Reviews: Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels

Pros. - The derivation of explicit formulae, which I think is correct, is neat and requires work. Since the title is GNTK, it may be better to include analysis of graph convolutional layers. Theorems 4.2 and 4.3 are more like technical lemmas whose usefulnesses are in doubt per se. Furthermore, the argument of polynomial'' sample complexity relies on the condition given in line 193 (please consider numbering the equations). There is no discussion in relation to the applicability of this condition.