Different Speech Translation Models Encode and Translate Speaker Gender Differently
Dennis Fucci, Marco Gaido, Matteo Negri, Luisa Bentivogli, Andre Martins, Giuseppe Attanasio
Abstract
Recent studies on interpreting the hidden states of speech models have shown their ability to capture speaker-specific features, including gender. Does this finding also hold for speech translation (ST) models? If so, what are the implications for the speaker’s gender assignment in translation? We address these questions from an interpretability perspective, using probing methods to assess gender encoding across diverse ST models. Results on three language directions (English → French/Italian/Spanish) indicate that while traditional encoder-decoder models capture gender information, newer architectures—integrating a speech encoder with a machine translation system via adapters—do not. We also demonstrate that low gender encoding capabilities result in systems’ tendency toward a masculine default, a translation bias that is more pronounced in newer architectures.
BibTeX
@inproceedings{fucci-etal-2025-different,
title = "Different Speech Translation Models Encode and Translate Speaker Gender Differently",
author = "Fucci, Dennis and
Gaido, Marco and
Negri, Matteo and
Bentivogli, Luisa and
Martins, Andre and
Attanasio, Giuseppe",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-short.78/",
doi = "10.18653/v1/2025.acl-short.78",
pages = "1005--1019",
ISBN = "979-8-89176-252-7"
}