Session

At a time in which Machine Translation (MT) is becoming more and more integrated into translation workflows, the challenges translators are required to face are evolving at an unprecedented pace. Starting from an overview of the main shortcomings associated with the use of MT and other widely-used translation technologies, notably Computer-Assisted Translation tools, this course will showcase IT tools and methods that make it possible to overcome such shortcomings. By adopting a practical, problem-based approach, we will focus on translation-relevant problems such as the retrieval of high-quality texts for documentation purposes, the identification of appropriate terminological equivalents, as well as stylistic checks on the ‘naturalness’ or idiomaticity of target texts.

In the first part of the course, we will cover well-established technologies and methods, in particular search engines (focusing on advanced search functionalities) and text corpora, i.e. large collections of texts collected for translation-inherent purposes (including translation, revision and terminology work). Participants will experience first-hand how to consult multilingual corpora using freely available software and how to search large reference corpora, progressing from simpler searches to more sophisticated ones.

Once we have laid the grounds, the second part of the course will be devoted to a comparison between these technologies and new systems based on generative Artificial Intelligence (e.g. ChatGPT) when applied to documentation and terminology tasks. The ultimate aim will be that of exploring their potential, but also raising awareness on their limitations.

The course is aimed at practising translators and translation students interested in how technologies can be integrated in the translation workflow.