alleviate the data sparseness in Chinese-Mongolian statistical machine translation (SMT). Besides, we present source-side reordering as the pre-processing model to verify the extensibility of our method. Experi-mental results show that the morpheme-based weighting can substantially improve the translation quality.

This paper describes the design and evaluation of a method for developing a chat-oriented dialog system by utilizing real human-to-human conversation examples from movie scripts and Twitter conversations. The aim of the proposed method is to build a conversational agent that can interact with users in as natural a fashion as possible, while reducing the time requirement for database design and collection. A number of the challenging design issues we faced are described, including (1) constructing an appropriate dialog corpora from raw movie scripts and Twitter data, and (2) developing an multi domain chat-oriented dialog management system which can retrieve a proper system response based on the current user query. To build a dialog corpus, we propose a unit of conversation called a tri-turn (a trigram conversation turn), as well as extraction and semantic similarity analysis techniques to help ensure that the content extracted from raw movie/drama script files forms appropriate dialog-pair (query-response) examples. The constructed dialog corpora are then utilized in a data-driven dialog management system. Here, various approaches are investigated including example-based (EBDM) and response generation using phrase-based statistical machine translation (SMT). In particular, we use two EBDM: syntactic-semantic similarity retrieval and TF-IDF based cosine similarity retrieval. Experiments are conducted to compare and contrast EBDM and SMT approaches in building a chat-oriented dialog system, and we investigate a combined method that addresses the advantages and disadvantages of both approaches. System performance was evaluated based on objective metrics (semantic similarity and cosine similarity) and human subjective evaluation from a small user study. Experimental results show that the proposed filtering approach effectively improve the performance. Furthermore, the results also show that by combing both EBDM and SMT approaches, we could overcome the shortcomings of each.

In this study, we have developed a translation repair method to automatically improve the accuracy of translations. Machine translation (MT) supports multilingual communication; however, it cannot achieve high accuracy. MT creates only one translated sentence; therefore, it is difficult to improve the accuracy of translated sentences. Our method creates multiple translations by adding personal pronouns to the source sentence and by using a word dictionary and a parallel corpus. In addition, it selects an accurate translation from among the multiple translations using the results of a Web search. As a result, the translation repair method improved the accuracy of translated sentences, and its accuracy is greater than that of MT.

In phrase-based statistical machine translation, long distance reordering problem is one of the most challenging issues when translating syntactically distant language pairs. In this paper, we propose a novel reordering model to solve this problem. In our model, reordering is affected by the overall structures of sentences such as listings, reduplications, and modifications as well as the relationships of adjacent phrases. To this end, we reflect global syntactic contexts including the parts that are not yet translated during the decoding process.

We describe a novel idea to improve machine translation by combining multiple candidate translations and extra translations. Without manual work, extra translations can be generated by identifying and hybridizing the syntactic equivalents in candidate translations. Candidate and extra translations are then combined on sentence level for better general translation performance.

This paper explores the problem of semantic enrichment of mathematical expressions. We formulate this task as the translation of mathematical expressions from presentation markup to content markup. We use MathML, an application of XML, to describe both the structure and content of mathematical notations. We apply a method based on statistical machine translation to extract translation rules automatically. This approach contrasts with previous research, which tends to rely on manually encoded rules. We also introduce segmentation rules used to segment mathematical expressions. Combining segmentation rules and translation rules strengthens the translation system and archives significant improvements over a prior rule-based system.

In most phrase-based statistical machine translation (SMT) systems, the translation model relies on word alignment, which serves as a constraint for the subsequent building of a phrase table. Word alignment is usually inferred by GIZA++, which implements all the IBM models and HMM model in the framework of Expectation Maximum (EM). In this paper, we present a fully Bayesian inference for word alignment. Different from the EM approach, the Bayesian inference makes use of all possible parameter values rather than estimating a single parameter value, from which we expect a more robust inference. After inferring the word alignment, current SMT systems usually train the phrase table from Viterbi word alignment, which is prone to learn incorrect phrases due to the word alignment mistakes. To overcome this drawback, a new phrase extraction method is proposed based on multiple Gibbs samples from Bayesian inference for word alignment. Empirical results show promising improvements over baselines in alignment quality as well as the translation performance.

The performance of English automatic speech recognition systems decreases when recognizing spontaneous speech mainly due to multiple pronunciation variants in the utterances. Previous approaches address this problem by modeling the alteration of the pronunciation on a phoneme to phoneme level. However, the phonetic transformation effects induced by the pronunciation of the whole sentence have not yet been considered. In this article, the sequence-based pronunciation variation is modeled using a noisy channel approach where the spontaneous phoneme sequence is considered as a “noisy” string and the goal is to recover the “clean” string of the word sequence. Hereby, the whole word sequence and its effect on the alternation of the phonemes will be taken into consideration. Moreover, the system not only learns the phoneme transformation but also the mapping from the phoneme to the word directly. In this study, first the phonemes will be recognized with the present recognition system and afterwards the pronunciation variation model based on the noisy channel approach will map from the phoneme to the word level. Two well-known natural language processing approaches are adopted and derived from the noisy channel model theory: Joint-sequence models and statistical machine translation. Both of them are applied and various experiments are conducted using microphone and telephone of spontaneous speech.

While phrase-based statistical machine translation systems prefer to translate with longer phrases, this may cause errors in a free word order language, such as Japanese, in which the order of the arguments of the predicates is not solely determined by the predicates and the arguments can be placed quite freely in the text. In this paper, we propose to reorder the arguments but not the predicates in Japanese using a dependency structure as a kind of reordering. Instead of a single deterministically given permutation, we generate multiple reordered phrases for each sentence and translate them independently. Then we apply a re-ranking method using a discriminative approach by Ranking Support Vector Machines (SVM) to re-score the multiple reordered phrase translations. In our experiment with the travel domain corpus BTEC, we gain a 1.22% BLEU score improvement when only 1-best is used for re-ranking and 4.12% BLEU score improvement when n-best is used for Japanese-English translation.