Abstract: Motivated by the cross-modal information processing mechanism of human brain, vertebrates, and invertebrates, we propose a multiple-input–multiple-output (MIMO) variational autoencoder (VAE) ...