Fig. 1
From: A Graph Feature Auto-Encoder for the prediction of unobserved node features on biological networks
A Depicts the GAE scheme of [12] tailored to the reconstruction of the adjacency matrix of the graph. In B we take the embedding matrix of a GAE and train an indirect regression task for the prediction of the expression values. C Illustrates our proposed GFAE approach for end-to-end learning of graph node features