Stochastic finite-state models for spoken language machine translation

Machine translation is used to implement the translation between different languages. Neural machine translation is one of the most popular machine translation methods which have made a great progress in recent years especially on universal languages. However, domestic translation software for non-universal languages is limited and also needs improving. In this paper, we carry out a practice on neural machine translation from Indonesian to Chinese. We first build a bilingual corpus of Indonesian and Chinese. After that, we train translation models using neural machine translation method, with traditional statistical machine translation models as baselines. In the later stages of our practice, we develop a Web software, named as Lore Translator, basing on our so-far best translation model. The performance of our model is comparable to previous work.

Problem statement: Any Arabic to English Machine Translation (MT) system should be capable of dealing with word order which Arabic exhibits. This poses a significant challenge to MT due to the vast number of ways to express the same sentence in Arabic. The ordering features are very important and should be carefully applied to ensure the generation of sentence in the target language. Because they apply to the target language, it shoul d fulfill the specific requirement of this language . Mistakes in the MT output can be either the result of analysis problems at the source language level, or due to generation problem at target language level. Word order rules are crucial for the generation of sentences in the target language. They also serve a s rules for the ordering of sentence constituents. These rules draw their information from the syntact ic knowledge. The word order problem becomes more obvious when making machine translation between languages that have rich morphological variations. Approach: The main objective of this research is to develop a machine translation that translates Arabic noun phrases into English by usin g transfer-based approach. A system called Npae- Rbmt has been developed in this research. Transfer- based machine translation is one instance of rule- based machine-translation approaches and is current ly one of the most widely used methods of machine translation. The idea of transfer-based mac hine translation it is necessary to have an intermediate representation that captures the mean ing of the original sentence in order to generate the correct translation. Using advantages of transf er-based machine translation such as analysis step, the Transfer-based becomes simpler as linguistic an alysis goes deeper-as the representation of analysi s step becomes more abstract. In fact, a major goal o f MT research is to define a level of analysis whic h is so deep in which transfer-based machine translat ion is able to do. Results: The method was tested on 88 thesis titles and journals from the computer science domain. The accuracy of the result was 94.6%. These results proved the viability of this a pproach for distant languages. Conclusion: Based on the achieved results, we have managed to perform th e syntactic reordering within an Arabic noun phrases to English translation task by using transf er-based machine translation and also achieved reasonable improvements in translation quality over related approach.

Machine translation is a computer-based translation process that receives a set of words from a particular humanreadable language as input and outputs a second set of words in the intended human-readable language. Machine translation has assisted linguists and sociologists all across the world. A machine translation model may be statistical, rule-based, or based on neural networks. The Neural Machine Translation (NMT) model was introduced in response to the numerous drawbacks of rule- based and statistical-based machine translation models. This study adds a parallel corpus of Assamese and English and develops the NMT system, a transformer model with a self- attention mechanism. For translation from English to Assamese, the system received a higher bilingual evaluation understudy (BLEU) score of 7.6, and for translation from Assamese to English,a BLEU score of 23.2.

This work aims to present an overview of using some artificial intelligence (AI) models in machine translation (MT). This work aims to integrate machine learning (ML), deep learning (DL), large language models (LLMs) to enhance machine translation between natural languages. The focus is directed to present the neural-based machine translation (NMT), and some DL models are presented. The bidirectional-encoder-representation from transformer (BERT) and LLMs are presented to utilize the big amount of textual data to learn translation patterns. The main measurable criteria that are used to evaluate the performance of MT and Arabic machine translation (AMT) are also presented. Some linguistic and technical challenges of MT and AMT are discussed. Some key points of future works in NMT are mentioned. A comparative study among some recent published related works is presented. A critical survey is presented to show the important role of DL and LLMs in MT. Some open-source toolkits, datasets and some commercial MT systems are collected and briefly presented. This work is expected to be useful for those people interested to know the up-to-date knowledge of MT and the potential role of DL and LLM in automatic translation.

This paper presents a practical machine translation system based on sentence types for economic news stories.Conventional English-to-Japanese machine translation (MT) systems which are rule-based approaches, are difficult to translate certain types of Associated Press (AP) wire service news stories, such as economics and sports, because these topics include many fixed expressions (such as compound words or collocations) which are difficult to be processed by conventional syntactic analysis and/or word selection methods.The proposed MT system, an economic-news stories machine translation system (ENTS), can translate economic news sentences with fixed expressions. The system consists of three processes, to handle different types of sentences, fixed type, economics-specific type and general type. This paper focuses mainly on the translation method for fixed-type sentences, which is a kind of example-based approach. In this translation method, fixed sentence translation (STRA) data plays a key role. The STRA data is a set of bilingual templates, which is built automatically from fixed English sentences and their Japanese translation equivalents. The fixed English sentences are extracted automatically from the AP corpus.A series of experiments to evaluate ENTS using economic news in the AP news stories showed the translation accuracy was about 50% higher than with our conventional rule-based MT method.

In this study, we explore the implications of dataset limitations in semantic knowledge-driven machine translation (MT) for intelligent virtual assistants (IVA). Our approach diverges from traditional single-best translation techniques, utilizing a multi-variant MT method that generates multiple valid translations per input sentence through a constrained beam search. This method extends beyond the typical constraints of specific verb ontologies, embedding within a broader semantic knowledge framework. We evaluate the performance of multi-variant MT models in translating training sets for Natural Language Understanding (NLU) models. These models are applied to semantically diverse datasets, including a detailed evaluation using the standard MultiATIS++ dataset. The results from this evaluation indicate that while multi-variant MT method is promising, its impact on improving intent classification (IC) accuracy is limited when applied to conventional datasets such as MultiATIS++. However, our findings underscore that the effectiveness of multi-variant translation is closely associated with the diversity and suitability of the datasets utilized. Finally, we provide an in-depth analysis focused on generating variant-aware NLU datasets. This analysis aims to offer guidance on enhancing NLU models through semantically rich and variant-sensitive datasets, maximizing the advantages of multi-variant MT.

This paper proposes an English-Chinese machine translation research method based on transfer learning. First, it expounds the theory of neural machine translation and transfer learning and related technologies. Neural machine translation is discussed, the advantages and disadvantages of various models are introduced, and the transformer neural machine translation model framework is selected. For low-resource Chinese-English parallel corpus and Tibetan-Chinese parallel corpus, 30 million Chinese-English parallel corpora, 100,000 Chinese-English low-resource parallel corpora, and 100,000 Tibetan-Chinese parallel corpora were used to pretrain the transformer machine translation architecture. The decoders are all composed of 6 identical hidden layers, the initialization of the model parameters is done by the transformer uniform distribution, and the model training uses Adam as the optimizer. In the model transfer part, the parameters with the better effect of the pretrained model are transferred to the low-resource Chinese-English and Tibetan-Chinese machine translation model training, so as to achieve the purpose of knowledge transfer. The results show that the model transfer learning of low-resource Chinese-English parallel corpus improves the translation system's translation by 3.97 BLEU values compared with the translation system without transfer learning at 0.34 BLEU values. Model transfer learning on low-resource Tibetan-Chinese parallel corpus increases the BLEU value by 2.64 BLEU compared to the translation system without transfer learning. The neural machine translation system that uses BPE technology for preprocessing plus model transfer learning is compared to the translation system that only performs transfer learning and shows an improved 0.26 BLEU value.It is verified that the transfer learning method proposed in this paper has a certain improvement in the effect of low-resource Chinese-English and Tibetan-Chinese neural machine translation models.

While various neural machine translation (NMT) methods have integrated multiple prior knowledge to guide the translation, no research is available on combining with source-target bilingual translation template. In this paper, we firstly propose a maximal-length noun phrase template (MNP-Template), which constructs a novel translation template focusing on the constituency syntactic structure. Secondly, building on the multi-source transformer framework, we design a template-based machine translation (TBMT) model to integrate the syntactic knowledge of the retrieved target template similar to the ground-truth translation in the NMT decoder. Experiment results show the effectiveness of MNP- Template and TBMT on test subsets filtered by the fuzzy match score. Moreover, our method achieves significant improvement in out-of-domain test sets, which well-validated the university across diverse domains.

Background: Questionnaires are forward-translated and then back-translated by human translators during adaptation of a survey questionnaire in the local language. The machine translation method is now available free from “Google Translate” website and it can be used for translation from English to Bengali and vice versa. Using machine translation method may save researchers from investing money and time in translation by human experts. Aim: The aim of this pilot study was to test the feasibility of “Google Translate” in forward-translation and back-translation during adaptation of a survey questionnaire from English to Bengali. Materials and Methods: A survey questionnaire, originally developed in English, was forward-translated to Bengali and then back-translated to English by machine translation method and manual translation method. Both the versions were checked for errors expressed in subjective sentence error rate (SSER). Back-translated English questionnaire obtained from machine translation and manual translation was checked for equivalence with the original questionnaire. Results: SSER for forward-translated Bengali questionnaire by machine translation method was 66.60 ± 4.04 and manual translation method was 117.20 ± 1.30 (P

Template-based Machine Translation model has been adopted in practical translation systems. In order to build a practical Chinese-Japanese Machine Translation system, we continue to extend Template-based Model. In order to make Template-based Model more practical and specific, original Template is divided into Template for sentences and Template for phrases to build nested Template-based Machine Translation Model. The existing scope and definition of the variable for original Template-based Model is also extended. From the experiment results, Template-based model is not only effective but also feasible and with considerable precision for Chinese Japanese Machine Translation. The Template-based model can even translate long sentences with complex structure.

While multilingual training is now an essential ingredient in machine translation (MT) systems, recent work has demonstrated that it has different effects in different multilingual settings, such as many-to-one, one-to-many, and many-to-many learning. These training settings expose the encoder and the decoder in a machine translation model with different data distributions. In this paper, we examine how different varieties of multilingual training contribute to learning these two components of the MT model. Specifically, we compare bilingual models with encoders and/or decoders initialized by multilingual training. We show that multilingual training is beneficial to encoders in general, while it only benefits decoders for low-resource languages (LRLs). We further find the important attention heads for each language pair and compare their correlations during inference. Our analysis sheds light on how multilingual translation models work and enables us to propose methods to improve performance by training with highly related languages. Our many-to-one models for high-resource languages and one-to-many models for LRL outperform the best results reported by Aharoni et al. (2019)

Classifiers in natural language processing (NLP) often have a large number of output classes. For example, neural language models (LMs) and machine translation (MT) models both predict tokens from a vocabulary of thousands. The Softmax output layer of these models typically receives as input a dense feature representation, which has much lower dimensionality than the output. In theory, the result is some words may be impossible to be predicted via argmax, irrespective of input features, and empirically, there is evidence this happens in small language models (Demeter et al., 2020). In this paper we ask whether it can happen in practical large language models and translation models. To do so, we develop algorithms to detect such unargmaxable tokens in public models. We find that 13 out of 150 models do indeed have such tokens; however, they are very infrequent and unlikely to impact model quality. We release our algorithms and code so that others can test their models.

This paper presents example-based machine translation (MT) based on syntactic transfer, which selects the best translation by using models of statistical machine translation. Example-based MT sometimes generates invalid translations because it selects similar examples to the input sentence based only on source language similarity. The method proposed in this paper selects the best translation by using a language model and a translation model in the same manner as statistical MT, and it can improve MT quality over that of 'pure' example-based MT. A feature of this method is that the statistical models are applied after word re-ordering is achieved by syntactic transfer. This implies that MT quality is maintained even when we only apply a lexicon model as the translation model. In addition, translation speed is improved by bottom-up generation, which utilizes the tree structure that is output from the syntactic transfer.

In order to solve the problems of time-consuming and frequent mistranslations in traditional translation methods, this study designed an English-Chinese machine translation method based on transfer learning. On the basis of analyzing the basic principles and specific strategies of English-Chinese machine translation, the traditional neural machine translation methods are analyzed, and then the translation process is optimized by transfer learning. On the basis of preprocessing English-Chinese translation text data, features of English-Chinese translation text are extracted, features of English-Chinese translation text are rapidly classified by feature transfer learning, and machine models of English-Chinese translation are constructed based on the classification results. The objective of feature transfer learning is to reuse the past knowledge obtained in the form of dataset to be utilized for another target data. The findings of the experiment show the effectiveness of the proposed method in achieving the design expectation. The benefit of the proposed method includes a short translation time and fewer mistranslations.

In recent years, with the development of deep learning, machine translation using neural network has gradually become the mainstream method in industry and academia. The existing Chinese-English machine translation models generally adopt the deep neural network architecture based on attention mechanism. However, it is still a challenging problem to model short and long sequences simultaneously. Therefore, a bidirectional LSTM model integrating attention mechanism is proposed. Firstly, by using the word vector as the input data of the translation model, the linguistic symbols used in the translation process are mathematized. Secondly, two attention mechanisms are designed: local attention mechanism and global attention mechanism. The local attention mechanism is mainly used to learn which words or phrases in the input sequence are more important for modeling, while the global attention mechanism is used to learn which layer of expression vector in the input sequence is more critical. Bidirectional LSTM can better fuse the feature information in the input sequence, while bidirectional LSTM with attention mechanism can simultaneously model short and long sequences. The experimental results show that compared with many existing translation models, the bidirectional LSTM model with attention mechanism can effectively improve the quality of machine translation.