Machine Translation Research Articles

History of the development of machine translation systems (from the early stages to the beginning of the 21st century)

The article examines certain periods of machine translation development and the methods used to create various machine translation systems. To ensure correct historical periodization, the author defines the criteria for classifying the development of machine translation. Special attention is paid to the development of individual machine translation systems, as well as the events that had a significant impact on both the development and the hindrance of progress in the field of machine translation. Additionally, the author considers the ALPAC report and its subsequent “quiet period” of machine translation development, analyses the key specialists who have been involved in the development and improvement of machine translation systems and machine translation technologies, categorizing them into four large groups. The article reviews the dominant basis of machine translation research until the late 1980s, rooted in various linguistic rules such as parsing rules, lexical rules, lexical transfer rules, syntactic generation rules, morphology rules, etc.The author traces the emergence of the corpus-based method in machine translation, as well as its further expansion through experiments in computational modelling of cognition and perception, including the study of parallel computing, neural networks and connectionism, which offered the prospect of making the systems “learn” from past successes and failures. Besides, the author also examines the emergence of the interlingualapproach for multilingual translation of technical manuals and the history of the “pivot language”, along with the use of SLLP (specialized languages for linguistic programming).

Error Analysis of Pretrained Language Models (PLMs) in English-to-Arabic Machine Translation

Advances in neural machine translation utilizing pretrained language models (PLMs) have shown promise in improving the translation quality between diverse languages. However, translation from English to languages with complex morphology, such as Arabic, remains challenging. This study investigated the prevailing error patterns of state-of-the-art PLMs when translating from English to Arabic across different text domains. Through empirical analysis using automatic metrics (chrF, BERTScore, COMET) and manual evaluation with the Multidimensional Quality Metrics (MQM) framework, we compared Google Translate and five PLMs (Helsinki, Marefa, Facebook, GPT-3.5-turbo, and GPT-4). Key findings provide valuable insights into current PLM limitations in handling aspects of Arabic grammar and vocabulary while also informing future improvements for advancing English–Arabic machine translation capabilities and accessibility.

Improving machine translation and post-editing for Chinese tourism texts using transformer-based models

As the digital age and globalization continue to evolve, the demand for accurate machine translation of tourism texts has increased substantially. This paper investigates how to improve the quality of machine translation (MT) and machine translation post-editing (MTPE) of Chinese tourism texts for non-native speakers. A review of the machine translation literature reveals a significant progression in translation methods from rule-based to corpus-based, statistical, and finally to the current neural machine translation (NMT) models. Despite its advanced capabilities, NMT requires large amounts of parallel data for training, which often presents challenges. This study proposes the use of Transformer-based models for MT and MTPE to improve translation quality. A dataset was curated from online sources, mainly Chinese tourism websites. The methodology involved pre-processing the data, performing machine translation using the Transformer model, and post-editing the results. The experiment demonstrated an increase in the BLEU score, suggesting an improvement in translation quality. However, challenges such as the handling of synonyms and geographical nouns were encountered, indicating the need for further research and model optimization.

Bibliometric analysis and visualization of translation assessment: Research theme, evolution and hotspots

Translation assessment, referring to evaluating various aspects related to translation, is crucial to the improvement of translation competence and quality. This study, taking valid papers on translation assessment published in WoS core collection from 2000 to 2022 as research samples, visualizes and reviews the research theme, research evolutions and emerging hot topics of translation assessment through bibliometric analysis. The number of annual publications shows that despite the general growth, the number of articles on translation assessment fluctuates from year to year. By Cite Space-based analysis of keywords co-occurrence, keyword clustering, time zone and burst detection, it finds that translation assessment mainly covers five themes, namely translation competence, translation quality, machine translation, translation teaching and training, and others. The overall evolution has been a gradual shift from concepts, metrics and evaluations of topics related to translation assessment into a topic of in-depth and interdisciplinary study. The hotspots in recent years include translation competence acquisition, neural machine translation and translation quality estimation. The findings reveal research evolutions and hot spots, propose implications for further translation assessment research and provide references for scholars who are interested in this topic.

A comparative analysis of lexical-based automatic evaluation metrics for different Indic language pairs

<p>With the rise of machine translation systems, it has become essential to evaluate the quality of translations produced by these systems. However, the existing evaluation metrics designed for English and other European languages may not always be suitable or apply to other Indic languages due to their complex morphology and syntax. Machine translation evaluation (MTE) is a process of assessing the quality and accuracy of the machine-translated text. MTE involves comparing the machine-translated output with the reference translation to calculate the level of similarity and correctness. Therefore, this study evaluates different metrics, namely, BLEU, METEOR, and TER to identify the most suitable evaluation metric for Indic languages. The study uses datasets for Indic languages and evaluates the metrics on various translation systems. The study contributes to the field of MT by providing insights into suitable evaluation metrics for Indic languages. This research paper aims to study and compare several lexical automatic machine translation evaluation metrics for Indic languages. For this research analysis, we have selected five language pairs of parallel corpora from the low-resource domain, such as English–Hindi, English-Punjabi, English-Gujarati, English-Marathi, and English-Bengali. All these languages belong to the Indo-Aryan language family and are resource-poor. A comparison of the state of art MT is presented and shows which translator works better on these language pairs. For this research work, the natural language toolkit tokenizers are used to assess the analysis of the experimental results. These results have been performed by taking two different datasets for all these language pairs using fully automatic MT evaluation metrics. The research study explores the effectiveness of these metrics in assessing the quality of machine translations between various Indic languages. Additionally, this dataset and analysis will make it easier to do future research in Indian MT evaluation.</p>

The impact of machine translation on the quality of translation with a sample study between English and Russian.

The impact of machine translation on the quality of translation with a sample study between English and Russian.

Student Performance in Machine Translation Training: Assessment and Exploration

The present research aims to conduct a process-oriented analysis to measure whether a group of graduate students enrolled in a translation course made steady progress in their performance of identifying machine translation (MT) errors and post-editing MT drafts of company web texts and news texts. A mixed methods approach consisting of quantitative and qualitative analyses was used. The findings show that there was a steady decline in the average number of MT errors that students could not spot or correctly identify in their three assignments. However, there was no significant improvement in student MTPE performance, with only a slight decrease in errors in the final MTPE assignment, which still remained worse than the first one. Finally, student responses in their reflection essays indicated that their reception of MT and MTPE had shifted from negative denial to positive acceptance. Overall, the findings of the present study reveal the need to extend the period of MTPE training for students. Incorporating MT training into the translation course has proven to be worthwhile for students, as it helps to dispel students’ previous misconception about MT and MTPE.

Performance of machine translators in translating French medical research abstracts to English: A comparative study of DeepL, Google Translate, and CUBBITT.

Non-English speaking researchers may find it difficult to write articles in English and may be tempted to use machine translators (MTs) to facilitate their task. We compared the performance of DeepL, Google Translate, and CUBBITT for the translation of abstracts from French to English. We selected ten abstracts published in 2021 in two high-impact bilingual medical journals (CMAJ and Canadian Family Physician) and used nine metrics of Recall-Oriented Understudy for Gisting Evaluation (ROUGE-1 recall/precision/F1-score, ROUGE-2 recall/precision/F1-score, and ROUGE-L recall/precision/F1-score) to evaluate the accuracy of the translation (scores ranging from zero to one [= maximum]). We also used the fluency score assigned by ten raters to evaluate the stylistic quality of the translation (ranging from ten [= incomprehensible] to fifty [= flawless English]). We used Kruskal-Wallis tests to compare the medians between the three MTs. For the human evaluation, we also examined the original English text. Differences in medians were not statistically significant for the nine metrics of ROUGE (medians: min-max = 0.5246-0.7392 for DeepL, 0.4634-0.7200 for Google Translate, 0.4815-0.7316 for CUBBITT, all p-values > 0.10). For the human evaluation, CUBBITT tended to score higher than DeepL, Google Translate, and the original English text (median = 43 for CUBBITT, vs. 39, 38, and 40, respectively, p-value = 0.003). The three MTs performed similarly when tested with ROUGE, but CUBBITT was slightly better than the other two using human evaluation. Although we only included abstracts and did not evaluate the time required for post-editing, we believe that French-speaking researchers could use DeepL, Google Translate, or CUBBITT when writing articles in English.

Unveiling the New Frontier: ChatGPT-3 Powered Translation for Arabic-English Language Pairs

This study evaluates the aptitude of ChatGPT for Arabic-English machine translation. The main objective of this research is to scrutinize the quality of ChatGPT's translations and compare its performance against machine translation systems, such as Google Translate, which are intricately tailored for translation purposes. In addition, the study seeks to investigate the potential integration of ChatGPT into translation workflows. Furthermore, we aspire to discern whether ChatGPT's translation efficacy harmonizes with or diverges from the profound finesse exhibited by human translation expertise. To accomplish this, a comparable corpus of 1000 English sentences and their corresponding Arabic translations was employed to evaluate the translation outputs of both machine translation systems alongside a human translation reference. The corpus was sourced from Tatoeba, an open online platform and underwent electronic assessment using the BLEU (Bilingual Evaluation Understudy) metric. The results indicate a marginal advantage of ChatGPT over Google Translate in delivering high-quality translations. Upon evaluating the corpus, we ascertain that ChatGPT performs impressively well compared to specialized translation systems like Google Translate. However, despite these promising findings, it is essential to acknowledge that even the most advanced machine translation technology, ChatGPT, cannot currently match the proficiency of human translation, at least not in the near future.

Evaluation of domain sulfur industry for DIA translator using bilingual evaluation understudy method

Evaluation is important part of our system development cycle; it also contributes to improving new machine translation (MT) technology optimum via comparing them with the traditional systems available to determine the weaknesses and the effectiveness to be improved in the proposed MT system. This work aiming to make a study that evaluate the performance and effectivness of the domain sulfur industry (DSI) for English-Arabic DIA translator quality. The recent study has conducted evaluating by making a comparison between this programme with the prominent Google translator through applying a rendering of 1,200 English sentences in bilingual evaluation understudy (BLUE) method. The obtain results show that the efficiency of Google translator is about 30.325%, while DIA translator efficiency in domain sulfur industry is about 73.325% and it’s more effective and give a better translation accuracy. The BLUE method efficiency is about (90.478%) compared with the human expert evaluator.

Spoken‐to‐written text conversion for enhancement of Korean–English readability and machine translation

AbstractThe Korean language has written (formal) and spoken (phonetic) forms that differ in their application, which can lead to confusion, especially when dealing with numbers and embedded Western words and phrases. This fact makes it difficult to automate Korean speech recognition models due to the need for a complete transcription training dataset. Because such datasets are frequently constructed using broadcast audio and their accompanying transcriptions, they do not follow a discrete rule‐based matching pattern. Furthermore, these mismatches are exacerbated over time due to changing tacit policies. To mitigate this problem, we introduce a data‐driven Korean spoken‐to‐written transcription conversion technique that enhances the automatic conversion of numbers and Western phrases to improve automatic translation model performance.

The translation teaching platform based on multilingual corpora of Xi Jinping: The Governance of China: Design, resources and applications

Over the last two decades, the domain of corpus-based translation studies has evolved into a significant paradigm, instrumental in probing a diverse array of scholarly inquiries. This article demonstrates the various uses of corpus analysis techniques that have gained popularity in translation studies, and it presents a new multilingual corpus. This corpus showcases the evolving nature of the field as it expands to accommodate new conceptual frameworks and methodologies. In addition to addressing the use of corpora in translating, learning, and analyzing translation within the context of multilingual corpora and machine translation, this article reports on a pilot project concerning the construction of an AI-based translation teaching platform. Utilizing multilingual corpora, the platform aims to create a foundation for extensive translation instruction and multilingual training at the tertiary level. Following a discussion on the evolution of corpus-based approaches in translation studies and their connection to the growth of translation education as a profession, the paper describes the actual design of the platform. This includes the theoretical framework, the methodology of annotation, and the execution of the software programme. These corpora are relevant not only to individuals studying translation but also to the related multilingual corpora, comprising collections of originals and their translations in the same language.

Non-Fluent Synthetic Target-Language Data Improve Neural Machine Translation.

When the amount of parallel sentences available to train a neural machine translation is scarce, a common practice is to generate new synthetic training samples from them. A number of approaches have been proposed to produce synthetic parallel sentences that are similar to those in the parallel data available. These approaches work under the assumption that non-fluent target-side synthetic training samples can be harmful and may deteriorate translation performance. Even so, in this paper we demonstrate that synthetic training samples with non-fluent target sentences can improve translation performance if they are used in a multilingual machine translation framework as if they were sentences in another language. We conducted experiments on ten low-resource and four high-resource translation tasks and found out that this simple approach consistently improves translation performance as compared to state-of-the-art methods for generating synthetic training samples similar to those found in corpora. Furthermore, this improvement is independent of the size of the original training corpus, the resulting systems are much more robust against domain shift and produce less hallucinations.

The Interaction Between Morphosyntactic Features and the Performance of Machine Translation Tools: The Case of Google Translate, Systran, and Microsoft Bing in English-Arabic Translation

The present study compares the performance of three machine translation tools in English-Arabic translation to answer the questions of (a) whether the three machine translation tools, Google Translate, Systran and Microsoft Bing can be ordered in a hierarchy of performance, and (b) whether they can handle lexically and structurally ambiguous sentences and garden path sentences. Using a number of constructed and selected English sentences, the morphosyntactic features of number, gender, case, definiteness, and humanness, agreement between cardinal numerals and their head nouns, lexically and structurally ambiguous sentences and garden path sentences are used to test the three machine translation tools for performance. The results show that (a) as far as handling the morphosyntactic features of subject-verb agreement in Standard Arabic, all three machine translation tools perform equally well, and no machine translation tool seems to perform significantly better than the other two (b) some marked features (e.g. dual number and humanness) of SA seem to pose some problems for machine translation tools, and (c) lexically and structurally ambiguous sentences and garden path sentences seem to be the most challenging sentences for the three machine translation tools.

Machine Translation and Natural Language Processing

Machine translation and natural language processing are powerful tools that allow for the automated translation of text from one language to another. Machine translation systems have evolved from rule-based systems to statistical and neural network-based approaches, with each approach having its own strengths and weaknesses. NLP, on the other hand, focuses on understanding the structure and meaning of natural language and includes tasks such as part-of-speech tagging, named entity recognition, and sentiment analysis. Thegoalofmachinetranslation,a subfield of natural language processing (NLP), is to create platform that translate text or speechautomatically.RBMT provides linguistic accuracy and control, but linguistic resources must be built and maintained meticulously. However, SMT makes use of large-scale parallel corpora to automatically learn translation patterns, making it adaptable to new language pairs and context-dependent translations.However, SMT may struggle with handling grammatical nuances and domain-specific vocabulary. Hybrid approaches combining the best of both RBMT andSMT and newer approaches like neural machine translation (NMT) are being explored to overcome the limitations and improve the quality of translation output.

GPT를 활용한 중-한 문학번역에 대한 고찰

This study investigates the use of advanced machine translation technologies like GPT in literary translation. It highlights the importance of conveying an author's intent and cultural nuances, a task traditionally reliant on human translators. The paper focuses on translating idioms from the Chinese novel “To Live(活著)” using GPT and traditional machine translation (MT), comparing their outputs. Findings indicate that GPT, through user interaction, offers a nuanced understanding of context and idioms, showcasing its effectiveness in handling complex expressions. The paper underscores the importance of appropriate prompt setting in AI translation, suggesting future research at the nexus of technology and literary expression. It concludes that the quest for more precise and nuanced translations will advance through a synergistic relationship between AI and human translators.

German-to-Basque Translation Analysis of Multiword Expressions in a Learner Translation Corpus

This paper presents the results of the annotation in a learner translation corpus consisting of German source texts and student translations into Basque. The analysis was carried out with the purpose of identifying trainee translators’ strengths and weaknesses when translating multiword expressions, such as compounds, collocations, and idioms. The data comprised eight German source texts and sixty-eight Basque translations from undergraduate students enrolled at the University of the Basque Country. From the total number of annotations (1214), which include not only errors but also cases of interference from the source language and positive outcomes, around 27 % are related to multiword expressions. The results of the translation analysis show that there are variables — such as the use of machine translation systems, the level of specialisation of the source text, the type of multiword expression to be translated or the absence of a literal counterpart in the target language — that may affect the translation of such units and lead to erroneous solutions and/or interference in the outputs produced by the trainee translators. From a pedagogical point of view, these findings will have a direct impact on the translation classes and will be very valuable for designing corpus-based in-class activities

Cultural Adaptation of Recipes

Abstract Building upon the considerable advances in Large Language Models (LLMs), we are now equipped to address more sophisticated tasks demanding a nuanced understanding of cross-cultural contexts. A key example is recipe adaptation, which goes beyond simple translation to include a grasp of ingredients, culinary techniques, and dietary preferences specific to a given culture. We introduce a new task involving the translation and cultural adaptation of recipes between Chinese- and English-speaking cuisines. To support this investigation, we present CulturalRecipes, a unique dataset composed of automatically paired recipes written in Mandarin Chinese and English. This dataset is further enriched with a human-written and curated test set. In this intricate task of cross-cultural recipe adaptation, we evaluate the performance of various methods, including GPT-4 and other LLMs, traditional machine translation, and information retrieval techniques. Our comprehensive analysis includes both automatic and human evaluation metrics. While GPT-4 exhibits impressive abilities in adapting Chinese recipes into English, it still lags behind human expertise when translating English recipes into Chinese. This underscores the multifaceted nature of cultural adaptations. We anticipate that these insights will significantly contribute to future research on culturally aware language models and their practical application in culturally diverse contexts.

Theorizing the Role of Machine Translation in L2 Reading Comprehension: Validating the Theory and Exploring the Learning Potential through MT Use

Theorizing the Role of Machine Translation in L2 Reading Comprehension: Validating the Theory and Exploring the Learning Potential through MT Use

When minoritized languages encounter MT

Machine translation (MT) is improving even for low-resource minoritized languages such as Basque, for which free online engines are available. However, the level of adoption and common practices involving the technology are unknown, even though it has the potential to disrupt a carefully planned Basque language revitalization and sustainability process. To shed light on MT usage habits and perceptions among the Basque community, we report on the results of a survey of language specialists and general users, and a focus group with professional translators and interpreters. The data shows that MT is already becoming more popular among users of all backgrounds and that, overall, the attitude towards the technology is positive, which might result in increased use in the future. However, participants express concern about the impact MT will have on the development of Basque. The results call for further research on the language impact of MT and MT literacy initiatives.