Erratum to: Latent discriminative representation learning for speaker recognition