Translation Memory

This dissertation aims to shed light on the impact of artificial intelligence and machine learning on translators. The current hype surrounding the recent progress in the field of artificial intelligence in this era of technology disruption has been welcomed very negatively by many translators, who feel their profession and position in society threatened by machines. Nevertheless, the figures demonstrate that this fear is unnecessary, since the number of job losses due to automation will be more than counterbalanced by the new ones where new skills will be required. The goal of the present work is assessing the progress of technology in the field of machine translation (MT) by comparing the performance of different MT tools, namely Systran, SDL Trados, DeepL, Google Translate and Amazon Translate, in the translation of snippets of text extracted from an e-commerce website. By illustrating with practical examples the improvements of neural machine translation -which bases its strength on machine learning -it is observed that the human factor will always be vital in translation, despite the high-quality output of neural MT tools. It is discussed that translators should rely on

This dissertation aims to shed light on the impact of artificial intelligence and machine learning on translators. The current hype surrounding the recent progress in the field of artificial intelligence in this era of technology disruption has been welcomed very negatively by many translators, who feel their profession and position in society threatened by machines. Nevertheless, the figures demonstrate that this fear is unnecessary, since the number of job losses due to automation will be more than counterbalanced by the new ones where new skills will be required. The goal of the present work is assessing the progress of technology in the field of machine translation (MT) by comparing the performance of different MT tools, namely Systran, SDL Trados, DeepL, Google Translate and Amazon Translate, in the translation of snippets of text extracted from an e-commerce website. By illustrating with practical examples the improvements of neural machine translation -which bases its strength on machine learning -it is observed that the human factor will always be vital in translation, despite the high-quality output of neural MT tools. It is discussed that translators should rely on

This thesis presents a study of the impact a translation memory system may have over the translation process in terms of cognitive segmentation, lexicogrammatical variation (agnation) and explicitation, when professional translators translate from German into Portuguese. The framework of the investigation attempts to merge the theoretical dichotomy between process and product, taking them as a single phenomenon, as the product is considered the final record of a complex process. Six professional translators took part on the experiment, being each of them submitted to two translation tasks -the first without a Translation Memory System, and the second one with it. The task realization in both environments was mapped (Dragsted, 2004) through pause analyses (DRAGSTED, 2004;, and interpreted through a systemic functional frame STEINER, 2001).The translation memory system's impact over the process is evaluated and the common sense assumption of interlinguistical standardization conditioned by the program is empirically investigated. Three linguistic categories are focused, namely nominalization, participle constructions, composed nouns and verbs with separable prefixes. The results show that, although the software tends to standardize not only choices, but also segmentation patterns, natural segmentation tends to impose itself. In many cases, translators negotiate their own choices with the options proposed by the memory. The analysis of this negotiation was focused on instances of explicitation, considered under the concept of grammatical metaphor, through its identification in the segmentation process of the target text, as well as through the contrast of different realizations among the translators.

Artificial Intelligence-generated translation tools are proliferating at an unprecedented pace. Translation job markets are changing and changing the required skills of graduates. The debate about replacing human translators by machines is intensifying. Questions about what translation education can offer students to be employable are legitimately raised by educators, students and their families, researchers, educational institutions, etc. Between the two poles of protagonists who promote for AI integration into education and opponents, many issues may unfold. For instance, is AI really useful or risky in educational contexts? How to stay 'connected' in a digital age and acquire the necessary skills for a dream job? Is the field going to be directed by the job market, or more accurately the international global trend of integrating AI into translation? Amid this turbulent, controversial environment, the present study aims to investigate the integration of Artificial Intelligence into translation education from a transdisciplinary approach to translation education and industry, taking educational theories, translation studies, politics and Artificial Intelligence as its point of departure to a discussion of the benefits and risks of that integration.

This project describes the design and implementation of a modular, filebased encryption and decryption system developed using modern C++ (C++17). The system is designed to securely encrypt and decrypt files stored within a directory by applying a key-based encryption algorithm. It supports recursive directory traversal, batch file processing, binary file handling, and memory-safe execution using smart pointers. The project emphasizes clean software architecture, separation of concerns, robust error handling, and extensibility. The internal task queue mechanism prepares the system for future enhancements such as multi-threaded and process-level parallel execution, making it scalable and efficient for large file sets.

Machine translation outputs are not correct enough to be used as it is, except for the very simplest translations. They only give the general meaning of a sentence not the exact translation. As Machine Translation (MT) is gaining a position in the whole world, there is a need for estimating the quality of machine translation outputs. Many prominent MT-Researchers are trying to make the MT-System, that produces very good and accurate translations and that also covers maximum language pairs. If good translations out of all translations can be categorized then the time and cost can be saved to a great extent. Now, Good quality translations will be sent for post-editing and rest will be sent for pre-editing or retranslation. In this paper, Kneser Ney smoothing language model is used to calculate the probability of machine translated output. But a translation cannot be said good or bad. Based on its probability score there are many other parameters that effect its quality. The quality of...

The growing use of computer technologies as well as the amount of available information has posed new challenges to humanmachine interaction. Applying natural language processing techniques to information systems has brought advantages to this emerging field. Question-answering (QA) systems have been developed in order to ease the information access process. However, most of today's QA systems retrieve a collection of documents whose contents may fulfill the answer to user's queries. Such systems usually consider unstructured information sources, namely Web or textual documents. Adopting a different approach, INQUER, the QA system described in this paper, makes use of a structured knowledge base: WordNet.PT (Portuguese WordNet). This system provides direct natural language answers to user's questions by applying inference and information extraction mechanisms that interact with WordNet.PT (Portuguese WordNet). Although INQUER deals specifically with Portuguese, the theoretical approach that underlies this application is language independent.

In the literature and discussion about information retrieval (IR) the words "string", "word", "term" and "concept" are used inconsistently. There is no common agreement on the concepts named by these words. Thus the communication in IR is inefficient, inaccurate and prone to errors. Based on the concept of term in the study of terminology and on a three level model of IR, the paper seeks to clarify the terminology and proposes the terms string, search key and concept, with their subterms, for IR. Consequently it is proposed that a number of other words commonly used, such as "controlled term", "uncontrolled term", "free term", should not be used in IR.

Unités de traduction et identité du traducteur : l'impact des outils d'aide à la traduction Résumé Les mémoires de traduction (MT) sont bien implantées dans le monde de la traduction spécialisées et font partie intégrante de la formation des futurs traducteurs. Néanmoins, elles gardent une mauvaise réputation, parce qu'elles forcent à travailler de manière fragmentée, au niveau du segment. Dès lors, les critiques subjectives ou objectives portent sur la tarification au mot, sur la décontextualisation et sur la dépossession de sa propre production. En forçant le trait, on pourrait dire qu'une traduction n'est plus un texte, mais un ensemble de segments et de termes. Nous allons ici étudier ces critiques, les problèmes techniques, les implications au niveau du droit d'auteur, ainsi que l'impact cognitif. Entre acceptation fataliste et révolte délétère, il y a des solutions techniques à des problèmes concrets, des améliorations à opérer aux MT, mais aussi, sans doute, d'autres modes de saisie, comme la dictée vocale, ou un autre processus de traduction dans son ensemble, pour retrouver la dimension discursive du texte, et en final, l'identité du traducteur.

The poor quality of user-generated content (UGC) found in forums hinders both readability and machine-translatability. To improve these two aspects, we have developed human-and machine-oriented pre-editing rules, which correct or reformulate this content. In this paper we present the results of a study which investigates whether pre-editing rules that improve the quality of statistical machine translation (SMT) output also have a positive impact on post-editing productivity. For this study, pre-editing rules were applied to a set of French sentences extracted from a technical forum. After SMT, the post-editing temporal effort and final quality are compared for translations of the raw source and its pre-edited version. Results obtained suggest that pre-editing speeds up post-editing and that the combination of the two processes is worthy of further investigation. J'ai redémarrer l'ordi (apparition de la croix rouge) mais pas besoin de restaurer le système:Toute ces mises à jour on été faite le 2013-03-13

This book series tackles prominent issues in Machine Translation (MT) at a depth which will allow these books to reflect the current state-of-the-art, while simultaneously standing the test of time. Each book topic will be introduced so it can be understood by a wide audience, yet will also include the cutting-edge research and development being conducted today. MT is being deployed for a range of use-cases by millions of people on a daily basis. Google Translate and Facebook provide billions of translations daily across many language pairs. Almost 1 billion users see these translations each month. With MT being embedded in platforms like this which are available to everyone with an internet connection, one no longer has to explain what MT is on a general level. However, the number of people who really understand its inner workings is much smaller. The series includes investigations of different MT paradigms (Syntax-based Statistical MT, Neural MT, Rule-Based MT), Quality issues (MT evaluation, Quality Estimation), modalities (Spoken Language MT) and MT in Practice (Post-Editing and Controlled Language in MT), topics which cut right across the spectrum of MT as it is used today in real translation workflows.

Writers and translators have difficulties treating proper nouns correctly. These designations represent concepts that are very likely not common knowledge. While humans can research, machines can only apply data provided. It is therefore important that proper nouns are documented in term bases and made available to MT engines.

The present study investigates routing of Iranian translator training procedures based on 'Dick and Carey model'. Since the development of psychology texts in theoretical and practical grounds is considered as an urgent need, translators and interpretators must use National Accreditation Authority to maintain the quality of professional translation (Sun, 2015). They can resolve the problem of translation practice and instruction or various demands of translation due to multiculturalism problems using the modeling of an instructional flowchart/diagram and utilizing of it as step by step through the rationale of training sessions. To this end, the researcher intends design it as a main module till the translator can be used the method even online during their task and learning steps, and the 5 faculty members must reach an agreement on confirmation of diagram. A list of psychological vocabularies will be attached to the diagram. At the end of the statements, the design proposal of this study makes it possible to draw the considered flowchart on a computer; even to be used or combined in software training applications as an advanced method.

This paper evaluates the impact of ma-chine translation on the software localiza-tion process and the daily work of profes-sional translators when SMT is applied to low-resourced languages with rich mor-phology. Translation from English into six low-resourced languages (Czech, Es-tonian, Hungarian, Latvian, Lithuanian and Polish) from different language groups are examined. Quality, usability and applicability of SMT for professional translation were evaluated. The building of domain and project tailored SMT sys-tems for localization purposes was evalu-ated in two setups. The results of the first evaluation were used to improve SMT systems and MT platform. The second evaluation analysed a more complex situ-ation considering tag translation and its effects on the translator’s productivity.

This paper reports the latest development of The Halliday Centre Tagger (the Tagger), an online platform provided with semi-automatic features to facilitate text annotation and analysis. The Tagger is featured for its web-based architecture with all functionalities and file storage space provided online, and a theory-neutral design where users can define their own labels for annotating various kinds of linguistic information. The Tagger is currently optimized for text annotation of Systemic Functional Grammar (SFG), providing by default a pre-defined set of SFG grammatical features, and the function of automatic identification of process types for English verbs. Apart from annotation, the Tagger also offers the features of visualization and summarization to aid text analysis. The visualization feature combines and illustrates multi-dimensional layers of annotation in a unified way of presentation, while the summarization feature categorizes annotated entries according to different S...

Semantic knowledge has been adopted recently for SMT preprocessing, decoding and evaluation, in order to be able to compare sentences based on their meaning rather than on mere lexical and syntactic similarity. Little attention has been paid to semantic knowledge in the context of integrating fuzzy matches from a translation memory with SMT. We present work in progress which focuses on semantics-based pretranslation before decoding in SMT. This involves applying fuzzy matching metrics based on lexical semantics and semantic roles, aligning parse trees based on semantic roles, and pretranslating matching source sentence parts using aligned tree nodes.

The current dynamic, rapidly changing labour market is influenced by the constant development of new technologies, globalisation, the changing nature of the economy, and changed demands on employees. In this context, university graduates are expected to be flexible, dynamic, and able to adapt effectively to new, rapidly changing conditions in the labour market, changes in individual job positions, and the changes and challenges that 21st-century society is undergoing. These circumstances force universities to respond to the situation as employers point to the disconnect between students' university training and practice. According to them, students come unprepared to a contemporary working environment, their skills and knowledge not corresponding with the needs of practice. The field of study of philology is no exception in this regard. The position of the translator (in the near future) will be different, their tasks will be more diverse, and they will be required to have different competencies and skills. The translator will have to interact with other experts or participants in the translation process and will have to be a team player who is proficient in using IT. This paper focuses on the role of interpersonal skills in the development of translation competence. The first part defines key terms: knowledge, skills, and interpersonal skills; in the second one, a teaching model for specialised translation is introduced. This model also enables the monitoring of the development of interpersonal skills in the process of acquiring translation competence, while also revealing a change in the paradigm of teacher-student interaction.

This paper reports the results of two studies carried out with two different group of professional translators to find out how professionals perceive and accept NMT in comparison with TM. The first group translated and post-edited segments from English into German, and the second group from English into Spanish. Both studies had equivalent settings in order to guarantee the comparability of the results. It will also help to shed light upon the real benefit of SMT from which translators may take advantage

This paper highlights the results and trends on post-editing and machine translation from the recent AMTA and SDL Automated Translation Survey. Then Continental Airlines and SDL share their experiences, and the benefits and challenges of human postediting.

This paper describes a database of translation memory, TotalRecall, developed to encourage authentic and idiomatic use in second language writing. TotalRecall is a bilingual concordancer that support search query in English or Chinese for relevant sentences and translations. Although initially intended for learners of English as Foreign Language (EFL) in Taiwan, it is a gold mine of texts in English or Mandarin Chinese. TotalRecall is particularly useful for those who write in or translate into a foreign language. We exploited and structured existing high-quality translations from bilingual corpora from a Taiwan-based Sinorama Magazine and Official Records of Hong Kong Legislative Council to build a bilingual concordance. Novel approaches were taken to provide highprecision bilingual alignment on the subsentential and lexical levels. A browserbased user interface was developed for ease of access over the Internet. Users can search for word, phrase or expression in English or Mandarin. The Web-based user interface facilitates the recording of the user actions to provide data for further research.