Parallel Corpus Research Articles

A neural machine translation system for Kreol Repiblik Moris and English

Although Google Translate is a widely used machine translation service that supports 133 languages, it does not incorporate support for the Kreol Repiblik Moris (KRM) language. Addressing this limitation, the current research focuses on enhancing the accuracy and fluency of machine translation between KRM and English through natural language processing and deep neural machine translation techniques. In this study, a machine translation system using a transformer model trained with a dataset of 50,000 parallel corpora has been developed. The model was evaluated using manual translations and the bilingual evaluation understudy (BLEU) score. A score of 31.46 for translating from KRM to English and 28.15 for translating from English to KRM was achieved. To our knowledge, these are the highest BLEU scores for translation between these two languages. This is due to utilising the largest dataset and extensive atomic words from the KRM dictionary. This successful interdisciplinary funded project led to the setting up of a free online translation service and a smartphone app for Mauritian citizens and tourists.

A Corpus-Based Analysis of Pauses in Chinese-English Consecutive Interpreting of Chinese English Majors

This study aims to investigate pause characteristics and probe into the main reason for unnatural pauses in Chinese-English consecutive interpreting. The data were from the Parallel Corpus of Chinese EFL Learners-Spoken (PACCEL-S), which includes phonetic materials in the interpretation part of TEM-8. Forty interpreters were divided into four groups according to their scores and gender. Their recordings and transcribed texts were annotated and analyzed to provide a comprehensive overview of pausing. The findings are: (1) Pauses often occur in C-E CI. Interpreters use more UPs and unnatural pauses. (2) There is no interaction between proficiency and gender in pause duration and frequency, as well as the two indicators of unnatural pauses. (3) UP duration is affected by proficiency, and females tend to produce more FPs and unnatural pauses. (4) FP and UP invariably occur together but do not show linear correlation. (5) Interpreters tend to pause before content words, and unnatural pauses are mainly induced by notional word retrieval difficulty. By discussing pauses and explaining the leading motivation for unnatural pauses from the perspective of lexical retrieval, this study informs readers of the nature and features of pauses and provides suggestions for stakeholders.

Auxiliaries in Old Dutch

Abstract This study explores the use of auxiliaries in the oldest text available for Old Dutch, the Wachtendonck Psalter, dating from the 10th century. Our aim is to understand why there are so few different auxiliaries in this text in comparison to other texts in Old Dutch. We tackle this question by taking a historical comparative perspective, using methodological insights and techniques from corpus-based contrastive linguistics and typology. More specifically, we build a diachronic parallel corpus of psalm translations and compare the contexts in which auxiliaries and inflectional alternatives are used in these parallel texts by means of multidimensional scaling. Our historical comparative method results in five proximity maps which allow us to explore and compare the inventory of verb constructions of the Wachtendonck Psalter both retrospectively, with its source text in Latin, and prospectively, with later translations in Dutch. Our analysis examines the role of grammaticalization as well as the specific nature of the text as an interlinear translation as possible motivations for the presence and absence of auxiliaries in the Wachtendonck Psalter.

Translation from Tunisian Dialect to Modern Standard Arabic: Exploring Finite-State Transducers and Sequence-to-Sequence Transformer Approaches

Translation from the mother tongue, including the Tunisian dialect, to modern standard Arabic is a highly significant field in natural language processing due to its wide range of applications and associated benefits. Recently, researchers have shown increased interest in the Tunisian dialect, primarily driven by the massive volume of content generated spontaneously by Tunisians on social media following the revolution. This article presents two distinct translators for converting the Tunisian dialect into Modern Standard Arabic. The first translator utilizes a rule-based approach, employing a collection of finite state transducers and a bilingual dictionary derived from the study corpus. On the other hand, the second translator relies on deep learning models, specifically the sequence-to-sequence transformer model and a parallel corpus. To assess, evaluate, and compare the performance of the two translators, we conducted tests using a parallel corpus comprising 8,599 words. The results achieved by both translators are noteworthy. The translator based on finite state transducers achieved a BLEU score of 56.65, while the transformer model-based translator achieved a higher score of 66.07.

Saudi Learner Translation Corpus: The design and compilation of an English-Arabic learner translation corpus.

This article introduces the Saudi Learner Translation Corpus (SauLTC), an innovative multi-version English-Arabic parallel corpus featuring part-of-speech tagging. We describe the corpus parameters and compilation process and explain how textual processing and sentence alignment are conducted. The participants include 366 student translators, 48 instructors, and 23 alignment verifiers. The corpus provides access to two target versions of every ST to allow the detection of the changes in the translation and revision processes from the initial to the final draft. The translations were collected over three years, yielding 5,160,386 tokens. The metadata of 23 sentence alignment verifiers were added to the analysis as a unique variable to investigate individual differences in the manual verification process. This unidirectional corpus can be used to identify student translators' strategies and errors in translation and analyze the efficacy of instructors' feedback. Furthermore, it is accessible via an application and a website. It provides translation teachers and researchers with a database that can help develop corpus-based and corpus-driven teaching materials.

Tickling translations: Small but mighty open-sourced transformers bring English PUN-ny entities to life in French!

Recent advancements in transformer-based language models have demonstrated substantial progress in producing good translations. Despite these achievements, challenges persist in translating playful requests, especially when users intentionally introduce humor. Deciphering the hidden pun among such playful requests is one of the major difficulties for modern language models, which causes user dissatisfaction. This paper targets a specific niche of humor translation, which is the translation of English-named entities containing puns into French using small-scale open-sourced transformer models. The transformer architecture serves as a foundation for popular language models like chatGPT. It allows learning long-range contextual relationships within sequences. The main novelty of the paper is the proposed extractive question/answering (Q/A) styled technique based on the transformers to find relevant translations for the provided English nouns using the openly available parallel corpora. To evaluate the effectiveness of our method, we utilize a dataset provided by the JOKER CLEF automatic pun and humor translation 2022 team. The dataset contains single-word nouns from popular novels, anime, movies, and games, each containing a pun. The discussed methodology and experimental framework are adaptable and can be extended to any language pair for which an open, available parallel corpus exists. This flexibility underscores the broader applicability of our findings and suggests the potential for enhancing humor translation across various language combinations.

A Transformer-Based Yoruba to English Machine Translation (TYEMT) System with Rouge Score

Automated translation systems for some indigenous Nigerian languages like the Yoruba, have historically been limited by the lack of large, high- quality bilingual text and effective approaches to modeling. This paper presents introduces an approach to bi-directional Yoruba-English text-to-text machine translation utilizing deep learning technique, specifically Transformer models. Transformer models, which utilizes self-attention mechanisms to improve translation quality and efficiency. The system was trained and evaluated on a newly curated Yoruba- English parallel corpus, which significantly augments existing resources. Experimental results demonstrate that the Transformer-based model performs translation accurately and fluently, achieving a ROUGE (Recall-Oriented Understudy for Gisting Evaluation) score improvement of 0.4649. This work not only advances the frontiers of Yoruba-English machine translation but also enriches a wider domain in the field of multilingual Natural Language processing (NLP) by addressing challenges associated with translating between languages with limited resources. Future studies include enhancing the available parallel corpus and exploring hybrid models that combine the strengths of both RNN and Transformer architectures.

Parallel corpus in analysing Czech spoken expressions and their equivalents in English, French, and Polish

Parallel corpus in analysing Czech spoken expressions and their equivalents in English, French, and Polish

Subtitling profanity into Malay: Balancing formal and creative taboo expressions

Introduction: Translation of swearwords has unceasingly gained traction in audiovisual translation. In subtitling, swearing does not only warrant linguistic consideration but other multimodal elements, such as intonation and body movement. Apart from subtitling technicalities, the difficulties of swearword translation are also impeded by their explicit and implicit forms, as well as the linguistic-cultural differences in source-target texts. This study aims to deliberate on the Malay translation of English swearwords of ‘fuck’ and ‘shit’ in two episodes of Netflix TV series The History of Swearwords.Methods: To this end, it employs a comparative and descriptive analysis by developing a parallel corpus by utilising AntConc and AntPConc software. The selection of Netflix posits no local law or policy restrictions and censorship.Results: The comparative analysis found a linkup between creative and formal subtitling of the swearwords, which signals a prevalent pragmatic and cultural foundation in the translation, especially vis-à-vis the deletion and neutralisation procedures. The study also raises other notable concerns, such as the identification and semantic comprehension of swearwords by the series subtitlers.Conclusions: Overall, the study posits that the socio-pragmatic factors, among others are still central for the English-Malay translation strategies of the swearwords in maintaining both linguistic accuracy and cultural appropriateness for target audience

Exploring Large Language Models to generate Easy to Read content

Ensuring text accessibility and understandability are essential goals, particularly for individuals with cognitive impairments and intellectual disabilities, who encounter problems accessing information across various mediums such as web pages, newspapers, online administrative tasks, or health documents. Initiatives like Easy to Read and Plain Language guidelines aim to simplify complex texts; however, standardizing these guidelines remains challenging and often involves manual processes. This work presents an exploratory investigation into leveraging Artificial Intelligence (AI) and Natural Language Processing (NLP) approaches to simplify Spanish texts into Easy to Read formats, with a focus on utilizing Large Language Models (LLMs) for creating accessible texts, especially in generating Easy to Read content. The study contributes a parallel corpus of Spanish adapted for Easy To Read format, which serves as a valuable resource for training and testing text simplification systems. Additionally, several text simplification experiments using LLMs and the collected corpus are conducted, involving fine-tuning and testing a Llama2 model to generate Easy to Read content. A qualitative evaluation, guided by an expert in text adaptation for Easy to Read content, is carried out to assess the automatically simplified texts. This research contributes to advancing text accessibility for individuals with cognitive impairments, highlighting promising strategies for leveraging LLMs while responsibly managing energy usage.

C anton MT: Investigating Back-Translation and Model-Switch Mechanisms for Cantonese-English Neural Machine Translation

This paper investigates the development and evaluation of machine translation models from Cantonese to English (and backward), where we propose a novel approach to tackle low-resource language translations. Despite recent improvements in Neural Machine Translation (NMT) models with Transformer-based architectures, Cantonese, a language with over 80 million native speakers, has below-par State-of-the-art commercial translation models due to a lack of resources. The main objectives of the study are to develop a model that can effectively translate Cantonese to English and evaluate it against state-of-the-art commercial models. To achieve this, a new parallel corpus has been created by combining different available corpora online with preprocessing and cleaning. In addition, a monolingual Cantonese dataset has been created through web scraping to aid the synthetic parallel corpus generation. Following the data collection process, several approaches, including fine-tuning models, back-translation , and model switch , have been used. The translation quality of models has been evaluated with multiple quality metrics, including lexicon-based metrics (SacreBLEU and hLEPOR) and embedding-space metrics (COMET and BERTscore). Based on the automatic metrics, the best model is selected and compared against the 2 best commercial translators using a new human evaluation framework HOPES . The best model proposed in this investigation (NLLB-mBART) with model switch mechanisms has reached comparable and even better automatic evaluation scores against State-of-the-art commercial models (Bing and Baidu Translators), with a SacreBLEU score of 16.8 on our test set. Furthermore, an open-source web application has been developed to allow users to translate between Cantonese and English, with the different trained models available for effective comparisons between models from this investigation and users. CantonMT is available at https://github.com/kenrickkung/CantoneseTranslation

Phrasem-Konstruktionen im Deutschen und Tschechischen. Eine Fallstudie zu der Phrasem-Konstruktion Es ist zum + Vinf [!] (substantivierter Verbinfinitiv)

The empirical study presents the problem of phraseme constructions (PhraCon) in current Germanic and international contrastive phraseological research. The article analyses the phraseme construction Es ist zum + Vinf [!] (substantivised verb infinitive). Es ist zum Heulen / Verzweifeln / Verrücktwerden / Lachen etc. will be focussed on from a contrastive German-Czech perspective with regard to its form and function in the text as well as its translation possibilities. Evidence from the parallel corpus InterCorp of the Czech National Corpus (CNK) will be analysed comparatively.

Constructing and Augmenting a Bidirectional Paraphrases Dataset from an English-Arabic Subtitling Parallel Corpus

Constructing and Augmenting a Bidirectional Paraphrases Dataset from an English-Arabic Subtitling Parallel Corpus

Study on the Corresponding Chinese Expressions of Korean Honorific Expressions ‘-nim/nim’ : Based on the Drama Parallel Corpus

Study on the Corresponding Chinese Expressions of Korean Honorific Expressions ‘-nim/nim’ : Based on the Drama Parallel Corpus

Indonesian to Bengkulu Malay Statistical Machine Translation System

Machine translation is an automatic tool that can process language translation from one language to another. This research focuses on developing Statistical Machine Translation (SMT) from Indonesian to Bengkulu Malay and evaluating the quality of the machine translation output. The training and testing data consist of parallel corpora taken from Bengkulu Malay dictionaries and online resources for Indonesian corpora, with a total of 5261 parallel sentence pairs. Several steps are performed in SMT. The initial step is preprocessing, aimed at preparing the parallel corpus. After that, a training phase is conducted, where the parallel corpus is processed to build language and translation models. Subsequently, a testing phase is carried out, followed by an evaluation phase. Based on the research results, various factors influence the quality of SMT translation output. The most important factor is the quantity and quality of the parallel corpus used as the foundation for developing translation and language models. The machine translation output is automatically evaluated using the Bilingual Evaluation Understudy (BLEU), indicating accuracy values observed when using 500 sentences, 1500 sentences, 2500 sentences, 4000 sentences, and 5261 sentences are 80.56%, 90.86%, 92.50%, 92.91%, and 94.48% respectively.

Harmony search for hyperparameters optimization of a low resource language transformer model trained with a novel parallel corpus Ocelotl Nahuatl – Spanish

Nahuatl, a low-resource language, does not have an online translator application. Instead, resources are limited to dictionaries, web pages, or digital books. Given this condition, it is vital to provide as much support to the language as possible. This research aims to enhance the BLEU score in machine translation by applying the harmony search heuristic method to state-of-the-art transformers models. This is conducted by finding the optimal hyperparameter settings for the models. Models are trained and tested using a fresh moderate-size parallel corpus of 1.5k phrases. By utilizing harmony search, the study shows an improvement in the BLEU score, enhancing it by 2.569%. In order to accomplish this, various factors related to the hyperparameters need to be considered. The application of harmony search with transformers can be extended to various parallel corpora or models, taking these considerations into account.

Efficient incremental training using a novel NMT-SMT hybrid framework for translation of low-resource languages.

The data-hungry statistical machine translation (SMT) and neural machine translation (NMT) models offer state-of-the-art results for languages with abundant data resources. However, extensive research is imperative to make these models perform equally well for low-resource languages. This paper proposes a novel approach to integrate the best features of the NMT and SMT systems for improved translation performance of low-resource English-Tamil language pair. The suboptimal NMT model trained with the small parallel corpus translates the monolingual corpus and selects only the best translations, to retrain itself in the next iteration. The proposed method employs the SMT phrase-pair table to determine the best translations, based on the maximum match between the words of the phrase-pair dictionary and each of the individual translations. This repeating cycle of translation and retraining generates a large quasi-parallel corpus, thus making the NMT model more powerful. SMT-integrated incremental training demonstrates a substantial difference in translation performance as compared to the existing approaches for incremental training. The model is strengthened further by adopting a beam search decoding strategy to produce k best possible translations for each input sentence. Empirical findings prove that the proposed model with BLEU scores of 19.56 and 23.49 outperforms the baseline NMT with scores 11.06 and 17.06 for Eng-to-Tam and Tam-to-Eng translations, respectively. METEOR score evaluation further corroborates these results, proving the supremacy of the proposed model.

Research on the Construction of a Sichuan Intangible Cultural Heritage Archive Chinese-English Parallel Corpus

This study aims to construct a Chinese-English parallel corpus of Sichuan's intangible cultural heritage (ICH) archives to facilitate the protection, inheritance, and international exchange of ICH. Through field research, literature review, and the application of digital technology, this project analyzes the current state of Sichuan's ICH archives, discusses the construction plan for the Chinese-English corpus, and proposes strategies to enhance the internationalization level of ICH inheritance and exchange. The research results indicate that the construction of the Chinese-English corpus is of great significance for the digital protection of ICH resources, optimization of translation practices, and promotion of cultural heritage.

Health text simplification: An annotated corpus for digestive cancer education and novel strategies for reinforcement learning

Objective:The reading level of health educational materials significantly influences the understandability and accessibility of the information, particularly for minoritized populations. Many patient educational resources surpass widely accepted standards for reading level and complexity. There is a critical need for high-performing text simplification models for health information to enhance dissemination and literacy. This need is particularly acute in cancer education, where effective prevention and screening education can substantially reduce morbidity and mortality. Methods:We introduce Simplified Digestive Cancer (SimpleDC), a parallel corpus of cancer education materials tailored for health text simplification research, comprising educational content from the American Cancer Society, Centers for Disease Control and Prevention, and National Cancer Institute. The corpus includes 31 web pages with the corresponding manually simplified versions. It consists of 1183 annotated sentence pairs (361 train, 294 development, and 528 test). Utilizing SimpleDC and the existing Med-EASi corpus, we explore Large Language Model (LLM)-based simplification methods, including fine-tuning, reinforcement learning (RL), reinforcement learning with human feedback (RLHF), domain adaptation, and prompt-based approaches. Our experimentation encompasses Llama 2, Llama 3, and GPT-4. We introduce a novel RLHF reward function featuring a lightweight model adept at distinguishing between original and simplified texts when enables training on unlabeled data. Results:Fine-tuned Llama models demonstrated high performance across various metrics. Our RLHF reward function outperformed existing RL text simplification reward functions. The results underscore that RL/RLHF can achieve performance comparable to fine-tuning and improve the performance of fine-tuned models. Additionally, these methods effectively adapt out-of-domain text simplification models to a target domain. The best-performing RL-enhanced Llama models outperformed GPT-4 in both automatic metrics and manual evaluation by subject matter experts. Conclusion:The newly developed SimpleDC corpus will serve as a valuable asset to the research community, particularly in patient education simplification. The RL/RLHF methodologies presented herein enable effective training of simplification models on unlabeled text and the utilization of out-of-domain simplification corpora.

Verb influence on French wh-placement: a parallel corpus study

Abstract Our study investigates the effect of French verb lemmata on the preverbal (QV) or postverbal (VQ) positioning of interrogative forms equivalent to English ‘what’ (que, quoi, and related forms) within a French–Spanish parallel corpus of subtitles. We highlight and illustrate the corpus’s utility for studying less frequent verbs in combination with specific wh-forms. Our findings suggest that less frequent French verbs exhibit weaker associations with QV compared to their more frequent counterparts. A post-hoc study using Spanish translations reveals that French verbs correlated with QV often denote observable actions involving directly accessible Q-referents. We hypothesise that queries concerning ‘situationally accessible’ referents are predominantly utilised for non-standard, evaluative, or challenging questions, which are typically QV in French.