Human Translation Research Articles

The Affordances of the Translator

This article explores affordance-theoretical readings of Walter Benjamin’s “Task of the Translator,” looking first at Aleksei Procyshyn’s mobilization of Anthony Chemero’s “radical embodied cognitive science” approach to affordances, in which, as Procyshyn summarizes it, “language use is an enactive process of meaning creation, which affords an appropriately situated and capable agent specific potentials for further action.” A closer look shows not only that Procyshyn has not drawn on the full potential of Chemero’s theorization, but that Chemero himself has not developed a 4EA-cogsci affordance theory fully—and that the application of affordance theory to Benjamin ultimately doesn’t work without a complex reframing of both Benjamin and affordance theory. Specifically, toward the end of Benjamin’s essay he moves toward a more personalized understanding of human translators as situated agents—notably Friedrich Hölderlin, but also Martin Luther, Johann Heinrich Voß, A. W. Schlegel, and Stefan George—and another pass through Wilhelm Dilthey’s hermeneutical theory of the Zusammenhang des Lebens (“nexus/intertwining of life”), which Benjamin invokes by name, helps flesh out both an affordance theory of translation and an extended application to Hölderlin’s Sophocles translations. The historical chain from Dilthey through Husserl and Merleau-Ponty to Varela, Thompson, and Rosch’s The Embodied Mind ties hermeneutics, phenomenology, and 4EA cognitive science together under the rubric of the affordances of the translator.

Validation of a 3D perfused cell culture platform as a tool for humanised preclinical drug testing in breast cancer using established cell lines and patient-derived tissues.

3D cell culture models of cancer are currently being developed to recapitulate in vivo physiological conditions and to assess therapeutic responses. However, most models failed to incorporate the biochemical and biophysical stimuli from fluid flow. In this study, a three-dimensional scaffold, SeedEZ was applied within the PerfusionPal perfused culture system to investigate how perfusion, and blood-like oxygen delivery influenced breast cancer cell growth and their responses to a commonly used breast cancer drug tamoxifen. Our results showed that breast cancer cells could be maintained over 3 weeks in PerfusionPal with increased cell viability compared to static 3D culture in fully humanised conditions. This platform also supported examining the effect of tamoxifen on breast cancer cell lines and in primary patient-derived breast cancer samples. Future work is warranted to further the adaption for fully humanised assessment of drug effectiveness in a patient personalized approach with the aim to reduce the burden of animal use in cancer research and increase the degree of human pre-clinical data translation to clinic.

Organic anion-transporting polypeptide 2B1 knockout and humanized mice; insights into the handling of bilirubin and drugs

Organic anion transporting polypeptide 2B1 (OATP2B1/SLCO2B1) facilitates uptake transport of structurally diverse endogenous and exogenous compounds. To investigate the roles of OATP2B1 in physiology and pharmacology, we established and characterized Oatp2b1 knockout (single Slco2b1-/- and combination Slco1a/1b/2b1-/-) and humanized hepatic and intestinal OATP2B1 transgenic mouse models. While viable and fertile, these strains exhibited a modestly increased body weight. In males, unconjugated bilirubin levels were markedly reduced in Slco2b1-/- compared to wild-type mice, whereas bilirubin monoglucuronide levels were modestly increased in Slco1a/1b/2b1-/- compared to Slco1a/1b-/- mice. Single Slco2b1-/- mice showed no significant changes in oral pharmacokinetics of several tested drugs. However, markedly higher or lower plasma exposure of pravastatin and the erlotinib metabolite OSI-420, respectively, were found in Slco1a/1b/2b1-/- compared to Slco1a/1b-/- mice, while oral rosuvastatin and fluvastatin behaved similarly between the strains. In males, humanized OATP2B1 strains showed lower conjugated and unconjugated bilirubin levels than control Slco1a/1b/2b1-deficient mice. Moreover, hepatic expression of human OATP2B1 partially or completely rescued the impaired hepatic uptake of OSI-420, rosuvastatin, pravastatin, and fluvastatin in Slco1a/1b/2b1-/- mice, establishing an important role in hepatic uptake. Expression of human OATP2B1 in the intestine was basolateral and markedly reduced the oral availability of rosuvastatin and pravastatin, but not of OSI-420 and fluvastatin. Neither lack of Oatp2b1, nor overexpression of human OATP2B1 had any effect on fexofenadine oral pharmacokinetics. While these mouse models still have limitations for human translation, with additional work we expect they will provide powerful tools to further understand the physiological and pharmacological roles of OATP2B1.

When Translation Goes Digital: Case Studies and Critical Reflections. Renée Desjardins, Claire Larsonneur, and Philippe Lacour

The emerging and growing field of ‘Digital Humanities’ (DH) has brought profound and comprehensive changes. Particular concern has been aroused by the impact of advanced technologies on translation and Translation Studies (TS). Cutting-edge technologies have developed effective online machine translation (MT) tools such as Google Translate and DeepL. Consequently, there has been an on-going debate about whether digital technologies will completely replace human translation in the near future, which seems unlikely to me. When Translation Goes Digital: Case Studies and Critical Reflections (Desjardins et al., 2021) supports this position and bolsters my confidence in human translation. Consisting of ten studies presented at the sixth International Association for Translation at Hong Kong Baptist University, the book is structured in three parts. Aiming to ‘explore how the digital landscape was impacting translation and TS, broadly and specifically’ (1), it is suitable for scholars, translators, or any translation practitioners curious about the latest studies regarding TS and the current digital world, the relationship between human translation and digital technologies, and how scholars define the relationship between DH and TS.

Metamorphic Robustness Testing for DeepL Translation

Due to the complexity and flexibility of natural languages, such robustness testing involving the ability of machine translation software to handle incorrect or unintended input are extremely challenging, especially if there is no human oracle or reference translation. In this paper, using metamorphic robustness testing methods, we compare the translation of the primitive input with the translation of a subsequent input that has different types of very small changes and may cause large changes in the translation results. The lack of robustness of the DeepL translation is shown by our experiments, these preliminary results show that metamorphic testing has potential in the field of natural language processing.

기계 번역과 자연어처리를 활용한 튜터링 활동 분석

This is a follow-up study by Yoon, Kim, Choi, and Kim (2021). In the current study, the same text as the previous study was machine-translated into English, and then the topics derived through a statistical model, which were compared with the categories derived from the Korean text of the previous study. By doing this, rather than simply evaluating the quality of machine translation, the present study examined 1) the natural language processing model applied to text, 2) the role of human researchers in the process of interpreting the results after fitting the model, and 3) the harmony between human and machine translation, The texts to be analyzed were collected from the liberal arts education tutoring program of D university. A theoretical description of the Latent Dirichlet Allocation model, which is often used for topic modeling, and an explanation of the use of the R library ‘topicmodels’ were added. The analyses revealed a somewhat different pattern of topics from the previous study, and the authors’ interpretation was presented. Finally, while acknowledging the explosive growth of natural language processing and machine translation as a branch of it that cannot be reversed, the room for human translators to complement machine translation was suggested. This study presents a new approach to liberal arts education research by combining data analysis techniques using natural language processing and machine translation discussions.

Google Translate Errors in Legal Texts: Machine Translation Quality Assessment

Machine translation received intensive research in different language pairs; yet, the quality of specialized translation; e.g., legal translation, from Arabic into English received little attention. The present study investigates the quality of machine translation of legal texts from Arabic into English. The paper aims at examining the errors found in the machine translation of legal texts from Arabic into English. It also studies the legal discourse features in machine-translation output. The research questions tackle the accuracy of machine translation of legal discourse and error categories and frequencies in machine translation. The researcher evaluated several factors to assess the quality of Google Translate; i.e., lexical, syntactic, and register-related errors. The study data consists of five legislative texts. The researcher conducted a manual error assessment and classification. To ensure the reliability of the error analysis, an existing human translation of the documents was used as a reference to ensure the reliability of the MT quality assessment and post-editing process. Later, the errors were classified, and their percentages and frequencies were calculated. A few examples of errors in each category were discussed and analysed. The highest error category was lexical errors scoring 43.4%. The last detected error category was deletion with a percentage of 1.7%. Syntactic errors constituted one-fourth of the errors found in the data. Legal register-related errors scored 30.2%. The subcategories of legal register-related errors varied in their occurrence; e.g., one-third of the errors related to legal discourse were legal terms. The study concluded that machine translation; though it provided a comprehensible output, could not translate legal structures and terminology perfectly.

Machine Translation of Selected Ghazals of Hafiz from Persian into English

Neural Machine Translation (NMT) is an end-to-end learning approach for automated translation with the potential to overcome many weaknesses in conventional phrase-based translation systems. However, there are still grey areas that need attention in the field of machine translation, mainly in the context of literary translation and poetic language. This research aims to identify the shortcomings of the NMT in addressing ambiguity and the semantic complexities of poetic expressions that are important in translating literature in a cross-cultural and multilingual context. Furthermore, while human translators struggle to retain naturalness and accuracy in translation, this research aims to identify to what extent machine translation could convey the message in translating literary-based texts from Persian into English. The corpus of this study consists of Forty ghazals from the Divan of Hafiz, which is translated by a computer-assisted translation software, memoQ and then compared to the authentic translation of Hafiz by Henry Wilberforce Clarke. This comparative descriptive study portrayed the challenges and issues related to Hafiz’s poems, which demonstrate that the literary touch is lost in MT products.

The Challenges of Using Machine Translation While Translating Polysemous Words

The disparities between machine translation and human translation have shown to be quite challenging when translating polysemous words. This study examines the challenges of translating polysemous words between French and English by finding an easy way to translate the polysemous terms which is the key objective. It would also help in dealing with the most effective method of translating polysemous terms, while a few succeed in translating the contexts and others struggle to come up with the right interpretations. As many words or phrases have numerous meanings, a machine translator is still unable to handle the ambiguity issues that arise and determine what the given context means. One issue that machine translators have yet to solve is the phenomena of polysemy, or numerous meanings, though it certainly isn't the only one. In this paper, instances of these machine translations of texts from several text genres—including texts from the journalistic, medical, legal, and dictionary definition genres—are examined. To clear up any misunderstanding, one must consider the situation’s background. One must also acknowledge that the situation makes sense. In the concluding section of this study, we will be shown some examples of significant issues that have arisen as a result of a lack of understanding of the term, including misunderstanding, lack of comprehension, ambiguity, and limitations on the heedless use of technical definitions in favour of common ones.

The position of machine translation in translation studies

Abstract Attempts to define the concept of translation have historically both reflected and driven developments and demarcations in the field. In light of the ubiquitous rise of machine translation (MT), the current article considers how definitional approaches to translation that preceded the MT era, and were formulated with human translation in mind, correspond with today’s MT. The article engages with two influential definitional strands in the discipline: a-priori prescriptive definitions, and descriptive definitions focused on the reception of translations. The general compatibility of both definitional approaches with MT, notwithstanding some empirical and moral criticisms pertaining to the first approach, encourages us to conceive of MT as a full-fledged translational object of inquiry, fully at home in translation studies. Finally, the article suggests that shifts in the professional status of human translators may lead to new definitions, aimed at differentiating human from machine translation by focusing on the notion of (conscious) agency.

Neural Machine Translation in Translation and Program Repair

Translation is a challenge for humans since it needs a good command of two or more languages. When it comes to computer programs, it is even more complex as it is difficult for computers to imitate human translators. With the emergence of deep learning algorithms, especially neural network architectures, neural machine translation (NMT) models gradually outperformed previous machine translation models and became the new mainstream in practical machine translation (MT) systems. Nowadays, NMT has been developing for several years and has been applied in many fields. This paper is focused on studies on four different application categories of NMT models: 1) Text NMT; 2) Automatic program repair (based on NMT); 3) Simultaneous translation. Our work provides a summary of the latest research on different applications of NMT and makes comments on their development in the future. This paper also mentioned the shortcomings of existing studies in this essay and pointed out some possible research directions.

A Survey on Evaluation Metrics for Machine Translation

The success of Transformer architecture has seen increased interest in machine translation (MT). The translation quality of neural network-based MT transcends that of translations derived using statistical methods. This growth in MT research has entailed the development of accurate automatic evaluation metrics that allow us to track the performance of MT. However, automatically evaluating and comparing MT systems is a challenging task. Several studies have shown that traditional metrics (e.g., BLEU, TER) show poor performance in capturing semantic similarity between MT outputs and human reference translations. To date, to improve performance, various evaluation metrics have been proposed using the Transformer architecture. However, a systematic and comprehensive literature review on these metrics is still missing. Therefore, it is necessary to survey the existing automatic evaluation metrics of MT to enable both established and new researchers to quickly understand the trend of MT evaluation over the past few years. In this survey, we present the trend of automatic evaluation metrics. To better understand the developments in the field, we provide the taxonomy of the automatic evaluation metrics. Then, we explain the key contributions and shortcomings of the metrics. In addition, we select the representative metrics from the taxonomy, and conduct experiments to analyze related problems. Finally, we discuss the limitation of the current automatic metric studies through the experimentation and our suggestions for further research to improve the automatic evaluation metrics.

Impact of National Institutes of Health and Food and Drug Administration Tobacco Research Funding: A Bibliometrics Analyses.

Conduct bibliometric analyses documenting the output of National Institutes of Health (NIH) tobacco-related and Food and Drug Administration (FDA) tobacco regulatory science (FDA-TRS) research portfolios. PubMed identifiers for publications between 2015 and 2020 citing tobacco funding by NIH and/or FDA were imported into NIH iCite generating measures of productivity and influence, including number of citations, journal, relative citation ratios (RCR), and comparison of research influence across Web of Science (WoS) disciplines. Coauthorship and measures of centrality among and between NIH and FDA-supported investigators gauged collaboration. Between FY 2015 and 2020, 8160 publications cited funding from NIH tobacco-related grants, 1776 cited FDA-TRS grants and 496 cited Common funding (ie, both NIH and FDA-TRS funding). The proportion of publications citing NIH grants declined while those citing FDA-TRS or Common funding rose significantly. Publications citing Common funding showed the highest influence (mean RCR = 2.52). Publications citing FDA-TRS funding displayed higher median RCRs than publications citing NIH funding in most WoS categories. Higher translational progress was estimated over time for FDA-TRS and Common publications compared to NIH publications. Authors citing Common funding scored highest across all collaboration measures. This study demonstrates the high bibliometric output of tobacco research overall. The rise in publications citing FDA-TRS and Common likely reflects increased funding for TRS research. Higher RCRs across WoS subject categories and trends towards human translation among FDA-TRS and Common publications indicate focus on research to inform regulation. This analysis suggests that FDA support for TRS has expanded the field of tobacco control resulting in sustained productivity, influence, and collaboration. This paper is the first effort to better describe the impact of tobacco research resulting from the addition of FDA funding for TRS in the past decade. The analysis provides impetus for further investigation into the publication topics and their focus which would offer insight into the specific evidence generated on tobacco control and regulation.

Human Interpretation and Machine Translations Based upon Interviews with Director Joon-ho Bong

Human Interpretation and Machine Translations Based upon Interviews with Director Joon-ho Bong

Comparing L2 translation, translation revision, and post-editing competences in translation trainees

AbstractTranslation proper is rarely the sole activity that translators undertake in today’s translation market. Translators regularly function as revisers or post-editors, requiring them to check human or machine translations to make or recommend changes to improve translation quality. Various construct and performance models of and studies into translation competence (TC), translation revision competence (TRC), and post-editing competence (PEC) exist. However, a fundamental question remains unanswered to date: how similar or different TC, TRC, and PEC are. Using indirect translations (L1 Dutch, L2 French), we collected and analyzed translation, translation revision (TR), and post-editing (PE) data from 11 graduate translation trainees. Our exploratory study shows that TRC and PEC appear to have different competences, with trainees performing better for TR than PE. However, TRC and PEC do appear to have a common core, which does not differ significantly across tasks: problem detection.

A Review of Machine Translation: Implications to Human Translators and Translation Teaching

A Review of Machine Translation: Implications to Human Translators and Translation Teaching

Design and Proofreading of the English-Chinese Computer-Aided Translation System by the Neural Network.

At present, complete machine translation (MT) cannot meet the needs of information communication and cultural exchange, and the speed of complete human translation is too slow. Therefore, if MT is used to assist in the process of English-Chinese translation, it can not only prove that machine learning (ML) can translate English to Chinese but also improve the translation efficiency and accuracy of translators through human-machine cooperation. The research on the mutual cooperation between ML and human translation has an important research significance for translation systems. An English-Chinese computer-aided translation (CAT) system is designed and proofread based on a neural network (NN) model. First, it gives a brief overview of CAT. Second, the related theory of the NN model is discussed. An English-Chinese CAT and proofreading system based on the recurrent neural network (RNN) is constructed. Finally, the translation accuracy and proofreading recognition rate of the translation files of 17 different projects under different models are studied and analyzed. The research results reveal that according to the different translation properties of different texts, the average accuracy rate of text translation under the RNN model is 93.96%, and the mean accuracy of text translation under the transformer model is 90.60%. The translation accuracy of the RNN model in the CAT system is 3.36% higher than that of the transformer model. The English-Chinese CAT system based on the RNN model has different proofreading results for sentence processing, sentence alignment, and inconsistency detection of translation files of different projects. Among them, the recognition rate for sentence alignment and the inconsistency detection of English-Chinese translation is high, and the expected effect is achieved. The design of the English-Chinese CAT and proofreading system based on the RNN can make the translation and proofreading be carried out simultaneously, which greatly improves the efficiency of translation work. Meanwhile, the above research methods can improve the problems encountered in the current English-Chinese translation, provide a path for the bilingual translation process, and have certain promotion prospects.

Sabão e água ou água e sabão? A diferença que um tradutor humano (ainda) pode fazer…

The world of professional translation is inextricably linked to the conquests of artificial intellig ence (AI), and human translators need to use machine translation tools and translatio n memories to increase their productivity . The aim of this paper is to highlig ht the importance of the human translator in the translatio n process and, consequently , to adapt the tra ining of young translators to the circumstances in which they will have to work in the future. Based on the work of final-year students of FLUP's Master's in Translation and Language Services, we carried out an analysis of several cases that revealed the weaknesses of machine translatio n systems, in o rder to illustrate how human intervention can contribute t o a more efficient tra nslation process and a higher quality of the final product and canalso help translators focus their efforts on areas where human intervention is most valuable. If the translator is aware of the limitations of AI systems in advance and if he/she knows what are the most appropria te lin guistic resources applied in the different text genres of specialised communication in the target cultu re and language, during the revision and post-editing process he/she may focus his efforts on areas where human intervention is most valuable: in producing texts that do not look like a translatio n and reveal the characteristic features of orig inal, specialised discourse in the target culture and language. By paying special attention to pronominal anaphora, collo cations, idiomatic expressions, or macro- and micro-textual conventionstranslators can optimize the revision process, increase productivity, and achieveimproved translations in less time.

The Impacts and Challenges of Artificial Intelligence Translation Tool on Translation Professionals

Machine translation, especially translation based on neural network technology has made a major breakthrough and is increasingly accepted and widely used. The development of artificial intelligence (AI) translation has had a definite impact on translation jobs. People, even professional translators, are relying on AI translation. But There is no research on whether machine translation software is superior to professional translators in translating various types of documents.. In this study, we design an experiment to determine the advantages and disadvantages between AI translations and human translations. The result shows the impact of the development of AI on the translation industry. To achieve better translation results and output highquality translations in the era of rapid development of AI, it will contribute to Human-AI partnerships.

ISF-GAN: An Implicit Style Function for High-Resolution Image-to-Image Translation

Recently, there has been an increasing interest in image editing methods that employ pre-trained unconditional image generators (e.g., StyleGAN). However, applying these methods to translate images to multiple visual domains remains challenging. Existing works do not often preserve the domain-invariant part of the image (e.g., the identity in human face translations), or they do not usually handle multiple domains or allow for multi-modal translations. This work proposes an implicit style function (ISF) to straightforwardly achieve multi-modal and multi-domain image-to-image translation from pre-trained unconditional generators. The ISF manipulates the semantics of a latent code to ensure that the image generated from the manipulated code lies in the desired visual domain. Our human faces and animal image manipulations show significantly improved results over the baselines. Our model enables cost-effective multi-modal unsupervised image-to-image translations at high resolution using pre-trained unconditional GANs.