Alina KARAKANTA

Poster and Demo chair

I am Alina KARAKANTA , Assistant Professor of Machine Translation at LULC, University of Leiden. My research is motivated by a natural curiosity around translation, its properties and on how to enhance and augment users of translation. Starting my career as a professional translator incentivised me to pursue ways to make my work more efficient. One of my goals is to study and extend automation in areas where language technology has been considered unfruitful, in particular to subtitling. The Human-in-the-loop Applied ML workshop is therefore an ideal opportunity to converse with researchers creating an interaction between humans and machine learning systems.

My research has been navigating disciplinary boundaries between translation studies, computational linguistics and computer science. I received my PhD in 2022 from the University of Trento and Fondazione Bruno Kessler on the topic of Automatic subtitling. The thesis contributions included: i) creating and publicly releasing new subtitling corpora and test benchmarks, such as MuST-Cinema, ii) developing new computational methods and models for different subtitling scenarios (offline and live subtitling), and iii) implementing novel evaluation metrics and conducting human evaluations of the developed technologies. The research outputs contributed to increasing mutual awareness between the fields of Audiovisual translation and MT. Dissemination of research outputs to the general public, students, professionals and academics has been an integral part of the thesis, through invited lectures (Ghent University, Warsaw University) and talks (TranslatingEurope workshop, TRAMA Setmana de la TAV).

In 2018 I was a research assistant for the SFB-funded project “Modelling human translation with a noisy channel”, where I studied variation in translation and interpreting. As a master’s student in Computational Linguistics at the University of Saarland, I conducted research projects on low-resource languages, extending machine translation to languages with little computational resources. Along these lines, between 2018-2020 I was a co-organiser of the Workshop on technologies for MT of low-resource languages (LoResMT) and editor of two special issues in the Machine Translation journal.