We propose a novel end-to-end graph recommendation model called the Collaborative Variational Graph Auto-Encoder (CVGA), which uses the information propagation and aggregation paradigms to encode user-item collaborative relationships on the user-item interaction bipartite graph. These relationships are ...
to solve graph semi-supervised learning Problem1(see Methods). Our Assumption1(see Methods) clarifies the capability of an autoencoder to obtain low-rank solution. Based on Assumption1, an autoencoder with manifold loss as we defined in
Variational autoencoder based bipartite network embedding by integrating local and global structureBipartite network embeddingLocal and global structureVariational autoencoderNonlinear structureAs a powerful tool for machine learning on the graph, network embedding, which projects nodes into low-dimensional ...