Research
2026-09
Paper on deepfake detector generalization submitted to ICASSP 2027
A full paper on how audio deepfake detectors generalize to data they were not trained on has been submitted to ICASSP 2027 and is under review.
We submitted a full paper to ICASSP 2027 on the generalization of audio deepfake detectors: how well a detector holds up on recordings, generators and channels it did not see during training.
The paper is under review. We will share more once the review is complete.
Think a voice sample might be synthetic?
1 free analysis. No account needed.
Analyze Audio