Adapting BERT as a text-only ASV attacker on VoicePrivacy datasets yields mean EER 35% (some speakers 2%), driven by semantic keyword overlaps from LibriSpeech curation, prompting calls to revise evaluation datasets and move beyond global EER.
Its aim is to develop techniques to compromise the privacy of speakers that have been processed by seven anonymization systems
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
You Are What You Say: Exploiting Linguistic Content for VoicePrivacy Attacks
Adapting BERT as a text-only ASV attacker on VoicePrivacy datasets yields mean EER 35% (some speakers 2%), driven by semantic keyword overlaps from LibriSpeech curation, prompting calls to revise evaluation datasets and move beyond global EER.