Variational graph autoencoders for multiview canonical correlation analysis

We present a novel approach for multiview canonical correlation analysis based on variational graph neural network model. propose nonlinear model which takes into account the available graph-based geometric constraints while being scalable to large-scale datasets with multiple views. This combines probabilistic interpretation of CCA an autoencoder architecture convolutional layers. Experiments proposed method are conducted classification, clustering, and recommendation tasks real datasets. The algorithm is competitive state-of-the-art representation learning techniques, in addition robust instances missing