Automatic Translation Research Articles

Structural and Parametric Synthesis of Neural Network Controllers for Control Objects with Limiters

The article presents a methodology for the synthesis of digital control systems for nonlinear objects with limiters under conditions of incomplete information. Closed-loop tracking systems with negative feedback are considered. Artificial neural networks are proposed to build a controller, which is included in series with the control object. This approach is effective when known classical methods do not allow to synthesize control. This is the case, for example, if the mathematical model is essentially nonlinear and is not fully defined. The developed methods allow us to expand the class of technical systems, for which the direct (without using various kinds of simplifications) synthesis of control laws that are close to optimal is possible. In addition, neural network controllers possess the properties of robustness, adaptivity, and are initially digital, i.e. those qualities, which are very much in demand in practice. In article main attention is given to such problems, as a choice of neural network structure for neural simulator and neural controller, construction of training sample, ensuring convergence of the process of weights correction. For training neural networks the method of back propagation of error is used as a basic one. The effectiveness of the proposed technique is demonstrated by the example of the synthesis of a neuroregulator for a specific technical object and its comparison with classical control systems. It should be noted that today neural network technologies are widespread enough in various spheres of activity. The successes demonstrated in sound processing, image processing, automatic translation, in navigation systems, in big data processing are impressive. However, their application in automatic control systems is not so widespread. The authors of this article believe that the potential of artificial neural networks can be used in this direction. It should be understood that the use of neural networks is effective only under certain conditions and properties of the control object.

Empowering Deaf Learners: The Promise of Sign Language MOOCs

In the realm of education, deaf students face significant communication barriers that limit their integration and success within intellectual communities. Due to the use of distinct mother languages, a communication gap arises between deaf and non-deaf communities. This poses serious difficulties for deaf students, who communicate primarily in sign language and are unable to fluently read materials written in spoken language. As most didactic materials in higher education are available exclusively in spoken languages, this further limits the opportunities and chances for deaf students to succeed. While technological advancements have made it possible to create tools and services that translate between spoken and sign languages, they are often expensive and not widely available in educational settings. The proposal is a pedagogical model that seeks to overcome these obstacles and foster inclusivity by offering deaf students access to educational resources in their mother language – sign language. This is accomplished using Massive Online Open Courses (MOOCs) paired with automatic sign language translation technology. To demonstrate this model's effectiveness, a digital literacy course for schoolteachers was developed and evaluated within a group of users, including deaf students. The paper discusses the state-of-the-art of digital educational content available to the deaf, the process of translating digital learning materials into sign languages, the specifications needed to structure educational content for the deaf, and an example of the development methodology for creating multilingual learning objects for the deaf. The evaluation results indicate that MOOC is inclusive and positively impacts deaf students' learning outcomes. The proposed pedagogical model has significant implications for improving access to education and promoting inclusivity for deaf students worldwide. By utilizing technological advancements and creating inclusive digital educational materials, it is possible to bridge the communication gap between deaf and non-deaf individuals, thereby enabling deaf students to acquire, expand, and improve their knowledge in various subjects.

Žodynai ir automatinis vertimas: ką renkasi besimokantieji lietuvių kalbos kaip svetimosios

This article presents the results of research on using electronic dictionaries and machine translation tools when learning Lithuanian as a foreign language. The aim of this study was to indicate which dictionaries and machine translation tools are used for which purposes and what kinds of difficulties learners experience while using them. The research data were obtained by conducting an anonymous online poll in Lithuanian and English (using Google Forms). The questionnaire contained 15 questions. The participants in the study were chosen by way of availability sampling. The survey was carried out at Vilnius University in March 2023 with 67 respondents: 11 in Lithuanian and 56 in English. The study revealed that when learning Lithuanian as a foreign language, the respondents used both dictionaries and automatic translation tools. Most often, both tools were used for translation from and into Lithuanian. The types of language activities for which these tools were used were reading, writing, and partly listening. Only the respondents to English questionnaires used the mentioned tools for speaking. When learning Lithuanian, both monolingual and bilingual dictionaries were used, but monolingual dictionaries were used more than bilingual ones: 46.27% of respondents used monolingual dictionaries, and 31.34% used bilingual dictionaries.The study confirmed the researchers’ claim that monolingual dictionaries are used more by respondents with a higher level of language proficiency. Automatic translation tools were used by all respondents. The most popular of them is Google Translate, followed by DeepL. Other tools – Reverso, Bing Microsoft Translator, Systran Translate, Yandex Translate, and Lingea – are used by a few and only English questionnaire respondents. Dictionaries were consulted for the meaning of an unknown word, for synonyms, to confirm the meaning and spelling, to find usage examples, the field of the word used, and stress. Automatic translation is used for unknown words, phrases, and sentences to hear the pronunciation and find usage examples. The most common difficulties in using dictionaries were: unclear definitions of words and abbreviations; no text-to-speech function; too few usage examples; and unclear or missing examples. Weaknesses of automatic translation tools indicated by respondents: poor translation quality; lack of a search function; no examples of usage; no text-to-speech.

Spatio-temporal representation learning enhanced speech emotion recognition with multi-head attention mechanisms

Speech emotion recognition (SER) systems have become essential in various fields, including intelligent healthcare, customer service, call centers, automatic translation systems, and human–computer interaction. However, current approaches predominantly rely on single frame-level or utterance-level features, offering only shallow or deep characterization, and fail to fully exploit the diverse types, levels, and scales of emotion features. The limited ability of single features to capture speech emotion information, along with the ineffective combination of different features’ complementary advantages through simple fusion, pose significant challenges. To address these issues, this paper presents a novel spatio-temporal representation learning enhanced speech emotion recognition with multi-head attention mechanisms(STRL-SER). The proposed technique integrates fine-grained frame-level features and coarse-grained utterance-level emotion features, while employing separate modules to extract deep representations at different levels. In the frame-level module, we introduce parallel networks and utilize a bidirectional long short-term memory network (BiLSTM) and an attention-based multi-scale convolutional neural network (CNN) to capture the spatio-temporal representation details of diverse frame-level signals. Consequently, we extract deep representations of utterance-level features to effectively learn global speech emotion features. To leverage the advantages of different feature types, we introduce a multi-head attention mechanism that fuses the deep representations from various levels. This fusion approach retains the distinctive qualities of each feature type. Finally, we employ segment-level multiplexed decision making to generate the ultimate classification results. We evaluate the effectiveness of our proposed method on two widely recognized benchmark datasets: IEMOCAP and RAVDESS. The results demonstrate that our method achieves notable performance improvements compared to previous studies. On the IEMOCAP dataset, our method achieves a weighted accuracy (WA) of 81.60% and an unweighted accuracy (UA) of 79.32%. Similarly, on the RAVDESS dataset, we achieve a WA of 88.88% and a UA of 87.85%. These outcomes confirm the substantial advancements realized by our proposed method.

Non-Axiomatic Logic Modeling of English Texts for Knowledge Discovery and Commonsense Reasoning

Non-axiomatic logic (NAL) is a term-based, non-monotonic, multi-valued logic with evidence-based formal semantics. All those characteristics position NAL as an excellent candidate for modeling natural language expressions and supporting artificial agents while performing knowledge discovery and commonsense reasoning tasks. In this article, we propose a set of rules for the automatic translation of natural language (NL) text into the formal language of non-axiomatic logic (NAL). Several free available tools are used to support a previous linguistic analysis, and a common sense ontology is used to populate a background knowledge base that helps to delimit the scope and the semantics of logical formulas translated. Experimentation shows our set to be the most comprehensive NL-to-NAL translation rule set known so far. Furthermore, we included an extensive set of examples to show how our proposed set of rules can be used for translating a wide range of English statements with varying grammatical structures.

What is the rate of text generated by artificial intelligence over a year of publication in Orthopedics & Traumatology: Surgery & Research? Analysis of 425 articles before versus after the launch of ChatGPT in November 2022

BackgroundThe use of artificial intelligence (AI) is soaring, and the launch of ChatGPT in November 2022 has accelerated this trend. This “chatbot” can generate complete scientific articles, with risk of plagiarism by mining existing data or downright fraud by fabricating studies with no real data at all. There are tools that detect AI in publications, but to our knowledge they have not been systematically assessed for publication in scientific journals. We therefore conducted a retrospective study on articles published in Orthopaedics & Traumatology: Surgery & Research (OTSR): firstly, to screen for AI-generated content before and after the publicized launch of ChatGPT; secondly, to assess whether AI was more often used in some countries than others to generate content; thirdly, to determine whether plagiarism rate correlated with AI-generation, and lastly, to determine whether elements other than text generation, and notably the translation procedure, could raise suspicion of AI use. HypothesisThe rate of AI use increased after the publicized launch of ChatGPT v3.5 in November 2022. Material and methodsIn all, 425 articles published between February 2022 and September 2023 (221 before and 204 after November 1, 2022) underwent ZeroGPT assessment of the level of AI generation in the final English-language version (abstract and body of the article). Two scores were obtained: probability of AI generation, in six grades from Human to AI; and percentage AI generation. Plagiarism was assessed on the Ithenticate application at submission. Articles in French were assessed in their English-language version as translated by a human translator, with comparison to automatic translation by Google Translate and DeepL. ResultsAI-generated text was detected mainly in Abstracts, with a 10.1% rate of AI or considerable AI generation, compared to only 1.9% for the body of the article and 5.6% for the total body+abstract. Analysis for before and after November 2022 found an increase in AI generation in body+abstract, from 10.30±15.95% (range, 0–100%) to 15.64±19.8% (range, 0–99.93) (p < 0.04; NS for abstracts alone). AI scores differed between types of article: 14.9% for original articles and 9.8% for reviews (p<0.01). The highest rates of probable AI generation were in articles from Japan, China, South America and English-speaking countries (p<0.0001). Plagiarism rates did not increase between the two study periods, and were unrelated to AI rates. On the other hand, when articles were classified as “suspected” of AI generation (plagiarism rate ≥ 20%) or “non-suspected” (rate<20%), the “similarity” score was higher in suspect articles: 25.7±13.23% (range, 10–69%) versus 16.28±10% (range, 0–79%) (p < 0.001). In the body of the article, use of translation software was associated with higher AI rates than with a human translator: 3.5±5% for human translators, versus 18±10% and 21.9±11% respectively for Google Translate and DeepL (p < 0.001). DiscussionThe present study revealed an increasing rate of AI use in articles published in OTSR. AI grades differed according to type of article and country of origin. Use of translation software increased the AI grade. In the long run, use of ChatGPT incurs a risk of plagiarism and scientific misconduct, and needs to be detected and signaled by a digital tag on any robot-generated text. Level of evidenceIII; case-control study.

A Review of Machine Translation Techniques and Proposal to Automatic Translation of Videos

Translation of videos including propulsion technology videos has been one of the toughest jobs to do during the eLearning development process. We have tried it earlier by involving mostly manual work and found it to be tiresome. After the fast evolution of Machine learning and Natural language processing techniques, it is feasible now to do the translation involving and inculcating intelligent techniques in the process. We have reviewed the technologies and techniques and presenting it in this paper.

In the Anatomic Studies, Is It Correct to Add an Artificial Intelligence Such as ChatGPT as a Co-Author?

Dear Editors, I follow your journal with pleasure. I enjoyed reading your articles on artificial intelligence [1,2]. I would like to share my ideas with you about the use of artificial intelligence, which many researchers are currently curious about. Health applications are being renewed day by day. New technology and new gains in treatment have made us curious about the use of artificial intelligence [3]. Health experts using methods such as machine learning and deep learning, which are sub-branches of artificial intelligence, are switching to new methods in processes such as diagnosis, treatment, rehabilitation, and health protection of diseases [4,5]. ChatGPT (Chat Generative Pre-trained Transformer), used in the studies and also added as a co-author, is ChatGPT, an artificial intelligence chatbot with the GPT-3.5.5 language model, which was released by the artificial intelligence research company OpenAI on November 30, 2022, and reached 1 million users in 5 days [3]. This robot can become a promising and powerful tool for tasks such as automatic drafting, article summarization, homework, and language translation, which can be useful for making writing work faster and easier in academic activities. Talan and Kalınkara [6] used ChatGPT in the Anatomy course and found that students performed better. Aktay et al. [7] stated that students found ChatGPT interesting and fun and that it increased students achievement. Saygın and Kabakcı [8] stated that the use of ChatGPT can be useful in information learning and in producing something new. On the other hand, in a few studies, the addition of artificial intelligence as a co-author has been stated to be worrying [1,9]. If the academician does not produce it himself, but the artificial intelligence he uses as a co-author does, it may create an obstacle in multidisciplinary studies. If ChatGPT is written instead of a collaborator when you want to study anatomy, we are concerned about the decrease in our assistants, students, and research scientists. As a result, ChatGPT can be used in the learning of the Anatomy course and can also be useful in increasing knowledge in academia, but I do not find it ethically correct for ChatGPT to do the work that master’s and doctoral students have done for years in master’s and doctoral studies in a short time. In addition, in the studies on Anatomy, while writing a scientist is a co-author, the inclusion of ChatGPT will pave the way for the production of studies without labor in a short time, and this will cause serious concerns in academia. It will save time in academia, but academics’ communication with each other will decrease, there will be no cooperation and multidisciplinary studies will decrease. Who will be responsible in case of any problems while participating as a co-author with ChatGPT? I think that the use of ChatGPT should be supervised and that it is worrying to be a co-author at the moment. Yours sincerely

Towards a methodology for translation activity: an activity theory approach

ABSTRACT The question of “How translation is done” is crucial in the field of translation studies as it serves as the foundation for both teaching and criticism. To address this question, one must theoretically explain the methodology behind translation activities, including how translators select methods, strategies, and techniques to achieve meaningful translations. Despite the existence of various cross-disciplinary paradigms in translation studies, a consensus has not yet been reached on this issue. According to Activity Theorists, a socio-cultural activity is realized through goal-directed actions that are ultimately carried out by conditioned operations. Based on this perspective, we propose an explanatory framework for understanding the methodology of translation activity. This framework may also be useful for revising and evaluating the output of automatic translation operations conducted by AI language models.

인공지능 기반 한문 번역을 위한 코퍼스 추출 및 정제 과정

It is fun to imagine how good it would be to be able to translate a lot of Chinese classics by using AI in an era where the field of AI is expanding day by day. This is why commercialization of machine translation called artificial intelligence can drastically reduce the manpower and time spent on classical Chinese translation. To this end, various institutions are improving their performance by developing artificial intelligence Chinese character recognition and Chinese character translation programs. AI-based Chinese character translation, research is concentrated in the field of technology, and there are no reported studies on the corpus extraction process and purification process.&#x0D; For deep learning, which trains artificial intelligence, automatic translation data, that is, corpus that connects Chinese text and translations in parallel, must be created. In order to make a corpus, a lot of corpus data is extracted through human translation, and a high-quality corpus is made through refining based on detailed guidelines. Here, we tried to confirm the purification process of how the extracted corpus data were selected.&#x0D; As a result of checking the corpus purification process, words such as various titles, items, and names, metrological units or number of people were all excluded. Short sentences containing place names, human names, government posts, places, products, dates, and proper nouns were also removed because they were not appropriate as corpus. In addition, prefixes consisting of one or two letters, such as adverbs, conjunctions, tense, and pronounciation at the beginning of the sentence, were deleted. In the future, it is expected that the time and expense of translation will be greatly reduced if extraction is carried out in consideration of these contents in the construction of corpus for AI-based Chinese character translation.

Retracted: Design of English Automatic Translation System Based on Machine Intelligent Translation and Secure Internet of Things

Retracted: Design of English Automatic Translation System Based on Machine Intelligent Translation and Secure Internet of Things

Metaverse and translation studies: analysis of machine translation

Although translation is fundamental, in collaboration with artificial intelligence, for the development and success of interaction and communication in the metaverse, little attention has been devoted to the direct influence of linguistic communication in three-dimensional space. Therefore, in this paper we have analyzed different research focused on the metaverse, coming from various disciplines, which have used the machine translation service. The aim is to know the perceptions about its use and quality, and thus understand what potential it would have in virtual reality and the challenges it poses. Documents were retrieved from the SciELO, Google Scholar and Dialnet databases, and those studies that have directly or indirectly used the automatic translation tool to favor communication and interaction between users were included. The results show a generally positive perception on the part of the studies analyzed, although the imprecision of the automatic translation engines and the lack of linguistic combinations are qualified. Finally, the importance of studying communication in the metaverse, opening up the number of available languages and research on the training of translation engines is highlighted, as the revenue of the machine translation market is bound to increase in the coming years due to the pace of production and the productivity offered.

An end-to-end framework for translation of American sign language to low-resource languages in Nigeria

Deaf and hard of hearing (DHH) is a communication impairment that affects millions of people worldwide, hindering communication through the vocal cavity and requiring the use of sign language. This communication gap often leads to limited access to education and job opportunities. While artificial intelligence (AI)-driven technologies have been investigated to address this issue, there has been no study focusing on the intelligent and automatic translation of American sign gestures to text in low-resource languages (LRL) specifically in Nigeria. To address this gap, we propose a novel end-to-end framework for translating American Sign Language (ASL) to LRLs such as Hausa, Ibo, and Yoruba in Nigeria. Our framework employs a Transformer-based model for ASL-to-Text generation and a translation model known as no language left behind for translating the generated text from English to LRL. We evaluated the performance of the ASL-to-Text model using BLEU scores, which resulted in BLEU-1, BLEU-2, BLEU-3, and BLEU-4 scores of 18.55, 8.99, 4.99, and 2.96, respectively. Qualitative analysis of the framework shows that the participants were able to comprehend the translated text and were also satisfied with both the ASL-to-Text and Text-to-LRL models. Our proposed framework demonstrates the potential for AI-driven technologies to foster inclusivity in sociocultural interactions and education, especially for people with DHH in low-resource settings.

The ISEBEL Project

AbstractISEBEL is an online database for belief legends. The acronym stands for: Intelligent Search Engine for Belief Legends. The database contains more than 70,000 traditional legends in Dutch, Frisian, Danish and German, while another 6,000 Icelandic legends are currently being added. The initiative for this project was taken several years ago by the Meertens Institute in Amsterdam, the University of Rostock/the Wossidlo Research Center and UCLA/UC Berkeley. The ambition is to create a European database, with an intelligent search function and geographical visualizations. What makes the search engine ‘intelligent’ is that it can always be searched in English, thanks to high-quality automatic translations in the background. The legend material that is brought together can also form the basis of sophisticated graphs that network themes, motifs, narrators, repertoires, and places. This article includes examples of international legends about mermaids and werewolves.

Reinforcement Learning for Omega-Regular Specifications on Continuous-Time MDP

Continuous-time Markov decision processes (CTMDPs) are canonical models to express sequential decision-making under dense-time and stochastic environments. When the stochastic evolution of the environment is only available via sampling, model-free reinforcement learning (RL) is the algorithm-of-choice to compute optimal decision sequence. RL, on the other hand, requires the learning objective to be encoded as scalar reward signals. Since doing such translations manually is both tedious and error-prone, a number of techniques have been proposed to translate high-level objectives (expressed in logic or automata formalism) to scalar rewards for discrete-time Markov decision processes. Unfortunately, no automatic translation exists for CTMDPs. We consider CTMDP environments against the learning objectives expressed as omega-regular languages. Omega-regular languages generalize regular languages to infinite-horizon specifications and can express properties given in popular linear-time logic LTL. To accommodate the dense-time nature of CTMDPs, we consider two different semantics of omega-regular objectives: 1) satisfaction semantics where the goal of the learner is to maximize the probability of spending positive time in the good states, and 2) expectation semantics where the goal of the learner is to optimize the long-run expected average time spent in the ''good states'' of the automaton. We present an approach enabling correct translation to scalar reward signals that can be readily used by off-the-shelf RL algorithms for CTMDPs. We demonstrate the effectiveness of the proposed algorithms by evaluating it on some popular CTMDP benchmarks with omega-regular objectives.

Основные направления развития современной армянской лингвистики

The present paper examines the topical areas of research in the field of linguistics in modern Armenia. These include, first of all, the continuation of work on the creation of a dialectological atlas of the Armenian language. Today, the problems of Armenian ethno- and glottogenesis are of great importance, where theoretical developments related to the description of the development of the Armenian language are of great importance. Scientists of the Institute of Language of the National Academy of Sciences of Armenia are doing a great work on the development of modern Armenian lexicography and terminology. Collaborative work of Armenian scientists is also carried out on the creation of linguistic base for automatic translation, linguistic processing of information.

Cross-lingual Text Reuse Detection at Document Level for English-Urdu Language Pair

In recent years, the problem of Cross-Lingual Text Reuse Detection (CLTRD) has gained the interest of the research community due to the availability of large digital repositories and automatic Machine Translation (MT) systems. These systems are readily available and openly accessible, which makes it easier to reuse text across languages but hard to detect. In previous studies, different corpora and methods have been developed for CLTRD at the sentence/passage level for the English-Urdu language pair. However, there is a lack of large standard corpora and methods for CLTRD for the English-Urdu language pair at the document level. To overcome this limitation, the significant contribution of this study is the development of a large benchmark cross-lingual (English-Urdu) text reuse corpus, called the TREU (Text Reuse for English-Urdu) corpus. It contains English to Urdu real cases of text reuse at the document level. The corpus is manually labelled into three categories (Wholly Derived = 672, Partially Derived = 888, and Non Derived = 697) with the source text in English and the derived text in the Urdu language. Another contribution of this study is the evaluation of the TREU corpus using a diversified range of methods to show its usefulness and how it can be utilized in the development of automatic methods for measuring cross-lingual (English-Urdu) text reuse at the document level. The best evaluation results, for both binary ( F 1 = 0.78) and ternary ( F 1 = 0.66) classification tasks, are obtained using a combination of all Translation plus Mono-lingual Analysis (T+MA) based methods. The TREU corpus is publicly available to promote CLTRD research in an under-resourced language, i.e., Urdu.

RETRACTED ARTICLE: Effectiveness model of automatic machine translation of publicity texts based on deep learning

RETRACTED ARTICLE: Effectiveness model of automatic machine translation of publicity texts based on deep learning

Attitudes towards machine translation and languages among travelers

Machine translation (MT), i.e., automatic translation, is a growing field in artificial intelligence with huge impacts on societies and businesses. Despite its importance for traveling and tourism communication, it has not been approached within tourism research. This study aims to fill this gap in knowledge by analyzing how attitudes toward machine translation are related to tourists’ profiles, travel behaviors, and language mindsets. It comprises two parts. The first one concerns a sample of 2535 individuals, while the second concerns a sub-sample of 907 language tourists (LTs). Specific research goals are set for each study: (1) to compare individuals with opposing viewpoints on the importance of MT in terms of profiles and attitudes toward languages; and (2) to understand how LTs’ profiles and travel experiences differed according to their agreement with the importance of MT in their most significant language trip. Statistical exploratory and inferential analyses have been conducted. We conclude that those with more positive views of MT tend to be younger and less educated, report poorer language skills, and attribute greater importance to the role of English as a lingua franca. Concerning LTs, those who rate MT as less important are more likely to have acquired language skills formally, engage more in cultural activities, and have closer contact with locals during their language trips. Acknowledging the role of MT in their most significant language trip is neither associated with a more unfavorable attitude towards the role of language in tourism nor with perceived diminished travel outcomes.

Onception: Active Learning with Expert Advice for Real World Machine Translation

Abstract Active learning can play an important role in low-resource settings (i.e., where annotated data is scarce), by selecting which instances may be more worthy to annotate. Most active learning approaches for Machine Translation assume the existence of a pool of sentences in a source language, and rely on human annotators to provide translations or post-edits, which can still be costly. In this article, we apply active learning to a real-world human-in-the-loop scenario in which we assume that: (1) the source sentences may not be readily available, but instead arrive in a stream; (2) the automatic translations receive feedback in the form of a rating, instead of a correct/edited translation, since the human-in-the-loop might be a user looking for a translation, but not be able to provide one. To tackle the challenge of deciding whether each incoming pair source–translations is worthy to query for human feedback, we resort to a number of stream-based active learning query strategies. Moreover, because we do not know in advance which query strategy will be the most adequate for a certain language pair and set of Machine Translation models, we propose to dynamically combine multiple strategies using prediction with expert advice. Our experiments on different language pairs and feedback settings show that using active learning allows us to converge on the best Machine Translation systems with fewer human interactions. Furthermore, combining multiple strategies using prediction with expert advice outperforms several individual active learning strategies with even fewer interactions, particularly in partial feedback settings.