Our voice encodes a uniquely identifiable pattern which can be used to infer private attributes, such as gender or identity, that an individual might wish not to reveal when using a speech recognition service. To prevent attribute inference attacks alongside speech recognition tasks, we present a generative adversarial network, GenGAN, that synthesises voices that conceal the gender or identity of a speaker. The proposed network includes a generator with a U-Net architecture that learns to fool a discriminator. We condition the generator only on gender information and use an adversarial loss between signal distortion and privacy preservation. We show that GenGAN improves the trade-off between privacy and utility compared to privacy-preserving representation learning methods that consider gender information as a sensitive attribute to protect.
Cite as: Stoidis, D., Cavallaro, A. (2022) Generating gender-ambiguous voices for privacy-preserving speech recognition. Proc. Interspeech 2022, 4237-4241, doi: 10.21437/Interspeech.2022-11322
@inproceedings{stoidis22_interspeech,
author={Dimitrios Stoidis and Andrea Cavallaro},
title={{Generating gender-ambiguous voices for privacy-preserving speech recognition}},
year=2022,
booktitle={Proc. Interspeech 2022},
pages={4237--4241},
doi={10.21437/Interspeech.2022-11322}
}