1 paper
Rongjin Li, Weibin Zhang, Dongpeng Chen
State-of-the-art speaker recognition systems are trained with a large amount of human-labeled training data set. Such a training set is usually composed of various data sources to…