sparsity, learning on graphs, kernel methods, high-dimensional data, feature selection, interpretability