Machine Translation

Celem niniejszego artykułu jest naświetlenie uwarunkowań związanych z korzystaniem z narzędzi do tłumaczeń maszynowych w procesie obsługi tłumaczeniowej polskich instytucji rządowych najwyższego szczebla oraz konieczność ich uwzględniania na etapie postępowania o udzielenie zamówienia publicznego w kontekście przesłanki odrzucenia oferty przewidzianej w art. 89 ust. 1 pkt 7d ustawy z dnia 29 stycznia 2004 r. – Prawo zamówień publicznych (Dz.U.2019.1843 t.j. z dnia 27.09.2019 r.).

Many Natural Language Processing (NLP) applications involve Named Entity Recognition (NER) as an important task, where it leads to improve the overall performance of NLP applications. In this paper the Deep learning techniques are used to perform NER task on Hindi text data as it found that as compared to English NER, Hindi language NER is not sufficiently done. This is a barrier for resource-scarce languages as many resources are not readily available. Many researchers use various techniques such as rule based, machine learning based and hybrid approaches to solve this problem. Deep learning based algorithms are being developed in large scale as an innovative approach now a days for the advanced NER models which will give the best results out of it. In this paper we devise a Novel architecture based on residual network architecture for preferably Bidirectional Long Short Term Memory (BiLSTM) with fasttext word embedding layers. For this purpose we use pre-trained word embedding to represent the words in the corpus where the NER tags of the words are defined as the used annotated corpora. BiLSTM Development of an NER system for Indian languages is a comparatively difficult task. In this paper, we have done the various experiments to compare the results of NER with normal embedding and fasttext embedding layers to analyse the performance of word embedding with different batch sizes to train the deep learning models. Here we present a state-of-the-art results with said approach F1 Score measures.

Machine translation is being carried out by the researchers from quite a long time. However, it is still a dream to materialize flawless Machine Translator and the small numbers of researchers has focussed at translating Marathi Text to English. Perfect Machine Translation Systems have not yet been fully built because the fact that languages differ syntactically as well as morphologically. Majority of the researchers have opted for Statistical Machine translation whereas in this paper we have addressed the challenges of Rule based Machine Translation. The paper describes the major divergences observed in language Marathi and English and many challenges encountered while attempting to build machine translation system form Marathi to English using rule based approach. As there are exceptions to the rules and limit to the feasibility of maintaining knowledgebase, the practical machine translation from Marathi to English is a complex task.

This paper focuses on a historical parallel corpus of Old Occitan and English. The 13th century Occitan narrative poem Roman de Flamenca holds a unique position in Provençal literature, and is a “universally acknowledged masterpiece of Old Occitan narrative” (Fleischmann, 1995). Flamenca’s unique character is revealed not only in its genre (narrative poem) and character portrayal, but also in its use of multiple dialogues, monologues, and its lexical variation. It clearly has a profound value for studies of language, stylistic, and social variation. First, we align the Occitan text with its English translation (Blodgett, 1995) using automatic alignment (Liang et al., 2006) and a lexicon-based approach, NATools.19 Second, we describe the morpho-syntactic and syntactic annotation of the Old Occitan, which was carried out via a resource-light method. Both types of information were annotated based on the annotations of an Old French corpus. Finally, we show how historical investigations may benefit from such a parallel corpus. The basic tenet of the parallel approach is that the original text and its translation equivalent have similar underlying contexts and functions (Enrique-Arias, 2012). Thus, using a translation from English, a non pro-drop language with a fixed word order, we can observe certain linguistic factors that would be otherwise overlooked. In order to illustrate the usefulness of this corpus, we investigate stylistic and social variation of subject pronouns, namely their presence and absence in Old Occitan. We show how different social, syntactic, and stylistic factors influence subject pronoun variation.

Translator behavior criticism theory is regarded as one of the main current translation studies theory in China. It is proposed by Professor Zhou Lingshun in 2010 and now it has been 10 years. This article starts from translators' position as the center, which provides the background for this theory. Then it moves on to make clear the main proposals and practice of this theory, especially his "truth-seeking and utility attaining" continuum mode of evaluation. Through this continuum mode, one can measure the degree of rationality of a translator's behavior by taking various factors into consideration, hence the evaluation of the translated version. After that this article talks about the significance and limitations of this theory. His theory is a breakthrough in translation studies, but still needs to be further polished.

This research examines the translation of the swear words of a Turkish film (GORA, 2004) into English subtitles. So far, the most closely related research on translation of swearing in the Turkish Soler-Pardo, 2015). The main reason for this paper is the current research gap in translating the Turkish swear words into another language (English). Throughout this research, the English subtitle translation of Cem Yılmaz's GORA (2004) was picked on Netflix and parts which contained swear words were documented on a .txt file, were marked and analyzed using swear words categories of Pinker (2007). Ameri & Ghazizadeh's (2014) swear word translation model, as well as Harvey's (2000) technique of 'functional equivalence' (2000, pp. 2-6) presented by the four main culture-based concepts of translation technique was used for the analysis. This article aims to explore and answer the following questions; (i) How have the Turkish swear words in GORA (2004) been translated into English? (ii) How and for what reason did changes in translation occur? The hypothesis of this research is as follows; (a) there will be a loss of meaning and function due to smoothing during the translation of swear words; (b) cultural contraction will be observed during the translation of swear words. The findings confirm the hypotheses and support that the words of curses are either completely omitted from the frame or changed in meaning by being softened. This shows that the translations also adopted a target language-oriented strategy during translation.

In order to solve the serious lack of information in ethnic minority languages, the paper proposes a machine translation system from Vietnamese into ethnic minority language in the restricted areas such as the weather forecast, forest fire warning, policies and laws of the government, farming experience, animal husbandry experience, natural disaster prevention methods, etc. The translation method is based on Vietnamese-ethnic minority language bilingual corpus in which the entries are indicated the field for ethnic minority language terminology. Applying this method, a machine translation system from Vietnamese into Ede language has been built for translating the weather bulletins by using a Vietnamese-Ede bilingual corpus with Ede language terminology in nature, geography, hydrometeorology, and weather forecast. The corpus has open structure and the major axis is in Vietnamese, that creates the ability to extend to other ethnic minority languages such as Cham, Ka Tu, Jarai, Muong etc. The initial tested results with the Dak Lak Radio and Television station are very positive. Besides, the proposed method has also contributed to solve the ambiguity of the word boundary, homonym word and polysemous word.

 Abstract-Processing the Vietnamese ethnic minorities Language in general and the Ede language in particular face many difficulties. So far there are few experts or scientists researching in this field. Especially, in the context of booming use of the internet and the preservating of ethnic minority languages are an urgent issue in the trend of cultural development and integration of ethnic minorities in Vietnam. In order to overcome the current situation of Vietnam ethnic minority languages in general and the Ede language in particular, deploy practical applications, to create conditions for further upgraded Ede language teaching quality, to help teachers in compiling Vietnamese-Ede bilingual lesson plans. The paper proposes building the application to support for Vietnamese-Ede machine translation basing on the Vietnamese-Ede bilingual vocabulary database. The translation system integrates the aid of users for processing ambiguous words in Ede language. This application has contributed the upgrading of the Ede language teaching levels and enriching Vietnamese-Ede bilingual subjects. Index Terms-machine translation, Vietnam ethnic minority, bilingual corpus, Vietnamese-Ede bilingual vocabulary database, Ede language.

Resume Les progrès technologiques et la mondialisation ont permis de rapprocher les universitaires de diverses disciplines sur le plan socioculturel. En raison de l'interdisciplinarité et de ces progrès dans le monde, de nouvelles approches théoriques de la traduction au XXIe siècle ont évolué, ce qui a conduit à des changements dans les attitudes à l'égard de la traduction. Cette étude portera sur les approches les plus récentes et les plus importantes en matière d'études de traduction au 21 ème siècle. La recherche vise à analyser les différents virages des études de traduction en mettant l'accent sur le virage qui est dû à la mondialisation. Il y a aussi une analyse des changements de perspective pour le virage dans les études en traduction. L'objectif de la recherche est d'analyser les différents changements de la traduction qui sont dus à la mondialisation ainsi que l'altération de l'attitude et du rôle des traducteurs afin de visualiser le futur de la traduction et ses développements. Cela permet de comprendre comment les théoriciens ont trouvé le dernier virage dans la discipline. Afin de fournir une analyse complète du changement d'attitude à l'égard de la traduction, il y a une étude approfondie des développements récents de la traduction qui sont principalement dus aux avancées technologiques. Celles-ci comprennent entre autres le Crowdsourcing, la traduction automatique en temps réel, la localisation, et la traduction assistée par l'ordinateur. Cette étude est descriptive, explicative, analytique et critique dans son examen des tendances de la traduction, elle suit un cadre de recherche théorique et analytique. Les virages de la discipline sont analysés dans un contexte socio-culturel et, à partir de cette information, le chercheur se rend compte de leur importance pour le développement de la traduction. La conclusion de l'étude est que la discipline croissante et inconstante de la traduction s'est heurtée à un certain nombre de menaces au 21ème siècle. Pour résoudre ces problèmes, toutes les organisations devraient être unies pour contribuer à la professionnalisation des traducteurs, afin de fixer des normes de qualité en tenant compte à la fois de l'impact de la technologie récente, du contexte et de l'effet de l'interdisciplinarité.

In the communication age we are in, the importance of translation programs and machine translation systems is increasing day by day. Google Translation which is one of the most preferred translation programs, has been examined in this research. The aim of the study is to examine the target texts created with Google Translation in terms of different text types regarding linguistics and evaluate the translations through error analysis. The data analysis techniques used in the research are based on the principles laid out by Rinsche (1993) and developed by Schäfer (2002) and Ramlow (2009). In the study, three different texts in German were determined according to their difficulty levels and these texts were translated into Turkish via Google Translation. As a result of the analysis, the source text and the translation text were compared and it was determined whether the difficulty level of the source text affects the quality of the target text in the translation process and the findings obtained from each text were evaluated in the light
of the questions at the starting point of the research.

Audiovisual translation (AVT) is a rather newly explored sub-discipline of translation studies. This paper compares the micro and macro translational strategies adopted in two types of AVT, namely dubbing and fansubbing. The corpus of study is a Disney cartoon, Toy Story1 dubbed from English into two Arabic varieties; Egyptian colloquial Arabic and Modern Standard Arabic (MSA), as well as a fansubbed version translated from English into MSA. The functional approach is realized by employing the points of analysis suggested by Chaume's (2004) model of analyzing audiovisual material and the problematic areas identified by Chiaro (2009), using the micro translational strategies pinpointed by Pedersen's (2005) extralinguistic cultural-bound references (ECRs) analytical model. The results of the research indicate that the fansubbed version tends to adopt a foreignized macro strategy with the direct or literal translation strategy as the prime micro source language (SL) strategy. The Egyptian colloquial Arabic dubbed version tends to adopt a domesticated macro strategy with the substitution strategy as the prime micro target language (TL) strategy. Finally, the MSA dubbed version tends to adopt a foreignized macro strategy with the direct translation strategy as the adopted main micro SL strategy.

The more language service companies (LSCs) include machine translation post-editing (MTPE) in their workflows, the more important it is to know how the PE task is performed, who the post-editors are, and what skills they should have. This research is designed to address such questions. It aims to deepen our knowledge of current practices to later create new training content and adapt existing training methodologies to different types of audiences. Based on the results of a survey of LSCs and other companies who currently use MTPE, we present a picture of evolving practices in the contemporary European MTPE market, and opinions held about this emerging métier. Our research finds that a high level of expertise in MTPE may not necessarily be indicative of the industry, and that the post-editor of MT has a multi- and transdisciplinary profile.

الحمد لله الذي خلق الإنسان وعلمه البيان، وخصصه بالقدرة الكلامية بين سائر مخلوقاته، ثم الصلاة والسلام على سيدنا محمد وعلى آله وصحبه أجمعين.
فهذا بحث - بعنوان نظريات تشومسكي اللسانية - عن أحد أعلام اللسانيات الحديثة في العالم، وهو العلامة الأمريكي المعاصر أفرنم نوم تشومسكي، ويهدف البحث إلى إبراز جهوده اللسانية التجددية التي أوجدت ثورة علمية -وما زالت- على المدارس والمناهج اللسانيات قبله، بداية من رائد اللسانيات الحديثة دي سوسير إلى اليوم. ويتناول البحث أهم مفاهيم نظريات تشومسكي، ومراحل تطورها وتوظيفها في اللسانيات التطبيقية وعلوم أخرى.
وقد قسمت البحث إلى مقدمة وتمهيد وخمسة فصول وخاتمة، ثم ذيلته بثبت المراجع وفهرس الموضوعات.
المقدمة: وفيه بيان هدف البحث وخطته.
التمهيد: وفيه نبذة يسيرة عن حياة تشومسكي ومؤلفاته
الفصل الأول: ما قبل نظريات تشومسكي، وفيه ثلاثة مباحث:
المبحث الأول: المدرسة الأمريكية اللسانية قبل تشومسكي
المبحث الثاني: علاقة تشومسكي بالمنهج البنيوي
المبحث الثالث: المنطلقات الفكرية للنحو التوليدي عند تشومسكي
الفصل الثاني: مبادئ وأسس النحو التوليدي
الفصل الثالث: مكونات القواعد التوليدية والتحويلية

الفصل الرابع: مراحل نظريات تشومسكي
الفصل الخامس: آثر نظريات تشومسكي، وفيه أربعة مباحث:
المبحث الأول: تأصيل بعض أراء تشومسكي عند العرب
المبحث الثاني: مميزيات نظريات تشومسكي
المبحث الثالث: المآخذ على نظريات تشومسكي
المبحث الرابع: تطبيق نظريات تشومسكي في اللسانيات وعلوم أخرى
الخاتمة: وفيها أهم النتائج التي توصل إليها البحث.
منهج البحث: استخدم البحث المنهج الوصفي التحليلي، وقام البحث بجمع فهم الباحثين لنظريات تشومسكي من خلال بعض كتبه المترجمة والمؤلفات حول نظرياته، وتطرق البحث إلى تناول مجالات توظيف بعض نظريات تشومسكي المباشر وغير المباشر في اللسانيات التطبيقية وعلوم أخرى. والبحث عبارة عن عرض مختصر لنظريات تشومسكي المتجددة، وقد ألف أكثر من عشرين كتب ودراسات عديدة يعزز فيها هذه النظريات، فلا يمكن جمعها في مثل هذه وريقات. وأسال الله تعالى أن يجعله عملا نافعا خاصا لوجه الكريم. والحمد لله في الأول والآخر.

The concept of transfer grammar is not a recent phenomenon. Even in 1954 Harris discusses about the importance transfer grammar in the context of translation, machine translation, language teaching and language learning.  We became aware of it recently due to our involvement in machine translation (MT).  Presently we are interested in preparing a transfer grammar for English Tamil MT.  Harris has proposed an elaborate methodology to prepare a transfer grammar (His idea of transfer grammar has been explained below under the heading “Transfer Grammar”. Here we narrow down our efforts just to correlate syntactic structure of English with that of Tamil from the point of view preparing a transfer grammar for English-Tamil machine translation. For this purpose the computable syntactic structures of English and Tamil have been worked out. These computational syntactic structure analyses are different form ordinary syntactic structure analyses in the sense that the computational syntactic structures are viable for computational processing. A transfer grammar for machine translation has to be prepared using these computational syntactic structure analyses. A transfer grammar is an important component in a machine translation system. This helps us to map one language structure into another language structure.
As English and Tamil belongs to two different types of language groups, that is English as predominantly SVO patterned language and Tamil a predominantly SOV patterned language showing unique characteristics which differentiate them drastically from one another, it is possible to manipulate these differences to form transfer rules. These transfer rules can be used to map the English structure into Tamil Structure and vice versa.

Natural Language Processing is a programmed approach to analyze text that is based on both a set of theories and a set of technologies. This forum aims to bring together researchers who have designed and build software that will analyze, understand, and generate languages that humans use naturally to address computers.

Por una parte, el objetivo de esta presentación es intentar explicar algunas nociones de la teoría de los sistemas de gramáticas referidas a algunos aspectos de las lenguas y lenguajes naturales. Y, por otra parte, podemos plantearnos una fascinante pregunta: ¿cómo computa la naturaleza? De manera similar al caso de la "computación natural", ¿es posible hablar de "traducción natural"? ¿Qué significa exactamente esta expresión?

In order to be able to understand how learning takes place in a given field it is necessary to accurately ascertain how people solve problems in that same area. The first step to do so is to try to reconstruct the mental process, or the programmes, which individuals use to solve a specific problem. In addition to problems specific to each subject matter the findings of research conducted within Cognitive Science point to the existence of cognitive or instructional cloggings of a general nature: very few students manage to progress from low-level skills to high-level skills (from basic and memorization skills to versatile high-level skills pertaining to an advanced rank). In this article we intend to identify those cognitive or instructional cloggings and to discover which of them are specific to computer science applied to translation. © Ciências & Cognição 2006; Vol. 09: 42-55.