SpeechT: Findings of the First Mentorship in Speech Translation

TL;DR

This paper reports on the first mentorship in speech translation, highlighting activities like data augmentation and system comparison across multiple languages, aiming to advance speech translation research.

Contribution

It introduces the first mentorship program in speech translation, involving diverse languages and exploring data techniques and system architectures.

Findings

Explored data augmentation techniques for speech translation.

Compared end-to-end and cascaded systems across languages.

Provided insights into multilingual speech translation challenges.

Abstract

This work presents the details and findings of the first mentorship in speech translation (SpeechT), which took place in December 2024 and January 2025. To fulfil the mentorship requirements, the participants engaged in key activities, including data preparation, modelling, and advanced research. The participants explored data augmentation techniques and compared end-to-end and cascaded speech translation systems. The projects covered various languages other than English, including Arabic, Bengali, Galician, Indonesian, Japanese, and Spanish.