Translation System Research Articles

Advanced Hybrid Neural Networks for Accurate Recognition of the Extended Alphabet and Dynamic Signs in Mexican Sign Language (MSL)

The Mexican deaf community primarily uses Mexican Sign Language (MSL) for communication, but significant barriers arise when interacting with hearing individuals unfamiliar with the language. Learning MSL requires a substantial commitment of at least 18 months, which is often impractical for many hearing people. To address this gap, we present an MSL-to-Spanish translation system that facilitates communication through a spelling-based approach, enabling deaf individuals to convey any idea while simplifying the AI’s task by limiting the number of signs to be recognized. Unlike previous systems that focus exclusively on static signs for individual letters, our solution incorporates dynamic signs, such as “k”, “rr”, and “ll”, to better capture the nuances of MSL and enhance expressiveness. The proposed Hybrid Neural Network-based algorithm integrates these dynamic elements effectively, achieving an F1 score of 90.91%, precision of 91.25%, recall of 91.05%, and accuracy of 91.09% in the extended alphabet classification. These results demonstrate the system’s potential to improve accessibility and inclusivity for the Mexican deaf community.

Stress response regulation of mRNA translation: Implications for antioxidant enzyme expression in cancer

From tumorigenesis to advanced metastatic stages, tumor cells encounter stress, ranging from limited nutrient and oxygen supply within the tumor microenvironment to extrinsic and intrinsic oxidative stress. Thus, tumor cells seize regulatory pathways to rapidly adapt to distinct physiologic conditions to promote cellular survival, including manipulation of mRNA translation. While it is now well established that metastatic tumor cells must up-regulate their antioxidant capacity to effectively spread and that regulation of antioxidant enzymes is imperative to disease progression, relatively few studies have assessed how translation and the hijacking of RNA systems contribute to antioxidant responses of tumors. Here, we review the major stress signaling pathways involved in translational regulation and discuss how these are affected by oxidative stress to promote prosurvival changes that manipulate antioxidant enzyme expression. We describe how tumors elicit these adaptive responses and detail how stress-induced translation can be regulated by kinases, RNA-binding proteins, RNA species, and RNA modification systems. We also highlight opportunities for further studies focused on the role of mRNA translation and RNA systems in the regulation of antioxidant enzyme expression, which may be of particular importance in the context of metastatic progression and therapeutic resistance.

Proteomimetic polymer blocks mitochondrial damage, rescues Huntington's neurons, and slows onset of neuropathology in vivo.

Recently, it has been shown that blocking the binding of valosin-containing protein (VCP) to mutant huntingtin (mtHtt) can prevent neuronal mitochondrial autophagy in Huntington's disease (HD) models. Herein, we describe the development and efficacy of a protein-like polymer (PLP) for inhibiting this interaction in cellular and in vivo models of HD. PLPs exhibit bioactivity in HD mouse striatal cells by successfully inhibiting mitochondrial destruction. PLP is notably resilient to in vitro enzyme, serum, and liver microsome stability assays, which render analogous control oligopeptides ineffective. PLP demonstrates a 2000-fold increase in circulation half-life compared to peptides, exhibiting an elimination half-life of 152 hours. In vivo efficacy studies in HD transgenic mice (R6/2) confirm the superior bioactivity of PLP compared to free peptide through behavioral and neuropathological analyses. PLP functions by preventing pathologic VCP/mtHtt binding in HD animal models; exhibits enhanced efficacy over the parent, free peptide; and implicates the PLP as a platform with potential for translational central nervous system therapeutics.

Translation Evaluation of Social Media Auto-Translate Feature toward News Headline

Machine translation has become an everyday part of society's news consumption patterns. Unfortunately, important news content is poorly mediated by social media, especially on the output of the translation system feature on the site. Therefore, further research is needed on the evaluation of the auto-translate system on social media. This study focuses on the results of linguistic evaluation on the output of social media machine translation in translating news headlines. The data were taken from news coverage by the Reuters channel in the period of September 1st–7th, 2023 and from social media such as Facebook and Twitter. The translation evaluation used linguistic elements consisting of orthography, morphology, lexical, semantic, and syntax. This study found various kinds of linguistic non-equivalence. Orthographic non-equivalence includes punctuation, capitalization, and apostrophes. Morphological non-equivalence in the two systems was found in number concordance, inflection, derivation, and compounding. Lexical errors occur in both systems in the form of no correspondence and missing target words, as well as untranslated words and proper noun. In terms of semantics, both systems had difficulties with polysemy, homonymy, and expression. Finally, syntactic errors occur with missing articles, prepositions, and inappropriate arrangements of syntactic elements.

The impact of contextual understanding on neural machine translation accuracy: A case study of Indonesian cultural idioms in English translation

The translation of Indonesian idiomatic expressions into English using Neural Machine Translation (NMT) systems presents significant challenges due to the intricate nature of idiomatic language. Idioms are culturally embedded constructs conveying meanings different from their literal interpretations. This study explores the effectiveness of NMT systems in capturing and translating Indonesian idiomatic expressions into English. Using a qualitative approach, 150 diverse Indonesian idioms were evaluated through Google Translate and DeepL. Each idiom was assessed for semantic accuracy, syntactic coherence, and contextual fidelity. Qualitative analysis provided a comprehensive evaluation of translation quality. Findings reveal significant challenges for NMT systems in translating idiomatic expressions. Both Google Translate and DeepL show strengths in some areas, but also have limitations. While they generally capture literal meanings, they often miss metaphorical nuances and cultural connotations. Syntactic errors such as incorrect word order and tense inconsistencies are common, especially in complex idioms. In addition, contextual fidelity analysis shows that NMT systems struggle with contextual appropriateness and pragmatic usage, resulting in translations lacking cultural sensitivity and relevance. These challenges highlight the need for improved algorithms to better interpret and translate idiomatic language across cultural contexts. Accurate idiom translation is crucial for language learners to understand cultural nuances and idiomatic usage. Enhancing the capabilities of NMT systems requires refining algorithms, developing context-aware models, and expanding training datasets with diverse idiomatic expressions. Overcoming these challenges will advance machine translation capabilities and improve cross-cultural communication and language learning experiences.

Towards effectively testing machine translation systems from white-box perspectives

Towards effectively testing machine translation systems from white-box perspectives

Application of AI in Turkish Sign Language Translation: A Case Study of its Use and Purpose

In Turkey, around 3 million people have hearing impairments, and the shift to digital platforms during the pandemic has worsened accessibility challenges as websites and apps often ignore their needs. Research shows that 50% of hearing-impaired individuals struggle to understand written text due to Turkish Sign Language being their first language, with Turkish serving as a second language. Differences in grammar between Turkish and Turkish Sign Language, along with a limited sign language vocabulary, further hinder comprehension. To solve this, we developed AI-powered sign language translation systems that allow people who are deaf or hard of hearing to access digital content in Turkish Sign Language. SignForDeaf's system translates text into sign language videos using Natural Language Processing (NLP) and generates seamless videos with smooth word transitions. Currently, this system supports Turkish Sign Language, with future plans to include other languages like Arabic, American, British, and Finnish Sign Language. The system was designed in collaboration with sign language experts to ensure accuracy and an inclusive development of a digital environment.

The Development of AI-Powered Automatic Video Sign Language Translation Systems

People who are deaf or hard of hearing often face challenges in fully understanding written subtitles in videos due to the differences between sign language and spoken language, each having unique grammar and structures. While many social media platforms provide automatic subtitles, they are often insufficient for accessibility. Sign language interpretation should be included to make video content fully accessible. Traditionally, adding sign language to videos involves a time-consuming process of recording and embedding separate translation videos, which must be redone with any changes to the original video. The plugin simplifies this process by dynamically translating updates directly from subtitle files, significantly reducing the time, effort, and cost involved. It allows seamless sign language support for any video without modifying the original content. The plugin integrates with video players, providing a user-controlled, customizable sign language window that can be activated, moved, resized, or turned off as desired.

Cardiomyocyte-Specific Overexpression of Activated Yes-Associated Protein Modified-RNA Promotes Cardiomyocyte Proliferation and Myocardial Regeneration.

The proliferative capacity of cardiomyocytes in adult mammalian hearts is far too low to replace the cells that are lost to myocardial infarction. Both cardiomyocyte proliferation and myocardial regeneration can be improved via the overexpression of a constitutively active variant of YAP5SA (Yes-associated protein, 5SA [active] mutant), but persistent overexpression of proliferation-inducing genes could lead to hypertrophy and arrhythmia, whereas off-target expression in fibroblasts and macrophages could increase fibrosis and inflammation. Transient overexpression of YAP5SA or GFP (green fluorescent protein; control) was targeted to cardiomyocytes via our cardiomyocyte-specific modified mRNA translation system (YAP5SACM-SMRTs or GFPCM-SMRTs, respectively). YAP5SA-cardiomyocyte specificity was confirmed via invitro experiments in cardiomyocytes and cardiac fibroblasts that had been differentiated from human induced- pluripotent stem cells and in human umbilical-vein endothelial cells, and the regenerative potency of YAP5SACM-SMRTs was evaluated in a mouse myocardial infarction model. In cultured human induced-pluripotent stem cells-cardiomyocytes, YAP was abundantly expressed for 3 days after YAP5SACM-SMRTs administration and was accompanied by increases in the expression of markers for proliferation, before declining to near-background levels after day 7, whereas GFP fluorescence remained high from days 1 to 3 after GFPCM-SMRTs treatment and then slowly declined. GFP fluorescence was also observed in human induced-pluripotent stem cells-cardiac fibroblasts and human umbilical-vein endothelial cells on day 1 after GFPCM-SMRTs administration but declined substantially by day 3. In the mouse myocardial infarction model, echocardiographic assessments of left-ventricular ejection fraction and fractional shortening were significantly greater, whereas infarct sizes were significantly smaller in YAP5SACM-SMRTs-treated mice than in vehicle-treated control animals, and YAP5SACM-SMRTs appeared to promote cardiomyocyte proliferation. The CM-SMRTs can be used to transiently and specifically overexpress YAP5SA in cardiomyocytes, and this treatment strategy significantly promoted cardiomyocyte proliferation and myocardial regeneration in a mouse myocardial infarction model.

Comprehensive Analysis of Codon Usage Bias in Human Papillomavirus Type 51.

Human papillomavirus type 51 (HPV-51) is associated with various cancers, including cervical cancer. Examining the codon usage bias of the organism can offer valuable insights into its evolutionary patterns and its relationship with the host. This study comprehensively analyzed codon usage bias in HPV-51 by examining 64 complete genome sequences sourced from the NCBI GenBank database. Our analysis revealed no noteworthy preference for codon usage in HPV-51 overall. However, there was a noticeable bias towards A/T-ending codons, accompanied by GC3s below 32%. Dinucleotide frequency analysis revealed reduced frequencies for ApA, CpG, and TpC dinucleotides, while CpA and TpG dinucleotides were more frequent than others. Relative Synonymous Codon Usage analysis revealed 30 favored codons, primarily concluding with A/T nucleotides. Further analysis using Parity Rule 2, Effective Number of Codons plot, and neutrality plot indicated a balance between mutational pressure and natural selection, with natural selection being the primary force shaping codon usage bias. The Isoacceptor tRNA Pool analysis indicates that HPV-51 has a higher translation efficiency within the human cellular translational system. Moreover, the Codon Adaptation Index and Relative Codon Deoptimization Index analyses suggested a moderate adaptation of HPV-51 to human codon preferences. Our discoveries offer valuable perspectives on how HPV-51 evolves and uses genetic codes, contributing to a deeper comprehension of its endurance and disease-causing potential.

Advancements in Key Technologies for Vibration Isolators Utilizing Electromagnetic Levitation

With the advancement of manufacturing, the precision requirements for various high-precision processing equipment and instruments have further increased. Due to its noncontact nature, simple structure, and controllable performance, electromagnetic levitation has broad application prospects in ultra-precision instruments and ground testing of aerospace equipment. Research on vibration isolation technology using electromagnetic levitation is imperative. This paper reviews the latest research achievements of three types of passive isolators and five active isolation actuators. It also summarizes the current research status of analytical methods for passive isolators and the impact of isolator layout. This study explores current isolators’ achievements, such as the development of passive isolators that generate negative stiffness and require mechanical springs for uniaxial translational vibrations, single-function actuators, and control systems focused on position and motion vibration control. Based on the current isolators’ characteristics, this review highlights future developments, including focusing on passive isolators for heavy loads and multi-axis isolation, addressing complex vibrations, including rotational ones, and developing methods to calculate forces and torques for arbitrary six-DOF movements while improving speed. Additionally, it emphasizes the importance of multifunctional actuators to simplify system structures and comprehensive control systems that consider more environmental factors. This provides significant reference value for vibration isolation technology using electromagnetic levitation.

Evaluating the Translation Accuracy of ChatGPT and DeepL Through the Lens of Implied Subjects

This research aims to offer deeper insights into the translation quality of large language models and neural machine translation systems. It compares the performance of two state-of-the-art models, ChatGPT and DeepL, to explore which system achieves higher translation accuracy. This research employs a highly focused and specialized method to evaluate translation performance, enabling the identification of nuanced differences between advanced AI/machine models. It focuses on implied subjects in contemporary Chinese writer Qiuyu Yu’s Iraq travelogue and examines how the two models handle these implied subjects when translating the text into English. The analysis shows that both ChatGPT and DeepL exhibit high-level intelligence when dealing with the implied subjects. They can flexibly adjust sentence structures to maintain the implicitness or to make the implied subjects explicit. The analysis further reveals that ChatGPT performs better in accurately inferring the implied subjects, which thereby reflects ChatGPT’s superior contextual understanding capability. The comparison of the source text and target texts also shows that ChatGPT’s translation is more succinct, nuanced, and easier to read. DeepL’s translation, by comparison, prioritizes reducing potential misunderstanding. This research tentatively concludes that, though not specialized for translation, ChatGPT translates with higher accuracy and demonstrates more sophisticated intelligence compared to DeepL when translating literary texts.

Diversification of sphingolipid synthase activities in kinetoplastid protozoa

Phosphosphingolipids (PSL) are essential components of eukaryotic membranes. The major PSL in fungi and protists is inositol phosphorylceramide (IPC), while sphingomyelin (SM), and to a lesser extent ethanolamine phosphorylceramide (EPC) predominate in mammals. Most kinetoplastid protozoa have a syntenic locus that encodes a single sphingolipid synthase (SLS) gene. Uniquely, among the kinetoplastids, the salivarian (African) trypanosomes have expanded this locus from a single gene in Trypanosoma vivax (TvSLS) to four genes in T. brucei (TbSLS1-4). We have previously shown that one of these is an IPC synthase, while the others are SM/EPC synthases, and that specificity is controlled by a single signature residue (IPC, serine; SM/EPC, phenylalanine). This residue is serine in T. cruzi and Leishmania major SLSs, both of which are demonstrated IPC synthases. However, T. vivax has a tyrosine at this residue raising the issue of specificity. Using a liposome-supplemented in vitro translation system we now show that T. vivax SLS is an SM/EPC synthase, and that the basal kinetoplastid Bodo saltans SLS is an IPC synthase (serine). We use these data, and a multiple alignment of available sequences, to discuss the evolution of kinetoplastid SLSs and their unique expansion in T. brucei and related salivarian trypanosomes.

BPE and morphologically segmented phrase based statistical machine translation system for Indian languages to resource constrained language Bodo

BPE and morphologically segmented phrase based statistical machine translation system for Indian languages to resource constrained language Bodo

Study on an Intelligent English Translation Method Using an Improved Convolutional Neural Network Model

This study uses an enhanced convolutional neural network (CNN) model for English translation. Traditional translation methods often struggle with complex language structures, prompting the adoption of deep learning techniques, particularly CNNs, in natural language processing. The research outlines CNN fundamentals and their relevance in language processing, elucidating the design and implementation of an improved CNN model. To address CNN limitations in maintaining coherence during the translation of lengthy texts, historical attention mechanisms are incorporated to enhance translation performance. Experimental validation conducted in MATLAB demonstrates notable improvements in translation task performance, evidenced by significant increases in BLEU scores. Results highlight the model's capacity to integrate contextual information, thereby enhancing translation coherence and accuracy. Additionally, the study establishes a mathematical framework for augmenting CNNs with attention mechanisms, providing valuable insights for the development of intelligent English translation systems.

Assessment of Computer-Aided Translation Quality Based on Large-Scale Corpora

CAT technology, utilizing translation memory and quality control tools, boosts translation efficiency and consistency. Yet, it faces challenges with cultural nuances, context, and creativity, requiring human intervention. This study explores leveraging large-scale corpora to enhance machine translation quality assessment accuracy and efficiency. Automated evaluation models surpass manual limitations, enhancing translation system performance and decision-making. Supervised and unsupervised machine translation quality assessment methods, coupled with data augmentation strategies, reduce model bias by reinforcing the connection between source and target languages. The proposed model achieves 94.8% and 92.6% correctness on Google Translate and Wikipedia datasets, slightly surpassing manual rates. Data augmentation and cascade evaluation notably improve translation evaluation accuracy, especially with large-scale corpora. Future research targets higher-quality pseudo-labeled data, refined evaluation models, and advanced assessment methods for diverse translation needs.

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.

Glycolytic pathway analysis and gene expression profiles of combination of aloe vera and paclitaxel on non-small cell lung cancer and breast cancer.

The purpose of this study is to enhance the effectiveness of known anticancer medications using natural compounds. The study investigated the impact of combining AVE with PAX on non-small cell lung cancer (A549) and breast cancer (MCF7). In this study, A549 and MCF7 cells were treated with PAX (5μM), AVE (24μg/mL), and a combination of PAX and AVE (5μM + 24μg/mL). The glucose consumption rates of the cells were determined by extracellular acidification rate (ECAR) thanks to the SeaHorse XFe24 instrument. In addition, gene expression profiles were determined by performing Total RNA sequencing with the Novaseq 6000 instrument. Finally, the expressions of GAPDH, BAX, and BCL-2 genes involved in the apoptotic pathway were detected by RT-qPCR. The combined application of PAX and AVE reduced the ECAR value in both cell lines. According to the RT-qPCR results, the expression level of the apoptotic gene BAX increased in both cell lines (p < 0.05). Total RNA sequencing revealed that the combination effects of PAX and AVE play a role in the ribosome mechanism, thereby affecting the protein translation system in MCF7 while apoptosis and cell cycle have come to the forefront in A549.

TOWARDS BUILDING A TAXONOMY OF MACHINE TRANSLATION ERRORS: VERSATILITY VS TARGETED APPROACHES

This study examines problems in existing approaches to classifying machine translation (MT) errors. Despite the fact that such classifications can be based on a taxonomy developed on the material of traditional translations, MT errors have their own specifics, which should also be considered in the classification system. The first part of the presented paper discusses the concepts of machine translation and translation error per se and provides an overview of the main approaches to building a taxonomy of translation errors in traditional, “human” translation. In the second part, we discuss the existing classification systems proposed for the analysis of machine translation output and some of their limitations. The research material in the discussed papers is mainly focused on translations in pairs with the Russian language and obtained through the use of the most popular in Russia machine translation services, viz. Yandex Translator, Google Translate and Promt. In particular, we discuss the main classes of errors identified by various authors, the frequency of the errors of the said classes, as well as their gravity in terms of their ability to cause a communication failure. This study argues that there is currently no unified approach to the construction of such classifications, as well as that they are inevitably dependent on the type of the analysed text, the language pair and the chosen automated translation system. It is concluded that the optimal approach to classifying MT errors is not a universal approach, but a targeted one, that is, depending on certain translation parameters. The possibility of practical applications of machine translation in teaching foreign languages and in the work of professional translators, as well as the need for post-editing of translated texts, are also discussed.

Dynamic Multi-Granularity Translation System: DAG-Structured Multi-Granularity Representation and Self-Attention

In neural machine translation (NMT), the sophistication of word embeddings plays a pivotal role in the model’s ability to render accurate and contextually relevant translations. However, conventional models with single granularity of word segmentation cannot fully embed complex languages like Chinese, where the granularity of segmentation significantly impacts understanding and translation fidelity. Addressing these challenges, our study introduces the Dynamic Multi-Granularity Translation System (DMGTS), an innovative approach that enhances the Transformer model by incorporating multi-granularity position encoding and multi-granularity self-attention mechanisms. Leveraging a Directed Acyclic Graph (DAG), the DMGTS utilizes four levels of word segmentation for multi-granularity position encoding. Dynamic word embeddings are also introduced to enhance the lexical representation by incorporating multi-granularity features. Multi-granularity self-attention mechanisms are applied to replace the conventional self-attention layers. We evaluate the DMGTS on multiple datasets, where our system demonstrates marked improvements. Notably, it achieves significant enhancements in translation quality, evidenced by increases of 1.16 and 1.55 in Bilingual Evaluation Understudy (BLEU) scores over traditional static embedding methods. These results underscore the efficacy of the DMGTS in refining NMT performance.