WORKSHOPS
The practices and challenges of post-editing machine translation at the European Parliament
This workshop will focus on the post-editing machine translation process and practices in the Romanian language Unit of the Directorate General for Translation, European Parliament. After a brief introduction of the translation process at the European Parliament, a classification of challenges posed by machine translation will be presented, with examples from translated texts. The second part of the workshop, more practical, will involve an analysis and discussion of the solutions proposed by machine translation in a text translated at the European Parliament.
luciana-sabina.tcaciuc@europarl.europa.eu
Nadia FARCAȘ, Quality Officer of the Romanian
Language Department in DGT, European Commission (Luxembourg)
Post-editing MT@DGT – challenges and practical approaches
As the volume of content in need of translation continues to grow and as the quality of machine translation output gets better and better, MT has become an indispensable daily tool for the DGT translators. Which does not mean that it is always reliable. And, seeing that our translations can have very real consequences for the European citizens and businesses, their quality and reliability are essential. How, then, do we make the best use of the MT potential, while making sure we avoid its potential pitfalls?
Nadia.FARCAS@ec.europa.eu
Veronica DURBACĂ, Lawyer-linguist with the Romanian Translation
Unit, Directorate-General for Multilingualism of the Court of Justice of the
European Union (Luxembourg)
The Limitations of GenAI, and MTPE while producing Romanian versions of the CJEU decisions drafted in French – translation workshop
We shall examine together a passage from a decision already delivered by the Court of Justice of the European Union and observe the cases where LLM (GPT-3.5 and Gemini) and machine translation engines (Google Translate, DeepL and eTranslation) would be useful, as well as cases where they would be less useful. This is a hypothetical scenario. Only DeepL and eTranslation are actually used in the translation process at the CJEU.
Veronica.Durbaca@curia.europa.eu