Table 1.

Graph kernels

Kernel	Formula
p-step random walk	$K={\left(aI-\tilde{L}\right)}^p$
Diffusion process	$K=\text{exp}\left(-{\sigma }^2/2\tilde{L}\right)$
Regularized Laplacian	$K={\left(I+{\sigma }^2\tilde{L}\right)}^{-1}$
Commute time kernel	$K=L^{+}$
Inverse cosine	$\text{cos}\tilde{L}\pi /4$

Definition of graph kernels based on the normalized Laplacian ($\tilde{L}$), and pseudoinverse of the Laplacian ($L^{+}$), where a, p, and σ are given parameters.