Link prediction, personalized graph search, fraud detection, and many such graph mining problems revolve around the computation of the most ``similar'' $k$ nodes to a given query node. One widely used
