Yasmin Moslem, a researcher at the ADAPT Centre and Trinity College Dublin (TCD), has had two papers accepted for presentation at the prestigious International Conference on Spoken Language Translation (IWSLT) 2025.
The first paper, Efficient Speech Translation through Model Compression and Knowledge Distillation, addresses the growing challenge of deploying large audio-language models for speech translation in resource-constrained settings. The study presents techniques to reduce the computational demands of these models (available online here).
The second paper, Bemba Speech Translation: Exploring a Low-Resource African Language, is co-authored with Muhammad Hazim Al-Farouq and Aman Kassahun Wassie. It focuses on building effective speech translation systems for Bemba, a low-resource language spoken in parts of southern Africa. This work contributes to the IWSLT 2025 track on low-resource language translation (available online here).
IWSLT is an annual scientific conference, associated with shared tasks on spoken language translation, where both scientific papers and system descriptions are presented and will take place this year in Vienna, Austria, from 31st July to 1 August 2025.
