In the ever-evolving realm of machine translation, the pursuit of excellence is a driving force. Every year, the European Association for Machine Translation (EAMT) recognizes outstanding contributions to the field through the Best Thesis Award. Today, we are proud to announce that our very own Marco Gaido has won the 2023 edition of the award for his thesis titled “Direct Speech Translation Toward High-Quality, Inclusive, and Augmented Systems.”

The thesis delves into enhancing direct models for speech translation on various aspects. Besides proposing technical solutions to increase their translation quality and computational efficiency, it investigates specific behaviours that range from ethical concerns about gender disparity in the generated output to exploring the systems’ potential in “augmented translation” scenarios, where the accuracy in recognising and translating named entities plays a pivotal role.

In addition to Marco’s award, the EAMT Executive Committee also recognized the outstanding contributions of two other theses: Jannis Vamvas’s thesis, “Model-based Evaluation of Multilinguality,” from the University of Zurich, Switzerland, supervised by Rico Sennrich and Lena A. Jäger, as well as Javier Iranzo-Sánchez’s thesis, “Streaming Neural Speech Translation,” from UPV, Spain, supervised by Jorge Civera and Alfons Juan, demonstrated excellence in their respective areas of research. As we celebrate Marco’s accomplishments, we extend our heartfelt congratulations to them for their dedication and scholarly contributions to the field of machine translation.

Marco will present his thesis at the upcoming EAMT Conference to be held in Sheffield, UK, from June 24th to 27th, 2024.

More details on the award con be found in the EAMT website: