Abstract This paper discusses neural machine translation (NMT), a new paradigm in the MT field, comparing quality of NMT systems with statistical by describing three studies using automatic and human evaluation methods. Automatic results presented for are very promising, however evaluations show mixed results. We report increases fluency but inconsistent adequacy post-editing effort. undoubtedly represents step forward one that community should be careful not to oversell.

The profession of translation is undergoing enormous change, expedited by the global recession that began in 2007–2008. Government policies and an intensive focus on cost has resulted a large scale move to freelance or contingent work, leaving worker precarious position with regard rights undermining his her agency. This shift exemplified particularly vendor model widespread specialised work. Related downward pressure costs productivity technologisation translation, tools becoming necessity...

Machine translation is currently undergoing a paradigm shift from statistical to neural network models. Neural machine (NMT) difficult conceptualise for students, especially without context. This article describes short in-class evaluation exercise compare and MT, including details of student results follow-on discussions. As part this exercise, students carry out evaluations two types MT output using three quality assurance (TQA) metrics: adequacy, post-editing productivity, simple error...

Abstract In the context of recent improvements in quality machine translation (MT) output and new use cases being found for that output, this article reports on an experiment using statistical neural MT systems to translate literature. Six professional translators with experience literary produced English-to-Catalan translations under three conditions: from scratch, post-editing, post-editing. They provided feedback before after via questionnaires interviews. While all participants prefer...

In the wake of platform economy's transformative influence on translation work, this study aims to address a critical concern: alignment workers' labour conditions with principles decent work. Through quantitative analysis subset questionnaire data collected from translators in Turkey engaging various digital platforms, findings suggest substantial disparities meeting six fundamental as defined by International Labour Organization (ILO). These include insufficient earnings, excessive and...

Abstract Translators have worked with the assistance of computers for many years, usually translating whole texts, divided into segments but in sequential order. In order to maximise efficiency and inspired by similar moves tech industry predictions Industry 4.0, large translation companies begun break tasks down smaller chunks rigidly define monitor processes. This is particularly true platform-mediated work, highly collaborative workflows, multimedia work that requires near-live turnaround...

This article is both an editorial introduction to the guest-edited special issue of JoSTrans on Translation Automation and Sustainability, a position paper in which we propose model for evaluating sustainable use automation technology translation beyond. As grounding notions, reviews definitions considers urgency sustainability. Thereafter adaptation Elkington’s (1997) triple bottom line, giving equal weight evaluation based people, planet, performance, describing each these elements turn....

Shared tasks are increasingly common in our field, and new challenges suggested at almost every conference workshop. However, as this has become an established way of pushing research forward, it is important to discuss how we researchers organise participate shared tasks, make that information available the community allow further improvements. In paper, present a number ethical issues along with other areas concern related competitive nature tasks. As such could potentially impact on...

AbstractThis study compares consistency in target texts produced using translation memory (TM) with that of statistical machine (SMT), where the SMT engine is trained on same as are reused TM workflow. These comparisons focus specifically noun and verb inconsistencies, such inconsistencies appear to be highly prevalent data. The substitutes inconsistent text nouns verbs for consistent from output test whether this results improvements overall an ‘laundered’ data performs better than baseline...

Preview this article: Fair MT, Page 1 of < Previous page | Next > /docserver/preview/fulltext/ts.00018.int-1.gif

This article presents the results of three AVT reception experiments with over 200 English-speaking participants who watched a 20-minute clip Mexican telenovela in different translation modalities: human-translated (HT), post-edited (PT) and machine-translated (MT). Participants answered questionnaire on narrative engagement, enjoyment, subtitles. The show that viewers have higher engagement PE than HT, but there is only statistically significant difference when compared to MT. When it comes...