Home
Resume
Publications
© 2026 Bartłomiej Marek
--:--
DeepFakes
Are Audio DeepFake Detection Models Polyglots?
Most audio deepfake detectors are trained on English. We benchmark zero-shot and adaptation strategies across languages and show that target-language data matters far more than English-only training.
Bartłomiej Marek
,
Piotr Kawa
,
Piotr Syga
arXiv
PDF
GitHub
Cite
×