1 paper
Yingjian Yu, Haiyan Guo, Tianshun Wang +2
The advancement of deep learning-based speech synthesis has significantly increased the diversity of deepfake speech, posing threats to voice authentication. While centralized trai…