Figure 2From: Deep neural network-based bottleneck feature and denoising autoencoder-based dereverberation for distant-talking speaker identificationGraphical representation of an RBM.Back to article page