A dual-branch multimodal model combining ECAPA-TDNN on anonymized audio and BERT on transcripts outperforms prior attackers on five of seven VPAC benchmarks and reaches SOTA with augmentation.
Can Authorship Attribution Models Distinguish Speakers in Speech Transcripts?
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
-
VoxATtack: A Multimodal Attack on Voice Anonymization Systems
A dual-branch multimodal model combining ECAPA-TDNN on anonymized audio and BERT on transcripts outperforms prior attackers on five of seven VPAC benchmarks and reaches SOTA with augmentation.