We propose a new hierarchical representation of discrete data sets living on graphs. The approach takes advantage of recent works on graph regularization. The role of the merging criterion that is common to hierarchical representations is greatly reduced due to the regularization step. The regularization is performed recursively with a decreasing fidelity parameter. This yields a robust represe...

Geometric matrix completion (GMC) has been proposed for recommendation by integrating the relationship (link) graphs among users/items into matrix completion (MC) . Traditional GMC methods typically adopt graph regularization to impose smoothness priors for MC. Recently, geometric deep learning on graphs (GDLG) is proposed to solve the GMC problem, showing better performance than existing GMC m...