Evaluating the Impact of Integrating Similar Translations into Neural Machine Translation

Previous research has shown that simple methods of augmenting machine translation training data and input sentences with translations of similar sentences (or fuzzy matches), retrieved from a translation memory or bilingual corpus, lead to considerable improvements in translation quality, as assessed by a limited set of automatic evaluation metrics. In this study, we extend this evaluation by calculating a wider range of automated quality metrics that tap into different aspects of translation quality and by performing manual MT error analysis. Moreover, we investigate in more detail how fuzzy matches influence translations and where potential quality improvements could still be made by carrying out a series of quantitative analyses that focus on different characteristics of the retrieved fuzzy matches. The automated evaluation shows that the quality of NFR translations is higher than the NMT baseline in terms of all metrics. However, the manual error analysis did not reveal a difference between the two systems in terms of total number of translation errors; yet, different profiles emerged when considering the types of errors made. Finally, in our analysis of how fuzzy matches influence NFR translations, we identified a number of features that could be used to improve the selection of fuzzy matches for NFR data augmentation.

Lexical Errors in English Students' Writing

This research analyzed lexical erros in writing of English students and focused on three types of errors; they were errors in literal translation, errors of word formation, and errors in word choice. This research aimed to explore the most frequent errors and the factors of errors made by English students. The design of this research was qualitative in which the data were colleted by using document in the form of students’ writings and interview. The informants of the research were thirty students majoring in English study program at Tadulako University. The result of this research showed that the most frequent errors made were in word choice and followed by errors in literal translation and errors in word formation. The factors causing the errors were students’ inability to consider the context of the topics when writing. They only focused on using dictionary, Google translate, or simply the words that they had in mind to express ideas in their writing. The students also never used words that were related to the topic in their daily interaction.

Machine translation literacy and undergraduate students in applied languages: Report on an exploratory study

This article reports on an exploratory study conducted on applied languages undergraduate students’ use of machine translation. Starting from the observation that they make extensive use of free tools available online, our aim was to understand whether they are capable of identifying and correcting machine translation errors, and if so, to what extent.

ANALYZING THE TRANSLATION OF INDONESIAN CULTURAL TERMS IN NOVEL ‘TARIAN BUMI’ TRANSLATED INTO ‘EARTH DANCE’

This research aims to obtain a deeper understanding of the Indonesian cultural terms in the novel Tarian Bumi and their translation into English, Earth Dance. This research focuses on the use of translation strategies, equivalences, errors, and factors affecting the errors. It uses a qualitative approach with the content analysis method. Both of the novels were read to find data. The collected data was divided, analyzed, and compared with their translation. The data analysis and explanation indicate that the strategy used in the translation of the cultural terms is a translation procedure amounted to 11 procedures, couplets and triplets are also used. The translation is oriented towards the source language because the translator often used procedures and equivalence oriented to the source language. The result of this research is also to look for translation errors of cultural terms and factors affecting the errors and to find how to solve them. It is caused by knowledge of the translator and different cultures. This research will give some benefits for those who are interested in studying the translation theory and other students who want to research Indonesian cultural terms translated into English.

Exploration of Effective Attention Strategies for Neural Automatic Post-editing with Transformer

Automatic post-editing (APE) is the study of correcting translation errors in the output of an unknown machine translation (MT) system and has been considered as a method of improving translation quality without any modification to conventional MT systems. Recently, several variants of Transformer that take both the MT output and its corresponding source sentence as inputs have been proposed for APE; and models introducing an additional attention layer into the encoder to jointly encode the MT output with its source sentence recorded a high-rank in the WMT19 APE shared task. We examine the effectiveness of such joint-encoding strategy in a controlled environment and compare four types of decoder multi-source attention strategies that have been introduced into previous APE models. The experimental results indicate that the joint-encoding strategy is effective and that taking the final encoded representation of the source sentence is the more proper strategy than taking such representation within the same encoder stack. Furthermore, among the multi-source attention strategies combined with the joint-encoding, the strategy that applies attention to the concatenated input representation and the strategy that adds up the individual attention to each input improve the quality of APE results over the strategy using the joint-encoding only.

Translating Multicultural Texts: Challenges and Solutions

Translating literary texts involves decoding not only the language but also the cultural elements integrated into the text. The situation becomes even more challenging if the translated text is multicultural. The original text may be full of references, allusions and subtexts that can present certain difficulties even for native speakers. The translator faces the double challenge: not only to convey the plot, which is the least difficult part but make the reader feel the subtleties and nuances of the cultures presented in the text. Cultures are not necessarily associated with different ethnic and religious identities. They may also refer to cultures of certain periods in history, cultures of age groups, cultures of local communities. Oversimplifications made by the translator rob the reader of the pleasure of reading and may create a distorted image of the writer and the text. We aim to analyse typical translation errors made by translators of contemporary Russian fiction into English. The analysis is based on Narine Abgaryan’s Three Apples Fell from the Sky and Dmitry Novikov’s A Flame Out in the Sea, both originally written in Russian and characterised by a high degree of multiculturalism. We also suggest practical ways to overcome the difficulties arising in the process of translating multicultural texts.

Development of translation skills among Chinese students in the process of professional training of linguists majoring in “Russian as a foreign language”

Перевод — комплексная дисциплина, требующая систематического развития всех видов речевой деятельности и предполагающая сложную поэтапную работу с учащимися. Особенность обучения переводу состоит в том, что образовательный процесс происходит в условиях двуязычия и осуществляется средствами двух языков. Формирование навыков перевода происходит в процессе выполнения сложного комплекса упражнений, учитывающих специфику переводческой деятельности: микрореферирование, перевод-пересказ, реферативный письменный и устный перевод, составление аннотации и др. Необходимо использование упражнений, ориентированных на развитие аналитических навыков: определение достоверности перевода, выбор подходящих вариантов перевода, сопоставительное изучение переводческих ошибок и др. В тренировочный комплекс включаются упражнения на переводческие трансформации: перевод с применением приемов калькирования, конкретизации, генерализации, антонимического перевода и др. Совершенствование переводческих навыков учащихся происходит одновременно с формированием лингвокультурологической компетенции, знакомством с языковой картиной мира русских. При разработке заданий целесообразной представляется работа с культурно-специфическими словами, «ложными друзьями переводчика», что способствует расширению языковых и фоновых знаний студентов, знакомству с реалиями и формированию необходимых переводческих навыков и умений. Основными задачами обучения являются: а) знание языкового материала, доведенное до уровня автоматического перевода; б) формирование фоновой эрудиции переводчика. Translation is a complex discipline that requires systematic development of all types of speech activity and involves complex step-by-step work with students. The specific feature of teaching translation is that the educational process occurs in a bilingual environment and applies two languages. Translation skills are formed in the process of performing a complex set of exercises that take into account the peculiarities of translation activities. It is necessary to use exercises focused on the development of analytical skills, for instance, determining the reliability of the translation, selecting the appropriate translation options, and comparative study of translation errors. The training complex includes exercises for translation transformations, for example, loan translation, concretization, generalization, and antonymic translation. The ability to operate with a large amount of information develops on the basis of tasks that train memory, such as micro-referencing, translation-retelling, abstract translating and interpreting, and writing summaries. In the process of training translators, special attention is paid to the formation of background erudition. The improvement of students' translation skills is concurrent with developing linguistic and culturological competence along with learning the linguistic picture of the Russian world. When designing assignments, it seems expedient to work with culturally sensitive words and “false friends of the translator”, which contributes to the growth of the students’ linguistic and background knowledge, exploration of the realities and the formation of the necessary translation skills and abilities. The main objectives of training are knowledge of the language material, brought to the level of automatic translation as well as the formation of the background erudition of the translator.

The Hieroglyphic Luwian genitive case

Descriptions of Hieroglyphic Luwian grammar assert that the genitive endings ‑as(a) and ‑asi are interchangeable; their distribution is said to be random rather than governed by any conditioning factor. However, recent studies have shown that the ending ‑asi is geographically and chronologically restricted in the corpus of Hieroglyphic Luwian texts, which suggests that the distribution is not entirely random. In this paper, it is argued that in texts from the “‑asi area” - where texts containing both endings are found - the genitive endings are distributed in a non-random way. Genitives in ‑as(a) are dependent on neuter gender head nouns, whereas genitives in ‑asi are dependent on common gender head nouns. This means that, syntactically, Hieroglyphic Luwian genitives resemble genitival adjectives by agreeing with the gender of their head. Although several counterexamples exist, they are probably caused by translation errors, by language change in late Hieroglyphic Luwian, and possibly by the fact that -as(a) may reflect -asa as well as -as. Finally, a new account of the historical development of Luwian genitives is presented.