Fig. 6: A basic variogram auto-encoder (VGAE).
From: Application of Artificial Intelligence In Drug-target Interactions Prediction: A Review
The input graph’s adjacency matrix A and the node’s identity matrix X are used to learn the mean μ and variance σ of the node’s low-dimensional vector representation through an encoder (graph convolutional network), which is then used to generate the graph with a decoder (link prediction).
