We address performance fairness for speaker verification using the
adver...
This paper proposes a novel formulation of prototypical loss with mixup ...
We present a two-stage framework for deep one-class classification. We f...
Thus far, end-to-end (E2E) models have not been shown to outperform
stat...