Dr. Natalia Resende

The MTrill is a process-oriented research that aims at investigating how online freely available Machine Translation (MT) systems are impacting the acquisition and processing of English as a second language. The research will shed light on issues involving a central aspect of language acquisition: the so-called language binding, i.e., the ability to combine single words properly in a grammatical sentence in a second language. To pursue this goal, a syntactic priming experiment will be carried out in which participants will be tested as whether they produce the same syntactic structures that had been previously seen during a translation task using a MT. The project brings a brand-new methodological approach within MT process-oriented evaluation research and it goes beyond the existing state-of-the-art approaches since it will focus on translation processing complemented by product analysis (oral production). The project will be conducted under the supervision of Professor Andy Way and Doctor Monica Ward in the ADAPT Centre based in Dublin City university. A four- months secondment will be carried out in the Max Planck Institute of Psycholinguistics (MPI) in Nijmegen, The Netherlands, in the Neurobiology of Language research group under the supervision of Professor Peter Hagoort. The knowledge acquired during the secondment will be consolidated in the return phase through implementation of the experiments carried out for this project as well as through teaching and co-supervision activities. Results of this research will be disseminated through conferences and journal paper publications and will be communicated to multiple audiences through articles published in non-academic magazines, blogs and social media. The fellowship will attract academic networking opportunities, skill-set improvements and a personal and professional transformative experience to the researcher since it will establish her as the new leading light in research of an interdisciplinary field.

this project aimed at seeking for the typical features of post-edited texts and compared to human-translated texts and machine translated texts.

this is an ongoing project which has already resulted in outputs such as a conference paper publication and three talks on the topic. The goal of the project is to test the advantages and limitations of translating poetry with large language models and the best prompt techniques for the task.

This is an ongoing research project for which a funding proposal has been submitted. The primary aim is to investigate whether translators engage in critical thinking when post"editing AI"generated translations, or whether their decision"making is influenced by the AI output, thus constraining their choices.

This project explores how different instructional approaches impact students" use of Large Language Models (LLMs) in literary translation. As generative AI tools become increasingly integrated into translator training, as reflected in the 2024 European Language Industry Survey reporting that 63% of students now use LLMs, there is an urgent need to understand how pedagogical strategies shape students" interactions with these technologies. Using a pre-test/test mixed-methods design, this study triangulates data from keystroke logging (KL), eye-tracking (ET), and think-aloud protocols (TAP) to gain a multidimensional view of students" translation processes. The study compares the effects of direct (teacher-led) versus indirect (task-based) instruction on students" cognitive effort and strategy use when working with LLMs. Findings will inform best practices in translator education and AI literacy, helping students not only to use LLMs effectively but also to critically assess AI-generated content for translation purposes. The interdisciplinary team combines expertise in literary translation, NLP and psycholinguistics. Results will be shared through journal publication and conference presentations, contributing to the development of evidence-based approaches in teaching literary translation with AI.